longitudinal MPC: use reset() function instead of recreating the solver in (#24091)

* scons: add acados_template as dependency for lat and long mpc

* long MPC: use acados reset instead of recreating the solver

* long MPC: print timings and reset commented

* update acados x86_64

* update acados include folder

* update acados Python interface

* update acados reference commit to latest acados/master

* update x86 libs

* update comma two

* update acados again with commit 8ea8827fafb1b23b4c7da1c4cf650de1cbd73584

* update comma two

* update comma three

* update x86

Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
Co-authored-by: Comma Device <device@comma.ai>

pyextra/acados_template/acados_layout.json

@@ -668,6 +668,9 @@
         "hessian_approx": [
             "str"
         ],
+        "hpipm_mode": [
+            "str"
+        ],
         "regularize_method": [
             "str"
         ],

pyextra/acados_template/acados_ocp.py

@@ -2155,6 +2155,7 @@ class AcadosOcpOptions:
         self.__globalization_use_SOC = 0
         self.__full_step_dual = 0
         self.__eps_sufficient_descent = 1e-4
+        self.__hpipm_mode = 'BALANCE'
 
 
     @property
@@ -2165,6 +2166,21 @@ class AcadosOcpOptions:
         """
         return self.__qp_solver
 
+    @property
+    def hpipm_mode(self):
+        """
+        Mode of HPIPM to be used,
+
+        String in ('BALANCE', 'SPEED_ABS', 'SPEED', 'ROBUST').
+
+        Default: 'BALANCE'.
+
+        see https://cdn.syscop.de/publications/Frison2020a.pdf
+        and the HPIPM code:
+        https://github.com/giaf/hpipm/blob/master/ocp_qp/x_ocp_qp_ipm.c#L69
+        """
+        return self.__hpipm_mode
+
     @property
     def hessian_approx(self):
         """Hessian approximation.
@@ -2541,6 +2557,15 @@ class AcadosOcpOptions:
             raise Exception('Invalid collocation_type value. Possible values are:\n\n' \
                     + ',\n'.join(collocation_types) + '.\n\nYou have: ' + collocation_type + '.\n\nExiting.')
 
+    @hpipm_mode.setter
+    def hpipm_mode(self, hpipm_mode):
+        hpipm_modes = ('BALANCE', 'SPEED_ABS', 'SPEED', 'ROBUST')
+        if hpipm_mode in hpipm_modes:
+            self.__hpipm_mode = hpipm_mode
+        else:
+            raise Exception('Invalid hpipm_mode value. Possible values are:\n\n' \
+                    + ',\n'.join(hpipm_modes) + '.\n\nYou have: ' + hpipm_mode + '.\n\nExiting.')
+
     @hessian_approx.setter
     def hessian_approx(self, hessian_approx):
         hessian_approxs = ('GAUSS_NEWTON', 'EXACT')
@@ -2890,10 +2915,10 @@ class AcadosOcp:
         self.solver_options = AcadosOcpOptions()
         """Solver Options, type :py:class:`acados_template.acados_ocp.AcadosOcpOptions`"""
 		
-        self.acados_include_path = f'{acados_path}/include'
+        self.acados_include_path = os.path.join(acados_path, 'include').replace(os.sep, '/') # the replace part is important on Windows for CMake
-        """Path to acados include directory, type: string"""
+        """Path to acados include directory (set automatically), type: `string`"""
-        self.acados_lib_path = f'{acados_path}/lib'
+        self.acados_lib_path = os.path.join(acados_path, 'lib').replace(os.sep, '/') # the replace part is important on Windows for CMake
-        """Path to where acados library is located, type: string"""
+        """Path to where acados library is located, type: `string`"""
 
         import numpy
         self.cython_include_dirs = numpy.get_include()

pyextra/acados_template/acados_ocp_solver.py

@@ -54,6 +54,7 @@ from .acados_model import acados_model_strip_casadi_symbolics
 from .utils import is_column, is_empty, casadi_length, render_template,\
      format_class_dict, ocp_check_against_layout, np_array_to_list, make_model_consistent,\
      set_up_imported_gnsf_model, get_ocp_nlp_layout, get_python_interface_path
+from .builders import CMakeBuilder
 
 
 def make_ocp_dims_consistent(acados_ocp):
@@ -132,6 +133,15 @@ def make_ocp_dims_consistent(acados_ocp):
                             f'\nGot W_0[{cost.W.shape}], yref_0[{cost.yref_0.shape}]\n')
         dims.ny_0 = ny_0
 
+    elif cost.cost_type_0 == 'EXTERNAL':
+        if opts.hessian_approx == 'GAUSS_NEWTON' and opts.ext_cost_num_hess == 0 and model.cost_expr_ext_cost_custom_hess_0 is None:
+            print("\nWARNING: Gauss-Newton Hessian approximation with EXTERNAL cost type not possible!\n"
+            "got cost_type_0: EXTERNAL, hessian_approx: 'GAUSS_NEWTON.'\n"
+            "GAUSS_NEWTON hessian is only supported for cost_types [NON]LINEAR_LS.\n"
+            "If you continue, acados will proceed computing the exact hessian for the cost term.\n"
+            "Note: There is also the option to use the external cost module with a numerical hessian approximation (see `ext_cost_num_hess`).\n"
+            "OR the option to provide a symbolic custom hessian approximation (see `cost_expr_ext_cost_custom_hess`).\n")
+
     # path
     if cost.cost_type == 'LINEAR_LS':
         ny = cost.W.shape[0]
@@ -161,6 +171,15 @@ def make_ocp_dims_consistent(acados_ocp):
                             f'\nGot W[{cost.W.shape}], yref[{cost.yref.shape}]\n')
         dims.ny = ny
 
+    elif cost.cost_type == 'EXTERNAL':
+        if opts.hessian_approx == 'GAUSS_NEWTON' and opts.ext_cost_num_hess == 0 and model.cost_expr_ext_cost_custom_hess is None:
+            print("\nWARNING: Gauss-Newton Hessian approximation with EXTERNAL cost type not possible!\n"
+            "got cost_type: EXTERNAL, hessian_approx: 'GAUSS_NEWTON.'\n"
+            "GAUSS_NEWTON hessian is only supported for cost_types [NON]LINEAR_LS.\n"
+            "If you continue, acados will proceed computing the exact hessian for the cost term.\n"
+            "Note: There is also the option to use the external cost module with a numerical hessian approximation (see `ext_cost_num_hess`).\n"
+            "OR the option to provide a symbolic custom hessian approximation (see `cost_expr_ext_cost_custom_hess`).\n")
+
     # terminal
     if cost.cost_type_e == 'LINEAR_LS':
         ny_e = cost.W_e.shape[0]
@@ -183,6 +202,14 @@ def make_ocp_dims_consistent(acados_ocp):
             raise Exception('inconsistent dimension: regarding W_e, yref_e.')
         dims.ny_e = ny_e
 
+    elif cost.cost_type_e == 'EXTERNAL':
+        if opts.hessian_approx == 'GAUSS_NEWTON' and opts.ext_cost_num_hess == 0 and model.cost_expr_ext_cost_custom_hess_e is None:
+            print("\nWARNING: Gauss-Newton Hessian approximation with EXTERNAL cost type not possible!\n"
+            "got cost_type_e: EXTERNAL, hessian_approx: 'GAUSS_NEWTON.'\n"
+            "GAUSS_NEWTON hessian is only supported for cost_types [NON]LINEAR_LS.\n"
+            "If you continue, acados will proceed computing the exact hessian for the cost term.\n"
+            "Note: There is also the option to use the external cost module with a numerical hessian approximation (see `ext_cost_num_hess`).\n"
+            "OR the option to provide a symbolic custom hessian approximation (see `cost_expr_ext_cost_custom_hess`).\n")
 
     ## constraints
     # initial
@@ -622,15 +649,25 @@ def ocp_generate_external_functions(acados_ocp, model):
         generate_c_code_external_cost(model, 'terminal', opts)
 
 
-def ocp_render_templates(acados_ocp, json_file):
+def ocp_get_default_cmake_builder() -> CMakeBuilder:
+    """
+    If :py:class:`~acados_template.acados_ocp_solver.AcadosOcpSolver` is used with `CMake` this function returns a good first setting.
+    :return: default :py:class:`~acados_template.builders.CMakeBuilder`
+    """
+    cmake_builder = CMakeBuilder()
+    cmake_builder.options_on = ['BUILD_ACADOS_OCP_SOLVER_LIB']
+    return cmake_builder
+
+
+def ocp_render_templates(acados_ocp, json_file, cmake_builder=None):
 
     name = acados_ocp.model.name
 
     # setting up loader and environment
-    json_path = os.path.join(os.getcwd(), json_file)
+    json_path = os.path.abspath(json_file)
 
     if not os.path.exists(json_path):
-        raise Exception('{} not found!'.format(json_path))
+        raise Exception(f'Path "{json_path}" not found!')
 
     code_export_dir = acados_ocp.code_export_directory
     template_dir = code_export_dir
@@ -652,9 +689,14 @@ def ocp_render_templates(acados_ocp, json_file):
     out_file = f'acados_solver.pxd'
     render_template(in_file, out_file, template_dir, json_path)
 
-    in_file = 'Makefile.in'
+    if cmake_builder is not None:
-    out_file = 'Makefile'
+        in_file = 'CMakeLists.in.txt'
-    render_template(in_file, out_file, template_dir, json_path)
+        out_file = 'CMakeLists.txt'
+        render_template(in_file, out_file, template_dir, json_path)
+    else:
+        in_file = 'Makefile.in'
+        out_file = 'Makefile'
+        render_template(in_file, out_file, template_dir, json_path)
 
     in_file = 'acados_solver_sfun.in.c'
     out_file = f'acados_solver_sfunction_{name}.c'
@@ -764,25 +806,30 @@ class AcadosOcpSolver:
     """
     Class to interact with the acados ocp solver C object.
 
-        :param acados_ocp: type AcadosOcp - description of the OCP for acados
+        :param acados_ocp: type :py:class:`~acados_template.acados_ocp.AcadosOcp` - description of the OCP for acados
         :param json_file: name for the json file used to render the templated code - default: acados_ocp_nlp.json
         :param simulink_opts: Options to configure Simulink S-function blocks, mainly to activate possible Inputs and Outputs
     """
     if sys.platform=="win32":
         from ctypes import wintypes
-        dlclose = ctypes.WinDLL('kernel32', use_last_error=True).FreeLibrary
+        from ctypes import WinDLL
+        dlclose = WinDLL('kernel32', use_last_error=True).FreeLibrary
         dlclose.argtypes = [wintypes.HMODULE]
     else:
         dlclose = CDLL(None).dlclose
         dlclose.argtypes = [c_void_p]
 
     @classmethod
-    def generate(cls, acados_ocp, json_file='acados_ocp_nlp.json', simulink_opts=None):
+    def generate(cls, acados_ocp, json_file='acados_ocp_nlp.json', simulink_opts=None, cmake_builder: CMakeBuilder = None):
         """
         Generates the code for an acados OCP solver, given the description in acados_ocp.
             :param acados_ocp: type AcadosOcp - description of the OCP for acados
-            :param json_file: name for the json file used to render the templated code - default: acados_ocp_nlp.json
+            :param json_file: name for the json file used to render the templated code - default: `acados_ocp_nlp.json`
-            :param simulink_opts: Options to configure Simulink S-function blocks, mainly to activate possible Inputs and Outputs
+            :param simulink_opts: Options to configure Simulink S-function blocks, mainly to activate possible inputs and
+                   outputs; default: `None`
+            :param cmake_builder: type :py:class:`~acados_template.builders.CMakeBuilder` generate a `CMakeLists.txt` and use
+                   the `CMake` pipeline instead of a `Makefile` (`CMake` seems to be the better option in conjunction with
+                   `MS Visual Studio`); default: `None`
         """
         model = acados_ocp.model
         acados_ocp.code_export_directory = os.path.abspath(acados_ocp.code_export_directory)
@@ -810,25 +857,32 @@ class AcadosOcpSolver:
         ocp_formulation_json_dump(acados_ocp, simulink_opts, json_file)
 
         # render templates
-        ocp_render_templates(acados_ocp, json_file)
+        ocp_render_templates(acados_ocp, json_file, cmake_builder=cmake_builder)
         acados_ocp.json_file = json_file
 
 
     @classmethod
-    def build(cls, code_export_dir, with_cython=False):
+    def build(cls, code_export_dir, with_cython=False, cmake_builder: CMakeBuilder = None):
         """
         Builds the code for an acados OCP solver, that has been generated in code_export_dir
             :param code_export_dir: directory in which acados OCP solver has been generated, see generate()
             :param with_cython: option indicating if the cython interface is build, default: False.
+            :param cmake_builder: type :py:class:`~acados_template.builders.CMakeBuilder` generate a `CMakeLists.txt` and use
+                   the `CMake` pipeline instead of a `Makefile` (`CMake` seems to be the better option in conjunction with
+                   `MS Visual Studio`); default: `None`
         """
+        code_export_dir = os.path.abspath(code_export_dir)
         cwd=os.getcwd()
         os.chdir(code_export_dir)
         if with_cython:
             os.system('make clean_ocp_cython')
             os.system('make ocp_cython')
         else:
-            os.system('make clean_ocp_shared_lib')
+            if cmake_builder is not None:
-            os.system('make ocp_shared_lib')
+                cmake_builder.exec(code_export_dir)
+            else:
+                os.system('make clean_ocp_shared_lib')
+                os.system('make ocp_shared_lib')
         os.chdir(cwd)
 
 
@@ -856,11 +910,11 @@ class AcadosOcpSolver:
                     acados_ocp_json['dims']['N'])
 
 
-    def __init__(self, acados_ocp, json_file='acados_ocp_nlp.json', simulink_opts=None, build=True, generate=True):
+    def __init__(self, acados_ocp, json_file='acados_ocp_nlp.json', simulink_opts=None, build=True, generate=True, cmake_builder: CMakeBuilder = None):
 
         self.solver_created = False
         if generate:
-            self.generate(acados_ocp, json_file=json_file, simulink_opts=simulink_opts)
+            self.generate(acados_ocp, json_file=json_file, simulink_opts=simulink_opts, cmake_builder=cmake_builder)
 
         # load json, store options in object
         with open(json_file, 'r') as f:
@@ -873,14 +927,22 @@ class AcadosOcpSolver:
         code_export_directory = acados_ocp_json['code_export_directory']
 
         if build:
-            self.build(code_export_directory, with_cython=False)
+            self.build(code_export_directory, with_cython=False, cmake_builder=cmake_builder)
+
+        # prepare library loading
+        lib_prefix = 'lib'
+        lib_ext = '.so'
+        if os.name == 'nt':
+            lib_prefix = ''
+            lib_ext = ''
+        # ToDo: check for mac
 
         # Load acados library to avoid unloading the library.
         # This is necessary if acados was compiled with OpenMP, since the OpenMP threads can't be destroyed.
         # Unloading a library which uses OpenMP results in a segfault (on any platform?).
         # see [https://stackoverflow.com/questions/34439956/vc-crash-when-freeing-a-dll-built-with-openmp]
         # or [https://python.hotexamples.com/examples/_ctypes/-/dlclose/python-dlclose-function-examples.html]
-        libacados_name = 'libacados.so'
+        libacados_name = f'{lib_prefix}acados{lib_ext}'
         libacados_filepath = os.path.join(acados_lib_path, libacados_name)
         self.__acados_lib = CDLL(libacados_filepath)
         # find out if acados was compiled with OpenMP
@@ -892,8 +954,8 @@ class AcadosOcpSolver:
             print('acados was compiled with OpenMP.')
         else:
             print('acados was compiled without OpenMP.')
-
+        libacados_ocp_solver_name = f'{lib_prefix}acados_ocp_solver_{self.model_name}{lib_ext}'
-        self.shared_lib_name = f'{code_export_directory}/libacados_ocp_solver_{self.model_name}.so'
+        self.shared_lib_name = os.path.join(code_export_directory, libacados_ocp_solver_name)
 
         # get shared_lib
         self.shared_lib = CDLL(self.shared_lib_name)
@@ -959,6 +1021,17 @@ class AcadosOcpSolver:
         return self.status
 
 
+    def reset(self):
+        """
+        Sets current iterate to all zeros.
+        """
+        getattr(self.shared_lib, f"{self.model_name}_acados_reset").argtypes = [c_void_p]
+        getattr(self.shared_lib, f"{self.model_name}_acados_reset").restype = c_int
+        getattr(self.shared_lib, f"{self.model_name}_acados_reset")(self.capsule)
+
+        return
+
+
     def set_new_time_steps(self, new_time_steps):
         """
         Set new time steps.
@@ -1364,12 +1437,12 @@ class AcadosOcpSolver:
         return out[0]
 
 
-    def get_residuals(self):
+    def get_residuals(self, recompute=False):
         """
         Returns an array of the form [res_stat, res_eq, res_ineq, res_comp].
         """
         # compute residuals if RTI
-        if self.solver_options['nlp_solver_type'] == 'SQP_RTI':
+        if self.solver_options['nlp_solver_type'] == 'SQP_RTI' or recompute:
             self.shared_lib.ocp_nlp_eval_residuals.argtypes = [c_void_p, c_void_p, c_void_p]
             self.shared_lib.ocp_nlp_eval_residuals(self.nlp_solver, self.nlp_in, self.nlp_out)
 
@@ -1403,7 +1476,7 @@ class AcadosOcpSolver:
         Set numerical data inside the solver.
 
             :param stage: integer corresponding to shooting node
-            :param field: string in ['x', 'u', 'pi', 'lam', 't', 'p']
+            :param field: string in ['x', 'u', 'pi', 'lam', 't', 'p', 'xdot_guess', 'z_guess']
 
             .. note:: regarding lam, t: \n
                     the inequalities are internally organized in the following order: \n
@@ -1419,7 +1492,7 @@ class AcadosOcpSolver:
         cost_fields = ['y_ref', 'yref']
         constraints_fields = ['lbx', 'ubx', 'lbu', 'ubu']
         out_fields = ['x', 'u', 'pi', 'lam', 't', 'z', 'sl', 'su']
-        mem_fields = ['xdot_guess']
+        mem_fields = ['xdot_guess', 'z_guess']
 
         # cast value_ to avoid conversion issues
         if isinstance(value_, (float, int)):
@@ -1663,11 +1736,21 @@ class AcadosOcpSolver:
         """
         Set options of the solver.
 
-            :param field: string, e.g. 'print_level', 'rti_phase', 'initialize_t_slacks', 'step_length', 'alpha_min', 'alpha_reduction', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC'
+            :param field: string, e.g. 'print_level', 'rti_phase', 'initialize_t_slacks', 'step_length', 'alpha_min', 'alpha_reduction', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC', 'qp_tol_stat', 'qp_tol_eq', 'qp_tol_ineq', 'qp_tol_comp', 'qp_tau_min', 'qp_mu0'
-            :param value: of type int, float
+
+            :param value: of type int, float, string
+
+            - qp_tol_stat: QP solver tolerance stationarity
+            - qp_tol_eq: QP solver tolerance equalities
+            - qp_tol_ineq: QP solver tolerance inequalities
+            - qp_tol_comp: QP solver tolerance complementarity
+            - qp_tau_min: for HPIPM QP solvers: minimum value of barrier parameter in HPIPM
+            - qp_mu0: for HPIPM QP solvers: initial value for complementarity slackness
+            - warm_start_first_qp: indicates if first QP in SQP is warm_started
         """
-        int_fields = ['print_level', 'rti_phase', 'initialize_t_slacks', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC']
+        int_fields = ['print_level', 'rti_phase', 'initialize_t_slacks', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC', 'warm_start_first_qp']
-        double_fields = ['step_length', 'tol_eq', 'tol_stat', 'tol_ineq', 'tol_comp', 'alpha_min', 'alpha_reduction', 'eps_sufficient_descent']
+        double_fields = ['step_length', 'tol_eq', 'tol_stat', 'tol_ineq', 'tol_comp', 'alpha_min', 'alpha_reduction', 'eps_sufficient_descent',
+        'qp_tol_stat', 'qp_tol_eq', 'qp_tol_ineq', 'qp_tol_comp', 'qp_tau_min', 'qp_mu0']
         string_fields = ['globalization']
 
         # check field availability and type

pyextra/acados_template/acados_ocp_solver_pyx.pyx

@@ -112,6 +112,13 @@ cdef class AcadosOcpSolverCython:
         return acados_solver.acados_solve(self.capsule)
 
 
+    def reset(self):
+        """
+        Sets current iterate to all zeros.
+        """
+        return acados_solver.acados_reset(self.capsule)
+
+
     def set_new_time_steps(self, new_time_steps):
         """
         Set new time steps.
@@ -450,12 +457,12 @@ cdef class AcadosOcpSolverCython:
         return out
 
 
-    def get_residuals(self):
+    def get_residuals(self, recompute=False):
         """
         Returns an array of the form [res_stat, res_eq, res_ineq, res_comp].
         """
         # compute residuals if RTI
-        if self.nlp_solver_type == 'SQP_RTI':
+        if self.nlp_solver_type == 'SQP_RTI' or recompute:
             acados_solver_common.ocp_nlp_eval_residuals(self.nlp_solver, self.nlp_in, self.nlp_out)
 
         # create output array
@@ -504,7 +511,7 @@ cdef class AcadosOcpSolverCython:
         cost_fields = ['y_ref', 'yref']
         constraints_fields = ['lbx', 'ubx', 'lbu', 'ubu']
         out_fields = ['x', 'u', 'pi', 'lam', 't', 'z', 'sl', 'su']
-        mem_fields = ['xdot_guess']
+        mem_fields = ['xdot_guess', 'z_guess']
 
         field = field_.encode('utf-8')
 
@@ -635,11 +642,21 @@ cdef class AcadosOcpSolverCython:
         """
         Set options of the solver.
 
-            :param field: string, e.g. 'print_level', 'rti_phase', 'initialize_t_slacks', 'step_length', 'alpha_min', 'alpha_reduction'
+            :param field: string, e.g. 'print_level', 'rti_phase', 'initialize_t_slacks', 'step_length', 'alpha_min', 'alpha_reduction', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC', 'qp_tol_stat', 'qp_tol_eq', 'qp_tol_ineq', 'qp_tol_comp', 'qp_tau_min', 'qp_mu0'
-            :param value: of type int, float
+
+            :param value: of type int, float, string
+
+            - qp_tol_stat: QP solver tolerance stationarity
+            - qp_tol_eq: QP solver tolerance equalities
+            - qp_tol_ineq: QP solver tolerance inequalities
+            - qp_tol_comp: QP solver tolerance complementarity
+            - qp_tau_min: for HPIPM QP solvers: minimum value of barrier parameter in HPIPM
+            - qp_mu0: for HPIPM QP solvers: initial value for complementarity slackness
+            - warm_start_first_qp: indicates if first QP in SQP is warm_started
         """
-        int_fields = ['print_level', 'rti_phase', 'initialize_t_slacks', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC']
+        int_fields = ['print_level', 'rti_phase', 'initialize_t_slacks', 'qp_warm_start', 'line_search_use_sufficient_descent', 'full_step_dual', 'globalization_use_SOC', 'warm_start_first_qp']
-        double_fields = ['step_length', 'tol_eq', 'tol_stat', 'tol_ineq', 'tol_comp', 'alpha_min', 'alpha_reduction', 'eps_sufficient_descent']
+        double_fields = ['step_length', 'tol_eq', 'tol_stat', 'tol_ineq', 'tol_comp', 'alpha_min', 'alpha_reduction', 'eps_sufficient_descent',
+        'qp_tol_stat', 'qp_tol_eq', 'qp_tol_ineq', 'qp_tol_comp', 'qp_tau_min', 'qp_mu0']
         string_fields = ['globalization']
 
         # encode

pyextra/acados_template/acados_sim.py

@@ -294,9 +294,9 @@ class AcadosSim:
         self.solver_options = AcadosSimOpts()
         """Solver Options, type :py:class:`acados_template.acados_sim.AcadosSimOpts`"""
 
-        self.acados_include_path = f'{acados_path}/include'
+        self.acados_include_path = os.path.join(acados_path, 'include').replace(os.sep, '/') # the replace part is important on Windows for CMake
-        """Path to acados include directors (set automatically), type: `string`"""
+        """Path to acados include directory (set automatically), type: `string`"""
-        self.acados_lib_path = f'{acados_path}/lib'
+        self.acados_lib_path = os.path.join(acados_path, 'lib').replace(os.sep, '/') # the replace part is important on Windows for CMake
         """Path to where acados library is located (set automatically), type: `string`"""
 
         self.code_export_directory = 'c_generated_code'

pyextra/acados_template/acados_sim_solver.py

@@ -47,6 +47,7 @@ from .acados_ocp import AcadosOcp
 from .acados_model import acados_model_strip_casadi_symbolics
 from .utils import is_column, render_template, format_class_dict, np_array_to_list,\
      make_model_consistent, set_up_imported_gnsf_model, get_python_interface_path
+from .builders import CMakeBuilder
 
 
 def make_sim_dims_consistent(acados_sim):
@@ -111,7 +112,17 @@ def sim_formulation_json_dump(acados_sim, json_file='acados_sim.json'):
         json.dump(sim_json, f, default=np_array_to_list, indent=4, sort_keys=True)
 
 
-def sim_render_templates(json_file, model_name, code_export_dir):
+def sim_get_default_cmake_builder() -> CMakeBuilder:
+    """
+    If :py:class:`~acados_template.acados_sim_solver.AcadosSimSolver` is used with `CMake` this function returns a good first setting.
+    :return: default :py:class:`~acados_template.builders.CMakeBuilder`
+    """
+    cmake_builder = CMakeBuilder()
+    cmake_builder.options_on = ['BUILD_ACADOS_SIM_SOLVER_LIB']
+    return cmake_builder
+
+
+def sim_render_templates(json_file, model_name, code_export_dir, cmake_options: CMakeBuilder = None):
     # setting up loader and environment
     json_path = os.path.join(os.getcwd(), json_file)
 
@@ -129,9 +140,15 @@ def sim_render_templates(json_file, model_name, code_export_dir):
     out_file = f'acados_sim_solver_{model_name}.h'
     render_template(in_file, out_file, template_dir, json_path)
 
-    in_file = 'Makefile.in'
+    # Builder
-    out_file = f'Makefile'
+    if cmake_options is not None:
-    render_template(in_file, out_file, template_dir, json_path)
+        in_file = 'CMakeLists.in.txt'
+        out_file = 'CMakeLists.txt'
+        render_template(in_file, out_file, template_dir, json_path)
+    else:
+        in_file = 'Makefile.in'
+        out_file = 'Makefile'
+        render_template(in_file, out_file, template_dir, json_path)
 
     in_file = 'main_sim.in.c'
     out_file = f'main_sim_{model_name}.c'
@@ -161,15 +178,19 @@ def sim_generate_casadi_functions(acados_sim):
     elif integrator_type == 'GNSF':
         generate_c_code_gnsf(model, opts)
 
+
 class AcadosSimSolver:
     """
     Class to interact with the acados integrator C object.
 
-    :param acados_sim: type :py:class:`acados_template.acados_ocp.AcadosOcp` (takes values to generate an instance :py:class:`acados_template.acados_sim.AcadosSim`) or :py:class:`acados_template.acados_sim.AcadosSim`
+        :param acados_sim: type :py:class:`~acados_template.acados_ocp.AcadosOcp` (takes values to generate an instance :py:class:`~acados_template.acados_sim.AcadosSim`) or :py:class:`~acados_template.acados_sim.AcadosSim`
-    :param json_file: Default: 'acados_sim.json'
+        :param json_file: Default: 'acados_sim.json'
-    :param build: Default: True
+        :param build: Default: True
+        :param cmake_builder: type :py:class:`~acados_template.utils.CMakeBuilder` generate a `CMakeLists.txt` and use
+            the `CMake` pipeline instead of a `Makefile` (`CMake` seems to be the better option in conjunction with
+            `MS Visual Studio`); default: `None`
     """
-    def __init__(self, acados_sim_, json_file='acados_sim.json', build=True):
+    def __init__(self, acados_sim_, json_file='acados_sim.json', build=True, cmake_builder: CMakeBuilder = None):
 
         self.solver_created = False
 
@@ -203,12 +224,16 @@ class AcadosSimSolver:
         code_export_dir = acados_sim.code_export_directory
         if build:
             # render templates
-            sim_render_templates(json_file, model_name, code_export_dir)
+            sim_render_templates(json_file, model_name, code_export_dir, cmake_builder)
 
-            ## Compile solver
+            # Compile solver
             cwd = os.getcwd()
+            code_export_dir = os.path.abspath(code_export_dir)
             os.chdir(code_export_dir)
-            os.system('make sim_shared_lib')
+            if cmake_builder is not None:
+                cmake_builder.exec(code_export_dir)
+            else:
+                os.system('make sim_shared_lib')
             os.chdir(cwd)
 
         self.sim_struct = acados_sim
@@ -234,8 +259,15 @@ class AcadosSimSolver:
             print('acados was compiled without OpenMP.')
 
         # Ctypes
-        shared_lib = f'{code_export_dir}/libacados_sim_solver_{model_name}.so'
+        lib_prefix = 'lib'
-        self.shared_lib = CDLL(shared_lib)
+        lib_ext = '.so'
+        if os.name == 'nt':
+            lib_prefix = ''
+            lib_ext = ''
+        self.shared_lib_name = os.path.join(code_export_dir, f'{lib_prefix}acados_sim_solver_{model_name}{lib_ext}')
+        print(f'self.shared_lib_name = "{self.shared_lib_name}"')
+        
+        self.shared_lib = CDLL(self.shared_lib_name)
 
 
         # create capsule

pyextra/acados_template/builders.py

@@ -0,0 +1,116 @@
+# -*- coding: future_fstrings -*-
+#
+# Copyright 2019 Gianluca Frison, Dimitris Kouzoupis, Robin Verschueren,
+# Andrea Zanelli, Niels van Duijkeren, Jonathan Frey, Tommaso Sartor,
+# Branimir Novoselnik, Rien Quirynen, Rezart Qelibari, Dang Doan,
+# Jonas Koenemann, Yutao Chen, Tobias Schöls, Jonas Schlagenhauf, Moritz Diehl
+#
+# This file is part of acados.
+#
+# The 2-Clause BSD License
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+# 1. Redistributions of source code must retain the above copyright notice,
+# this list of conditions and the following disclaimer.
+#
+# 2. Redistributions in binary form must reproduce the above copyright notice,
+# this list of conditions and the following disclaimer in the documentation
+# and/or other materials provided with the distribution.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+# POSSIBILITY OF SUCH DAMAGE.;
+#
+
+import os
+import sys
+from subprocess import call
+
+
+class CMakeBuilder:
+    """
+    Class to work with the `CMake` build system.
+    """
+    def __init__(self):
+        self._source_dir = None  # private source directory, this is set to code_export_dir
+        self.build_dir = 'build'
+        self._build_dir = None  # private build directory, usually rendered to abspath(build_dir)
+        self.generator = None
+        """Defines the generator, options can be found via `cmake --help` under 'Generator'. Type: string. Linux default 'Unix Makefiles', Windows 'Visual Studio 15 2017 Win64'; default value: `None`."""
+        # set something for Windows
+        if os.name == 'nt':
+            self.generator = 'Visual Studio 15 2017 Win64'
+        self.build_targets = None
+        """A comma-separated list of the build targets, if `None` then all targets will be build; type: List of strings; default: `None`."""
+        self.options_on = None
+        """List of strings as CMake options which are translated to '-D Opt[0]=ON -D Opt[1]=ON ...'; default: `None`."""
+
+    # Generate the command string for handling the cmake command.
+    def get_cmd1_cmake(self):
+        defines_str = ''
+        if self.options_on is not None:
+            defines_arr = [f' -D{opt}=ON' for opt in self.options_on]
+            defines_str = ' '.join(defines_arr)
+        generator_str = ''
+        if self.generator is not None:
+            generator_str = f' -G"{self.generator}"'
+        return f'cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX="{self._source_dir}"{defines_str}{generator_str} -Wdev -S"{self._source_dir}" -B"{self._build_dir}"'
+
+    # Generate the command string for handling the build.
+    def get_cmd2_build(self):
+        import multiprocessing
+        cmd = f'cmake --build "{self._build_dir}" --config Release -j{multiprocessing.cpu_count()}'
+        if self.build_targets is not None:
+            cmd += f' -t {self.build_targets}'
+        return cmd
+
+    # Generate the command string for handling the install command.
+    def get_cmd3_install(self):
+        return f'cmake --install "{self._build_dir}"'
+
+    def exec(self, code_export_directory):
+        """
+        Execute the compilation using `CMake` with the given settings.
+        :param code_export_directory: must be the absolute path to the directory where the code was exported to
+        """
+        if(os.path.isabs(code_export_directory) is False):
+            print(f'(W) the code export directory "{code_export_directory}" is not an absolute path!')
+        self._source_dir = code_export_directory
+        self._build_dir = os.path.abspath(self.build_dir)
+        try:
+            os.mkdir(self._build_dir)
+        except FileExistsError as e:
+            pass
+
+        try:
+            os.chdir(self._build_dir)
+            cmd_str = self.get_cmd1_cmake()
+            print(f'call("{cmd_str})"')
+            retcode = call(cmd_str, shell=True)
+            if retcode != 0:
+                raise RuntimeError(f'CMake command "{cmd_str}" was terminated by signal {retcode}')
+            cmd_str = self.get_cmd2_build()
+            print(f'call("{cmd_str}")')
+            retcode = call(cmd_str, shell=True)
+            if retcode != 0:
+                raise RuntimeError(f'Build command "{cmd_str}" was terminated by signal {retcode}')
+            cmd_str = self.get_cmd3_install()
+            print(f'call("{cmd_str}")')
+            retcode = call(cmd_str, shell=True)
+            if retcode != 0:
+                raise RuntimeError(f'Install command "{cmd_str}" was terminated by signal {retcode}')
+        except OSError as e:
+            print("Execution failed:", e, file=sys.stderr)
+        except Exception as e:
+            print("Execution failed:", e, file=sys.stderr)
+            exit(1)

pyextra/acados_template/c_templates_tera/CMakeLists.in.txt

@@ -0,0 +1,374 @@
+#
+# Copyright 2019 Gianluca Frison, Dimitris Kouzoupis, Robin Verschueren,
+# Andrea Zanelli, Niels van Duijkeren, Jonathan Frey, Tommaso Sartor,
+# Branimir Novoselnik, Rien Quirynen, Rezart Qelibari, Dang Doan,
+# Jonas Koenemann, Yutao Chen, Tobias Schöls, Jonas Schlagenhauf, Moritz Diehl
+#
+# This file is part of acados.
+#
+# The 2-Clause BSD License
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+# 1. Redistributions of source code must retain the above copyright notice,
+# this list of conditions and the following disclaimer.
+#
+# 2. Redistributions in binary form must reproduce the above copyright notice,
+# this list of conditions and the following disclaimer in the documentation
+# and/or other materials provided with the distribution.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+# POSSIBILITY OF SUCH DAMAGE.;
+#
+
+{%- if solver_options.qp_solver %}
+	{%- set qp_solver = solver_options.qp_solver %}
+{%- else %}
+	{%- set qp_solver = "FULL_CONDENSING_HPIPM" %}
+{%- endif %}
+
+{%- if solver_options.hessian_approx %}
+	{%- set hessian_approx = solver_options.hessian_approx %}
+{%- elif solver_options.sens_hess %}
+	{%- set hessian_approx = "EXACT" %}
+{%- else %}
+	{%- set hessian_approx = "GAUSS_NEWTON" %}
+{%- endif %}
+
+{%- if constraints.constr_type %}
+	{%- set constr_type = constraints.constr_type %}
+{%- else %}
+	{%- set constr_type = "NONE" %}
+{%- endif %}
+
+{%- if constraints.constr_type_e %}
+	{%- set constr_type_e = constraints.constr_type_e %}
+{%- else %}
+	{%- set constr_type_e = "NONE" %}
+{%- endif %}
+
+{%- if cost.cost_type %}
+	{%- set cost_type = cost.cost_type %}
+{%- else %}
+	{%- set cost_type = "NONE" %}
+{%- endif %}
+
+{%- if cost.cost_type_e %}
+	{%- set cost_type_e = cost.cost_type_e %}
+{%- else %}
+	{%- set cost_type_e = "NONE" %}
+{%- endif %}
+
+{%- if cost.cost_type_0 %}
+	{%- set cost_type_0 = cost.cost_type_0 %}
+{%- else %}
+	{%- set cost_type_0 = "NONE" %}
+{%- endif %}
+
+{%- if dims.nh %}
+	{%- set dims_nh = dims.nh %}
+{%- else %}
+	{%- set dims_nh = 0 %}
+{%- endif %}
+
+{%- if dims.nphi %}
+	{%- set dims_nphi = dims.nphi %}
+{%- else %}
+	{%- set dims_nphi = 0 %}
+{%- endif %}
+
+{%- if dims.nh_e %}
+	{%- set dims_nh_e = dims.nh_e %}
+{%- else %}
+	{%- set dims_nh_e = 0 %}
+{%- endif %}
+
+{%- if dims.nphi_e %}
+	{%- set dims_nphi_e = dims.nphi_e %}
+{%- else %}
+	{%- set dims_nphi_e = 0 %}
+{%- endif %}
+
+{%- if solver_options.model_external_shared_lib_dir %}
+	{%- set model_external_shared_lib_dir = solver_options.model_external_shared_lib_dir %}
+{%- endif %}
+
+{%- if solver_options.model_external_shared_lib_name %}
+	{%- set model_external_shared_lib_name = solver_options.model_external_shared_lib_name %}
+{%- endif %}
+
+{#- control operator #}
+{%- if os and os == "pc" %}
+	{%- set control = "&" %}
+{%- else %}
+	{%- set control = ";" %}
+{%- endif %}
+
+{%- if acados_link_libs and os and os == "pc" %}{# acados linking libraries and flags #}
+	{%- set link_libs = acados_link_libs.qpoases ~ " " ~ acados_link_libs.hpmpc ~ " " ~ acados_link_libs.osqp -%}
+	{%- set openmp_flag = acados_link_libs.openmp %}
+{%- else %}
+	{%- set openmp_flag = " " %}
+	{%- if qp_solver == "FULL_CONDENSING_QPOASES" %}
+		{%- set link_libs = "-lqpOASES_e" %}
+	{%- else %}
+		{%- set link_libs = "" %}
+	{%- endif %}
+{%- endif %}
+
+cmake_minimum_required(VERSION 3.10)
+
+project({{ model.name }})
+
+# build options.
+option(BUILD_ACADOS_SOLVER_LIB "Should the solver library acados_solver_{{ model.name }} be build?" OFF)
+option(BUILD_ACADOS_OCP_SOLVER_LIB "Should the OCP solver library acados_ocp_solver_{{ model.name }} be build?" OFF)
+option(BUILD_EXAMPLE "Should the example main_{{ model.name }} be build?" OFF)
+{%- if solver_options.integrator_type != "DISCRETE" %}
+option(BUILD_SIM_EXAMPLE "Should the simulation example main_sim_{{ model.name }} be build?" OFF)
+option(BUILD_ACADOS_SIM_SOLVER_LIB "Should the simulation solver library acados_sim_solver_{{ model.name }} be build?" OFF)
+{%- endif %}
+
+# object target names
+set(MODEL_OBJ model_{{ model.name }})
+set(OCP_OBJ ocp_{{ model.name }})
+set(SIM_OBJ sim_{{ model.name }})
+
+# model
+set(MODEL_SRC
+{%- if  solver_options.integrator_type == "ERK" %}
+    {{ model.name }}_model/{{ model.name }}_expl_ode_fun.c
+    {{ model.name }}_model/{{ model.name }}_expl_vde_forw.c
+	{%- if hessian_approx == "EXACT" %}
+    {{ model.name }}_model/{{ model.name }}_expl_ode_hess.c
+	{%- endif %}
+{%- elif solver_options.integrator_type == "IRK" %}
+    {{ model.name }}_model/{{ model.name }}_impl_dae_fun.c
+    {{ model.name }}_model/{{ model.name }}_impl_dae_fun_jac_x_xdot_z.c
+    {{ model.name }}_model/{{ model.name }}_impl_dae_jac_x_xdot_u_z.c
+	{%- if hessian_approx == "EXACT" %}
+    {{ model.name }}_model/{{ model.name }}_impl_dae_hess.c
+	{%- endif %}
+{%- elif solver_options.integrator_type == "LIFTED_IRK" %}
+    {{ model.name }}_model/{{ model.name }}_impl_dae_fun.c
+    {{ model.name }}_model/{{ model.name }}_impl_dae_fun_jac_x_xdot_u.c
+	{%- if hessian_approx == "EXACT" %}
+    {{ model.name }}_model/{{ model.name }}_impl_dae_hess.c
+	{%- endif %}
+{%- elif solver_options.integrator_type == "GNSF" %}
+	{% if model.gnsf.purely_linear != 1 %}
+    {{ model.name }}_model/{{ model.name }}_gnsf_phi_fun.c
+    {{ model.name }}_model/{{ model.name }}_gnsf_phi_fun_jac_y.c
+    {{ model.name }}_model/{{ model.name }}_gnsf_phi_jac_y_uhat.c
+		{% if model.gnsf.nontrivial_f_LO == 1 %}
+    {{ model.name }}_model/{{ model.name }}_gnsf_f_lo_fun_jac_x1k1uz.c
+		{%- endif %}
+	{%- endif %}
+    {{ model.name }}_model/{{ model.name }}_gnsf_get_matrices_fun.c
+{%- elif solver_options.integrator_type == "DISCRETE" %}
+	{%- if model.dyn_ext_fun_type == "casadi" %}
+    {{ model.name }}_model/{{ model.name }}_dyn_disc_phi_fun.c
+    {{ model.name }}_model/{{ model.name }}_dyn_disc_phi_fun_jac.c
+		{%- if hessian_approx == "EXACT" %}
+    {{ model.name }}_model/{{ model.name }}_dyn_disc_phi_fun_jac_hess.c
+		{%- endif %}
+	{%- else %}
+    {{ model.name }}_model/{{ model.dyn_source_discrete }}
+	{%- endif %}
+{%- endif -%}
+)
+add_library(${MODEL_OBJ} OBJECT ${MODEL_SRC} )
+
+# optimal control problem - mostly CasADi exports
+if(${BUILD_ACADOS_SOLVER_LIB} OR ${BUILD_ACADOS_OCP_SOLVER_LIB} OR ${BUILD_EXAMPLE})
+    set(OCP_SRC
+{%- if constr_type == "BGP" and dims_nphi > 0 %}
+        {{ model.name }}_constraints/{{ model.name }}_phi_constraint.c
+{%- endif %}
+{%- if constr_type_e == "BGP" and dims_nphi_e > 0 %}
+        {{ model.name }}_constraints/{{ model.name }}_phi_e_constraint.c
+{%- endif %}
+
+{%- if constr_type == "BGH" and dims_nh > 0 %}
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_fun_jac_uxt_zt.c
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_fun.c
+    {%- if hessian_approx == "EXACT" %}
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_fun_jac_uxt_zt_hess.c
+    {%- endif %}
+{%- endif %}
+
+{%- if constr_type_e == "BGH" and dims_nh_e > 0 %}
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_e_fun_jac_uxt_zt.c
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_e_fun.c
+    {%- if hessian_approx == "EXACT" %}
+        {{ model.name }}_constraints/{{ model.name }}_constr_h_e_fun_jac_uxt_zt_hess.c
+    {%- endif %}
+{%- endif %}
+
+{%- if cost_type_0 == "NONLINEAR_LS" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_y_0_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_0_fun_jac_ut_xt.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_0_hess.c
+{%- elif cost_type_0 == "EXTERNAL" %}
+    {%- if cost.cost_ext_fun_type_0 == "casadi" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_0_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_0_fun_jac.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_0_fun_jac_hess.c
+    {%- else %}
+        {{ model.name }}_cost/{{ cost.cost_source_ext_cost_0 }}
+    {%- endif %}
+{%- endif %}
+{%- if cost_type == "NONLINEAR_LS" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_y_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_fun_jac_ut_xt.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_hess.c
+{%- elif cost_type == "EXTERNAL" %}
+    {%- if cost.cost_ext_fun_type == "casadi" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_fun_jac.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_fun_jac_hess.c
+    {%- elif cost.cost_source_ext_cost != cost.cost_source_ext_cost_0 %}
+        {{ model.name }}_cost/{{ cost.cost_source_ext_cost }}
+    {%- endif %}
+{%- endif %}
+{%- if cost_type_e == "NONLINEAR_LS" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_y_e_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_e_fun_jac_ut_xt.c
+        {{ model.name }}_cost/{{ model.name }}_cost_y_e_hess.c
+{%- elif cost_type_e == "EXTERNAL" %}
+    {%- if cost.cost_ext_fun_type_e == "casadi" %}
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_e_fun.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_e_fun_jac.c
+        {{ model.name }}_cost/{{ model.name }}_cost_ext_cost_e_fun_jac_hess.c
+    {%- elif cost.cost_source_ext_cost_e != cost.cost_source_ext_cost_0 %}
+        {{ model.name }}_cost/{{ cost.cost_source_ext_cost_e }}
+    {%- endif %}
+{%- endif %}
+        acados_solver_{{ model.name }}.c)
+    add_library(${OCP_OBJ} OBJECT ${OCP_SRC})
+endif()
+
+{%- if solver_options.integrator_type != "DISCRETE" %}
+# for sim solver
+if(${BUILD_ACADOS_SOLVER_LIB} OR ${BUILD_EXAMPLE}
+    {%- if solver_options.integrator_type != "DISCRETE" %}
+    OR ${BUILD_SIM_EXAMPLE} OR ${BUILD_ACADOS_SIM_SOLVER_LIB}
+    {%- endif -%}
+    )
+    set(SIM_SRC acados_sim_solver_{{ model.name }}.c)
+    add_library(${SIM_OBJ} OBJECT ${SIM_SRC})
+endif()
+{%- endif %}
+
+# for target example
+set(EX_SRC main_{{ model.name }}.c)
+set(EX_EXE main_{{ model.name }})
+
+{%- if model_external_shared_lib_dir and model_external_shared_lib_name %}
+set(EXTERNAL_DIR {{ model_external_shared_lib_dir  }})
+set(EXTERNAL_LIB {{ model_external_shared_lib_name }})
+{%- else %}
+set(EXTERNAL_DIR)
+set(EXTERNAL_LIB)
+{%- endif %}
+
+# set some search paths for preprocessor and linker
+set(ACADOS_INCLUDE_PATH {{ acados_include_path }} CACHE PATH "Define the path which contains the include directory for acados.")
+set(ACADOS_LIB_PATH {{ acados_lib_path }} CACHE PATH "Define the path which contains the lib directory for acados.")
+
+# c-compiler flags for debugging
+set(CMAKE_C_FLAGS_DEBUG "-O0 -ggdb")
+
+set(CMAKE_C_FLAGS "
+{%- if qp_solver == "FULL_CONDENSING_QPOASES" -%}
+   -DACADOS_WITH_QPOASES
+{%- endif -%}
+{%- if qp_solver == "PARTIAL_CONDENSING_OSQP" -%}
+   -DACADOS_WITH_OSQP
+{%- endif -%}
+{%- if qp_solver == "PARTIAL_CONDENSING_QPDUNES" -%}
+   -DACADOS_WITH_QPDUNES
+{%- endif -%}
+    -fPIC -std=c99 {{ openmp_flag }}")
+#-fno-diagnostics-show-line-numbers -g
+
+include_directories(
+   ${ACADOS_INCLUDE_PATH}
+   ${ACADOS_INCLUDE_PATH}/acados
+   ${ACADOS_INCLUDE_PATH}/blasfeo/include
+   ${ACADOS_INCLUDE_PATH}/hpipm/include
+{%- if qp_solver == "FULL_CONDENSING_QPOASES" %}
+   ${ACADOS_INCLUDE_PATH}/qpOASES_e/
+{%- endif %}
+)
+
+# linker flags
+link_directories(${ACADOS_LIB_PATH})
+
+# link to libraries
+if(UNIX)
+    link_libraries(acados hpipm blasfeo m {{ link_libs }})
+else()
+    link_libraries(acados hpipm blasfeo {{ link_libs }})
+endif()
+
+# the targets
+
+# bundled_shared_lib
+if(${BUILD_ACADOS_SOLVER_LIB})
+    set(LIB_ACADOS_SOLVER acados_solver_{{ model.name }})
+    add_library(${LIB_ACADOS_SOLVER} SHARED $<TARGET_OBJECTS:${MODEL_OBJ}> $<TARGET_OBJECTS:${OCP_OBJ}>
+    {%- if solver_options.integrator_type != "DISCRETE" %}
+        $<TARGET_OBJECTS:${SIM_OBJ}>
+    {%- endif -%}
+    )
+    install(TARGETS ${LIB_ACADOS_SOLVER} DESTINATION ${CMAKE_INSTALL_PREFIX})
+endif(${BUILD_ACADOS_SOLVER_LIB})
+
+# ocp_shared_lib
+if(${BUILD_ACADOS_OCP_SOLVER_LIB})
+    set(LIB_ACADOS_OCP_SOLVER acados_ocp_solver_{{ model.name }})
+    add_library(${LIB_ACADOS_OCP_SOLVER} SHARED $<TARGET_OBJECTS:${MODEL_OBJ}> $<TARGET_OBJECTS:${OCP_OBJ}>)
+    # Specify libraries or flags to use when linking a given target and/or its dependents.
+    target_link_libraries(${LIB_ACADOS_OCP_SOLVER} PRIVATE ${EXTERNAL_LIB})
+    target_link_directories(${LIB_ACADOS_OCP_SOLVER} PRIVATE ${EXTERNAL_DIR})
+    install(TARGETS ${LIB_ACADOS_OCP_SOLVER} DESTINATION ${CMAKE_INSTALL_PREFIX})
+endif(${BUILD_ACADOS_OCP_SOLVER_LIB})
+
+# example
+if(${BUILD_EXAMPLE})
+    add_executable(${EX_EXE} ${EX_SRC} $<TARGET_OBJECTS:${MODEL_OBJ}> $<TARGET_OBJECTS:${OCP_OBJ}>
+    {%- if solver_options.integrator_type != "DISCRETE" %}
+        $<TARGET_OBJECTS:${SIM_OBJ}>
+    {%- endif -%}
+    )
+    install(TARGETS ${EX_EXE} DESTINATION ${CMAKE_INSTALL_PREFIX})
+endif(${BUILD_EXAMPLE})
+
+{% if solver_options.integrator_type != "DISCRETE" -%}
+# example_sim
+if(${BUILD_SIM_EXAMPLE})
+    set(EX_SIM_SRC main_sim_{{ model.name }}.c)
+    set(EX_SIM_EXE main_sim_{{ model.name }})
+    add_executable(${EX_SIM_EXE} ${EX_SIM_SRC} $<TARGET_OBJECTS:${MODEL_OBJ}> $<TARGET_OBJECTS:${SIM_OBJ}>)
+    install(TARGETS ${EX_SIM_EXE} DESTINATION ${CMAKE_INSTALL_PREFIX})
+endif(${BUILD_SIM_EXAMPLE})
+
+# sim_shared_lib
+if(${BUILD_ACADOS_SIM_SOLVER_LIB})
+    set(LIB_ACADOS_SIM_SOLVER acados_sim_solver_{{ model.name }})
+    add_library(${LIB_ACADOS_SIM_SOLVER} SHARED $<TARGET_OBJECTS:${MODEL_OBJ}> $<TARGET_OBJECTS:${SIM_OBJ}>)
+    install(TARGETS ${LIB_ACADOS_SIM_SOLVER} DESTINATION ${CMAKE_INSTALL_PREFIX})
+endif(${BUILD_ACADOS_SIM_SOLVER_LIB})
+{%- endif %}
+

pyextra/acados_template/c_templates_tera/acados_sim_solver.in.h

@@ -80,21 +80,21 @@ typedef struct sim_solver_capsule
 } sim_solver_capsule;
 
 
-int {{ model.name }}_acados_sim_create(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_sim_create(sim_solver_capsule *capsule);
-int {{ model.name }}_acados_sim_solve(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_sim_solve(sim_solver_capsule *capsule);
-int {{ model.name }}_acados_sim_free(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_sim_free(sim_solver_capsule *capsule);
-int {{ model.name }}_acados_sim_update_params(sim_solver_capsule *capsule, double *value, int np);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_sim_update_params(sim_solver_capsule *capsule, double *value, int np);
 
-sim_config * {{ model.name }}_acados_get_sim_config(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT sim_config * {{ model.name }}_acados_get_sim_config(sim_solver_capsule *capsule);
-sim_in * {{ model.name }}_acados_get_sim_in(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT sim_in * {{ model.name }}_acados_get_sim_in(sim_solver_capsule *capsule);
-sim_out * {{ model.name }}_acados_get_sim_out(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT sim_out * {{ model.name }}_acados_get_sim_out(sim_solver_capsule *capsule);
-void * {{ model.name }}_acados_get_sim_dims(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT void * {{ model.name }}_acados_get_sim_dims(sim_solver_capsule *capsule);
-sim_opts * {{ model.name }}_acados_get_sim_opts(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT sim_opts * {{ model.name }}_acados_get_sim_opts(sim_solver_capsule *capsule);
-sim_solver * {{ model.name }}_acados_get_sim_solver(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT sim_solver * {{ model.name }}_acados_get_sim_solver(sim_solver_capsule *capsule);
 
 
-sim_solver_capsule * {{ model.name }}_acados_sim_solver_create_capsule(void);
+ACADOS_SYMBOL_EXPORT sim_solver_capsule * {{ model.name }}_acados_sim_solver_create_capsule(void);
-int {{ model.name }}_acados_sim_solver_free_capsule(sim_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_sim_solver_free_capsule(sim_solver_capsule *capsule);
 
 #ifdef __cplusplus
 }

pyextra/acados_template/c_templates_tera/acados_solver.in.c

@@ -1862,6 +1862,11 @@ void {{ model.name }}_acados_create_6_set_opts({{ model.name }}_solver_capsule*
 {%- endif %}
 
 
+{%- if solver_options.qp_solver is containing("HPIPM") %}
+    // set HPIPM mode: should be done before setting other QP solver options
+    ocp_nlp_solver_opts_set(nlp_config, nlp_opts, "qp_hpipm_mode", "{{ solver_options.hpipm_mode }}");
+{%- endif %}
+
 {% if solver_options.nlp_solver_type == "SQP" %}
     // set SQP specific options
     double nlp_solver_tol_stat = {{ solver_options.nlp_solver_tol_stat }};
@@ -2071,6 +2076,63 @@ int {{ model.name }}_acados_update_qp_solver_cond_N({{ model.name }}_solver_caps
 }
 
 
+int {{ model.name }}_acados_reset({{ model.name }}_solver_capsule* capsule)
+{
+
+    // set initialization to all zeros
+{# TODO: use guess values / initial state value from json instead?! #}
+    const int N = capsule->nlp_solver_plan->N;
+    ocp_nlp_config* nlp_config = capsule->nlp_config;
+    ocp_nlp_dims* nlp_dims = capsule->nlp_dims;
+    ocp_nlp_out* nlp_out = capsule->nlp_out;
+    ocp_nlp_in* nlp_in = capsule->nlp_in;
+    ocp_nlp_solver* nlp_solver = capsule->nlp_solver;
+
+    int nx, nu, nv, ns, nz, ni, dim;
+
+    double* buffer = calloc(NX+NU+NZ+2*NS+2*NSN+NBX+NBU+NG+NH+NPHI+NBX0+NBXN+NHN+NPHIN+NGN, sizeof(double));
+
+    for(int i=0; i<N+1; i++)
+    {
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "x", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "u", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "sl", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "su", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "lam", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "t", buffer);
+        ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "z", buffer);
+        if (i<N)
+        {
+            ocp_nlp_out_set(nlp_config, nlp_dims, nlp_out, i, "pi", buffer);
+        {%- if solver_options.integrator_type == "IRK" %}
+            ocp_nlp_set(nlp_config, nlp_solver, i, "xdot_guess", buffer);
+            ocp_nlp_set(nlp_config, nlp_solver, i, "z_guess", buffer);
+        {% elif solver_options.integrator_type == "LIFTED_IRK" %}
+            ocp_nlp_set(nlp_config, nlp_solver, i, "xdot_guess", buffer);
+        {% elif solver_options.integrator_type == "GNSF" %}
+            ocp_nlp_set(nlp_config, nlp_solver, i, "gnsf_phi_guess", buffer);
+        {%- endif %}
+        }
+    }
+
+{%- if solver_options.qp_solver == 'PARTIAL_CONDENSING_HPIPM' %}
+    // get qp_status: if NaN -> reset memory
+    int qp_status;
+    ocp_nlp_get(capsule->nlp_config, capsule->nlp_solver, "qp_status", &qp_status);
+    if (qp_status == 3)
+    {
+        // printf("\nin reset qp_status %d -> resetting QP memory\n", qp_status);
+        ocp_nlp_solver_reset_qp_memory(nlp_solver, nlp_in, nlp_out);
+    }
+{%- endif %}
+
+    free(buffer);
+    return 0;
+}
+
+
+
+
 int {{ model.name }}_acados_update_params({{ model.name }}_solver_capsule* capsule, int stage, double *p, int np)
 {
     int solver_status = 0;
@@ -2096,7 +2158,7 @@ int {{ model.name }}_acados_update_params({{ model.name }}_solver_capsule* capsu
         {%- endif %}
     {% elif solver_options.integrator_type == "LIFTED_IRK" %}
         capsule->impl_dae_fun[stage].set_param(capsule->impl_dae_fun+stage, p);
-        capsule->impl_dae_fun_jac_x_xdot_z[stage].set_param(capsule->impl_dae_fun_jac_x_xdot_z+stage, p);
+        capsule->impl_dae_fun_jac_x_xdot_u[stage].set_param(capsule->impl_dae_fun_jac_x_xdot_u+stage, p);
     {% elif solver_options.integrator_type == "ERK" %}
         capsule->forw_vde_casadi[stage].set_param(capsule->forw_vde_casadi+stage, p);
         capsule->expl_ode_fun[stage].set_param(capsule->expl_ode_fun+stage, p);

pyextra/acados_template/c_templates_tera/acados_solver.in.h

@@ -34,6 +34,8 @@
 #ifndef ACADOS_SOLVER_{{ model.name }}_H_
 #define ACADOS_SOLVER_{{ model.name }}_H_
 
+#include "acados/utils/types.h"
+
 #include "acados_c/ocp_nlp_interface.h"
 #include "acados_c/external_function_interface.h"
 
@@ -172,38 +174,41 @@ typedef struct {{ model.name }}_solver_capsule
 
 } {{ model.name }}_solver_capsule;
 
-{{ model.name }}_solver_capsule * {{ model.name }}_acados_create_capsule(void);
+ACADOS_SYMBOL_EXPORT {{ model.name }}_solver_capsule * {{ model.name }}_acados_create_capsule(void);
-int {{ model.name }}_acados_free_capsule({{ model.name }}_solver_capsule *capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_free_capsule({{ model.name }}_solver_capsule *capsule);
+
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_create({{ model.name }}_solver_capsule * capsule);
+
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_reset({{ model.name }}_solver_capsule* capsule);
 
-int {{ model.name }}_acados_create({{ model.name }}_solver_capsule * capsule);
 /**
  * Generic version of {{ model.name }}_acados_create which allows to use a different number of shooting intervals than
  * the number used for code generation. If new_time_steps=NULL and n_time_steps matches the number used for code
  * generation, the time-steps from code generation is used.
  */
-int {{ model.name }}_acados_create_with_discretization({{ model.name }}_solver_capsule * capsule, int n_time_steps, double* new_time_steps);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_create_with_discretization({{ model.name }}_solver_capsule * capsule, int n_time_steps, double* new_time_steps);
 /**
  * Update the time step vector. Number N must be identical to the currently set number of shooting nodes in the
  * nlp_solver_plan. Returns 0 if no error occurred and a otherwise a value other than 0.
  */
-int {{ model.name }}_acados_update_time_steps({{ model.name }}_solver_capsule * capsule, int N, double* new_time_steps);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_update_time_steps({{ model.name }}_solver_capsule * capsule, int N, double* new_time_steps);
 /**
  * This function is used for updating an already initialized solver with a different number of qp_cond_N.
  */
-int {{ model.name }}_acados_update_qp_solver_cond_N({{ model.name }}_solver_capsule * capsule, int qp_solver_cond_N);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_update_qp_solver_cond_N({{ model.name }}_solver_capsule * capsule, int qp_solver_cond_N);
-int {{ model.name }}_acados_update_params({{ model.name }}_solver_capsule * capsule, int stage, double *value, int np);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_update_params({{ model.name }}_solver_capsule * capsule, int stage, double *value, int np);
-int {{ model.name }}_acados_solve({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_solve({{ model.name }}_solver_capsule * capsule);
-int {{ model.name }}_acados_free({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT int {{ model.name }}_acados_free({{ model.name }}_solver_capsule * capsule);
-void {{ model.name }}_acados_print_stats({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT void {{ model.name }}_acados_print_stats({{ model.name }}_solver_capsule * capsule);
-
+                     
-ocp_nlp_in *{{ model.name }}_acados_get_nlp_in({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_in *{{ model.name }}_acados_get_nlp_in({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_out *{{ model.name }}_acados_get_nlp_out({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_out *{{ model.name }}_acados_get_nlp_out({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_out *{{ model.name }}_acados_get_sens_out({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_out *{{ model.name }}_acados_get_sens_out({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_solver *{{ model.name }}_acados_get_nlp_solver({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_solver *{{ model.name }}_acados_get_nlp_solver({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_config *{{ model.name }}_acados_get_nlp_config({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_config *{{ model.name }}_acados_get_nlp_config({{ model.name }}_solver_capsule * capsule);
-void *{{ model.name }}_acados_get_nlp_opts({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT void *{{ model.name }}_acados_get_nlp_opts({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_dims *{{ model.name }}_acados_get_nlp_dims({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_dims *{{ model.name }}_acados_get_nlp_dims({{ model.name }}_solver_capsule * capsule);
-ocp_nlp_plan_t *{{ model.name }}_acados_get_nlp_plan({{ model.name }}_solver_capsule * capsule);
+ACADOS_SYMBOL_EXPORT ocp_nlp_plan_t *{{ model.name }}_acados_get_nlp_plan({{ model.name }}_solver_capsule * capsule);
 
 #ifdef __cplusplus
 } /* extern "C" */

pyextra/acados_template/c_templates_tera/acados_solver.in.pxd

@@ -48,6 +48,7 @@ cdef extern from "acados_solver_{{ model.name }}.h":
 
     int acados_update_params "{{ model.name }}_acados_update_params"(nlp_solver_capsule * capsule, int stage, double *value, int np_)
     int acados_solve "{{ model.name }}_acados_solve"(nlp_solver_capsule * capsule)
+    int acados_reset "{{ model.name }}_acados_reset"(nlp_solver_capsule * capsule)
     int acados_free "{{ model.name }}_acados_free"(nlp_solver_capsule * capsule)
     void acados_print_stats "{{ model.name }}_acados_print_stats"(nlp_solver_capsule * capsule)
 

pyextra/acados_template/c_templates_tera/acados_solver_sfun.in.c

@@ -126,10 +126,14 @@ static void mdlInitializeSizes (SimStruct *S)
   {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W_0 %}  {#- cost_W_0 #}
     {%- set n_inputs = n_inputs + 1 %}
   {%- endif -%}
-  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W -%}  {#- cost_W #}
+  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {#- cost_W #}
     {%- set n_inputs = n_inputs + 1 %}
   {%- endif -%}
-  {%- if dims.ny_e > 0 and simulink_opts.inputs.cost_W_e -%}  {#- cost_W_e #}
+  {%- if dims.ny_e > 0 and simulink_opts.inputs.cost_W_e %}  {#- cost_W_e #}
+    {%- set n_inputs = n_inputs + 1 -%}
+  {%- endif -%}
+
+  {%- if simulink_opts.inputs.reset_solver -%}  {#- reset_solver #}
     {%- set n_inputs = n_inputs + 1 -%}
   {%- endif -%}
 
@@ -256,7 +260,7 @@ static void mdlInitializeSizes (SimStruct *S)
     ssSetInputPortVectorDimension(S, {{ i_input }}, {{ dims.ny_0 * dims.ny_0 }});
   {%- endif %}
 
-  {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W %}  {#- cost_W #}
+  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {#- cost_W #}
     {%- set i_input = i_input + 1 %}
     // cost_W
     ssSetInputPortVectorDimension(S, {{ i_input }}, {{ dims.ny * dims.ny }});
@@ -268,6 +272,12 @@ static void mdlInitializeSizes (SimStruct *S)
     ssSetInputPortVectorDimension(S, {{ i_input }}, {{ dims.ny_e * dims.ny_e }});
   {%- endif %}
 
+  {%- if simulink_opts.inputs.reset_solver -%}  {#- reset_solver #}
+    {%- set i_input = i_input + 1 %}
+    // reset_solver
+    ssSetInputPortVectorDimension(S, {{ i_input }}, 1);
+  {%- endif -%}
+
   {%- if simulink_opts.inputs.x_init -%}  {#- x_init #}
     {%- set i_input = i_input + 1 %}
     // x_init
@@ -406,13 +416,13 @@ static void mdlOutputs(SimStruct *S, int_T tid)
     {%- set buffer_sizes = buffer_sizes | concat(with=(dims.ny_e)) %}
   {%- endif %}
 
-  {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W_0 %}  {# cost_W_0 #}
+  {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W_0 %}  {#- cost_W_0 #}
     {%- set buffer_sizes = buffer_sizes | concat(with=(dims.ny_0 * dims.ny_0)) %}
   {%- endif %}
-  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {# cost_W #}
+  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {#- cost_W #}
     {%- set buffer_sizes = buffer_sizes | concat(with=(dims.ny * dims.ny)) %}
   {%- endif %}
-  {%- if dims.ny_e > 0 and simulink_opts.inputs.cost_W_e %}  {# cost_W_e #}
+  {%- if dims.ny_e > 0 and simulink_opts.inputs.cost_W_e %}  {#- cost_W_e #}
     {%- set buffer_sizes = buffer_sizes | concat(with=(dims.ny_e * dims.ny_e)) %}
   {%- endif %}
 
@@ -602,7 +612,7 @@ static void mdlOutputs(SimStruct *S, int_T tid)
         ocp_nlp_constraints_model_set(nlp_config, nlp_dims, nlp_in, ii, "uh", buffer);
   {%- endif -%}
 
-  {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W_0 %}  {# cost_W_0 #}
+  {%- if dims.ny_0 > 0 and simulink_opts.inputs.cost_W_0 %}  {#- cost_W_0 #}
     // cost_W_0
     {%- set i_input = i_input + 1 %}
     in_sign = ssGetInputPortRealSignalPtrs(S, {{ i_input }});
@@ -612,7 +622,7 @@ static void mdlOutputs(SimStruct *S, int_T tid)
     ocp_nlp_cost_model_set(nlp_config, nlp_dims, nlp_in, 0, "W", buffer);
   {%- endif %}
 
-  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {# cost_W #}
+  {%- if dims.ny > 0 and simulink_opts.inputs.cost_W %}  {#- cost_W #}
     // cost_W
     {%- set i_input = i_input + 1 %}
     in_sign = ssGetInputPortRealSignalPtrs(S, {{ i_input }});
@@ -633,6 +643,17 @@ static void mdlOutputs(SimStruct *S, int_T tid)
     ocp_nlp_cost_model_set(nlp_config, nlp_dims, nlp_in, {{ dims.N }}, "W", buffer);
   {%- endif %}
 
+  {%- if simulink_opts.inputs.reset_solver %}  {#- reset_solver #}
+    // reset_solver
+    {%- set i_input = i_input + 1 %}
+    in_sign = ssGetInputPortRealSignalPtrs(S, {{ i_input }});
+    double reset = (double)(*in_sign[0]);
+    if (reset)
+    {
+        {{ model.name }}_acados_reset(capsule);
+    }
+  {%- endif %}
+
   {%- if simulink_opts.inputs.x_init %}  {#- x_init #}
     // x_init
     {%- set i_input = i_input + 1 %}

pyextra/acados_template/c_templates_tera/make_sfun.in.m

@@ -264,6 +264,11 @@ input_note = strcat(input_note, num2str(i_in), ') cost_W_e in column-major forma
 i_in = i_in + 1;
 {%- endif %}
 
+{%- if simulink_opts.inputs.reset_solver %}  {#- reset_solver #}
+input_note = strcat(input_note, num2str(i_in), ') reset_solver determines if iterate is set to all zeros before other initializations (x_init, u_init) are set and before solver is called, size [1]\n ');
+i_in = i_in + 1;
+{%- endif %}
+
 {%- if simulink_opts.inputs.x_init %}  {#- x_init #}
 input_note = strcat(input_note, num2str(i_in), ') initialization of x for all shooting nodes, size [{{ dims.nx * (dims.N+1) }}]\n ');
 i_in = i_in + 1;

pyextra/acados_template/generate_c_code_discrete_dynamics.py

@@ -32,12 +32,12 @@
 #
 
 import os
-from casadi import *
+import casadi as ca
 from .utils import ALLOWED_CASADI_VERSIONS, casadi_length, casadi_version_warning
 
 def generate_c_code_discrete_dynamics( model, opts ):
 
-    casadi_version = CasadiMeta.version()
+    casadi_version = ca.CasadiMeta.version()
     casadi_opts = dict(mex=False, casadi_int='int', casadi_real='double')
 
     if casadi_version not in (ALLOWED_CASADI_VERSIONS):
@@ -49,13 +49,12 @@ def generate_c_code_discrete_dynamics( model, opts ):
     p = model.p
     phi = model.disc_dyn_expr
     model_name = model.name
-    nx = x.size()[0]
+    nx = casadi_length(x)
 
-
+    if isinstance(phi, ca.MX):
-    if isinstance(phi, casadi.MX):
+        symbol = ca.MX.sym
-        symbol = MX.sym
+    elif isinstance(phi, ca.SX):
-    elif isinstance(phi, casadi.SX):
+        symbol = ca.SX.sym
-        symbol = SX.sym
     else:
         Exception("generate_c_code_disc_dyn: disc_dyn_expr must be a CasADi expression, you have type: {}".format(type(phi)))
 
@@ -63,12 +62,12 @@ def generate_c_code_discrete_dynamics( model, opts ):
     lam = symbol('lam', nx, 1)
 
     # generate jacobians
-    ux = vertcat(u,x)
+    ux = ca.vertcat(u,x)
-    jac_ux = jacobian(phi, ux)
+    jac_ux = ca.jacobian(phi, ux)
     # generate adjoint
-    adj_ux = jtimes(phi, ux, lam, True)
+    adj_ux = ca.jtimes(phi, ux, lam, True)
     # generate hessian
-    hess_ux = jacobian(adj_ux, ux)
+    hess_ux = ca.jacobian(adj_ux, ux)
     
     ## change directory
     code_export_dir = opts["code_export_directory"]
@@ -85,15 +84,15 @@ def generate_c_code_discrete_dynamics( model, opts ):
 
     # set up & generate Functions
     fun_name = model_name + '_dyn_disc_phi_fun'
-    phi_fun = Function(fun_name, [x, u, p], [phi])
+    phi_fun = ca.Function(fun_name, [x, u, p], [phi])
     phi_fun.generate(fun_name, casadi_opts)
 
     fun_name = model_name + '_dyn_disc_phi_fun_jac'
-    phi_fun_jac_ut_xt = Function(fun_name, [x, u, p], [phi, jac_ux.T])
+    phi_fun_jac_ut_xt = ca.Function(fun_name, [x, u, p], [phi, jac_ux.T])
     phi_fun_jac_ut_xt.generate(fun_name, casadi_opts)
 
     fun_name = model_name + '_dyn_disc_phi_fun_jac_hess'
-    phi_fun_jac_ut_xt_hess = Function(fun_name, [x, u, lam, p], [phi, jac_ux.T, hess_ux])
+    phi_fun_jac_ut_xt_hess = ca.Function(fun_name, [x, u, lam, p], [phi, jac_ux.T, hess_ux])
     phi_fun_jac_ut_xt_hess.generate(fun_name, casadi_opts)
     
     os.chdir(cwd)

pyextra/acados_template/simulink_default_opts.json

@@ -32,6 +32,7 @@
         "cost_W_0": 0,
         "cost_W": 0,
         "cost_W_e": 0,
+        "reset_solver": 0,
         "x_init": 0,
         "u_init": 0
     },

pyextra/acados_template/utils.py

@@ -1,3 +1,4 @@
+# -*- coding: future_fstrings -*-
 #
 # Copyright 2019 Gianluca Frison, Dimitris Kouzoupis, Robin Verschueren,
 # Andrea Zanelli, Niels van Duijkeren, Jonathan Frey, Tommaso Sartor,
@@ -49,7 +50,7 @@ def get_acados_path():
         ACADOS_PATH = os.path.realpath(acados_path)
         msg = 'Warning: Did not find environment variable ACADOS_SOURCE_DIR, '
         msg += 'guessed ACADOS_PATH to be {}.\n'.format(ACADOS_PATH)
-        msg += 'Please export ACADOS_SOURCE_DIR to not avoid this warning.'
+        msg += 'Please export ACADOS_SOURCE_DIR to avoid this warning.'
         print(msg)
     return ACADOS_PATH
 
@@ -74,7 +75,7 @@ def get_tera_exec_path():
 platform2tera = {
     "linux": "linux",
     "darwin": "osx",
-    "win32": "window.exe"
+    "win32": "windows"
 }
 
 
@@ -212,16 +213,14 @@ def render_template(in_file, out_file, template_dir, json_path):
     template_glob = os.path.join(acados_path, 'c_templates_tera', '*')
 
     # call tera as system cmd
-    os_cmd = "{tera_path} '{template_glob}' '{in_file}' '{json_path}' '{out_file}'".format(
+    os_cmd = f"{tera_path} '{template_glob}' '{in_file}' '{json_path}' '{out_file}'"
-        tera_path=tera_path,
+    # Windows cmd.exe can not cope with '...', so use "..." instead:
-        template_glob=template_glob,
+    if os.name == 'nt':
-        json_path=json_path,
+        os_cmd = os_cmd.replace('\'', '\"')
-        in_file=in_file,
+
-        out_file=out_file
-    )
     status = os.system(os_cmd)
     if (status != 0):
-        raise Exception('Rendering of {} failed!\n\nAttempted to execute OS command:\n{}\n\nExiting.\n'.format(in_file, os_cmd))
+        raise Exception(f'Rendering of {in_file} failed!\n\nAttempted to execute OS command:\n{os_cmd}\n\nExiting.\n')
 
     os.chdir(cwd)
 
@@ -235,9 +234,7 @@ def np_array_to_list(np_array):
     elif isinstance(np_array, (DM)):
         return np_array.full()
     else:
-        raise(Exception(
+        raise(Exception(f"Cannot convert to list type {type(np_array)}"))
-            "Cannot convert to list type {}".format(type(np_array))
-        ))
 
 
 def format_class_dict(d):

selfdrive/controls/lib/lateral_mpc_lib/SConscript

@@ -44,11 +44,14 @@ generated_files = [
 ] + build_files
 
 acados_dir = '#third_party/acados'
+acados_templates_dir = '#pyextra/acados_template/c_templates_tera'
+
 source_list = ['lat_mpc.py',
   f'{acados_dir}/include/acados_c/ocp_nlp_interface.h',
   f'{acados_dir}/aarch64/lib/libacados.so',
   f'{acados_dir}/x86_64/lib/libacados.so',
   f'{acados_dir}/larch64/lib/libacados.so',
+  f'{acados_templates_dir}/acados_solver.in.c',
 ]
 
 lenv = env.Clone()

selfdrive/controls/lib/longitudinal_mpc_lib/SConscript

@@ -51,11 +51,14 @@ generated_files = [
 ] + build_files
 
 acados_dir = '#third_party/acados'
+acados_templates_dir = '#pyextra/acados_template/c_templates_tera'
+
 source_list = ['long_mpc.py',
   f'{acados_dir}/include/acados_c/ocp_nlp_interface.h',
   f'{acados_dir}/aarch64/lib/libacados.so',
   f'{acados_dir}/x86_64/lib/libacados.so',
   f'{acados_dir}/larch64/lib/libacados.so',
+  f'{acados_templates_dir}/acados_solver.in.c',
 ]
 
 lenv = env.Clone()

selfdrive/controls/lib/longitudinal_mpc_lib/long_mpc.py

@@ -192,11 +192,14 @@ def gen_long_ocp():
 class LongitudinalMpc:
   def __init__(self, e2e=False):
     self.e2e = e2e
+    self.solver = AcadosOcpSolverCython(MODEL_NAME, ACADOS_SOLVER_TYPE, N)
     self.reset()
     self.source = SOURCES[2]
 
   def reset(self):
-    self.solver = AcadosOcpSolverCython(MODEL_NAME, ACADOS_SOLVER_TYPE, N)
+    # self.solver = AcadosOcpSolverCython(MODEL_NAME, ACADOS_SOLVER_TYPE, N)
+    self.solver.reset()
+    # self.solver.options_set('print_level', 2)
     self.v_solution = np.zeros(N+1)
     self.a_solution = np.zeros(N+1)
     self.prev_a = np.array(self.a_solution)
@@ -352,6 +355,8 @@ class LongitudinalMpc:
     self.run()
 
   def run(self):
+    # t0 = sec_since_boot()
+    # reset = 0
     for i in range(N+1):
       self.solver.set(i, 'p', self.params[i])
     self.solver.constraints_set(0, "lbx", self.x0)
@@ -386,6 +391,8 @@ class LongitudinalMpc:
         self.last_cloudlog_t = t
         cloudlog.warning(f"Long mpc reset, solution_status: {self.solution_status}")
       self.reset()
+      # reset = 1
+    # print(f"long_mpc timings: total internal {self.solve_time:.2e}, external: {(sec_since_boot() - t0):.2e} qp {self.time_qp_solution:.2e}, lin {self.time_linearization:.2e} qp_iter {qp_iter}, reset {reset}")
 
 
 if __name__ == "__main__":

third_party/acados/build.sh

@@ -17,8 +17,8 @@ if [ ! -d acados_repo/ ]; then
   # git clone https://github.com/commaai/acados.git $DIR/acados_repo
 fi
 cd acados_repo
-git fetch
+git fetch --all
-git checkout 105e06df87f06ea02df4af825867c946b31defdd
+git checkout 8ea8827fafb1b23b4c7da1c4cf650de1cbd73584
 git submodule update --recursive --init
 
 # build

third_party/acados/include/acados/dense_qp/dense_qp_qpoases.h

@@ -118,6 +118,8 @@ int dense_qp_qpoases(void *config, dense_qp_in *qp_in, dense_qp_out *qp_out, voi
 //
 void dense_qp_qpoases_eval_sens(void *config_, void *qp_in, void *qp_out, void *opts_, void *mem_, void *work_);
 //
+void dense_qp_qpoases_memory_reset(void *config_, void *qp_in, void *qp_out, void *opts_, void *mem_, void *work_);
+//
 void dense_qp_qpoases_config_initialize_default(void *config_);
 
 #ifdef __cplusplus

third_party/acados/include/acados/ocp_nlp/ocp_nlp_common.h

@@ -93,6 +93,7 @@ typedef struct ocp_nlp_config
                             char *field, int stage, int index, void *sens_nlp_out);
     // prepare memory
     int (*precompute)(void *config, void *dims, void *nlp_in, void *nlp_out, void *opts_, void *mem, void *work);
+    void (*memory_reset_qp_solver)(void *config, void *dims, void *nlp_in, void *nlp_out, void *opts_, void *mem, void *work);
     // initalize this struct with default values
     void (*config_initialize_default)(void *config);
     // general getter
@@ -306,6 +307,7 @@ typedef struct ocp_nlp_res
     struct blasfeo_dvec *res_eq;  // dynamics
     struct blasfeo_dvec *res_ineq;  // inequality constraints
     struct blasfeo_dvec *res_comp;  // complementarity
+    struct blasfeo_dvec tmp;  // tmp
     double inf_norm_res_stat;
     double inf_norm_res_eq;
     double inf_norm_res_ineq;
@@ -402,7 +404,7 @@ ocp_nlp_workspace *ocp_nlp_workspace_assign(ocp_nlp_config *config, ocp_nlp_dims
 void ocp_nlp_alias_memory_to_submodules(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
             ocp_nlp_out *out, ocp_nlp_opts *opts, ocp_nlp_memory *mem, ocp_nlp_workspace *work);
 //
-void ocp_nlp_initialize_qp(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
+void ocp_nlp_initialize_submodules(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
             ocp_nlp_out *out, ocp_nlp_opts *opts, ocp_nlp_memory *mem, ocp_nlp_workspace *work);
 //
 void ocp_nlp_approximate_qp_matrices(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,

third_party/acados/include/acados/ocp_nlp/ocp_nlp_sqp.h

@@ -122,7 +122,9 @@ typedef struct
 acados_size_t ocp_nlp_sqp_memory_calculate_size(void *config, void *dims, void *opts_);
 //
 void *ocp_nlp_sqp_memory_assign(void *config, void *dims, void *opts_, void *raw_memory);
-
+//
+void ocp_nlp_sqp_memory_reset_qp_solver(void *config_, void *dims_, void *nlp_in_, void *nlp_out_,
+    void *opts_, void *mem_, void *work_);
 
 
 /************************************************

third_party/acados/include/acados/ocp_qp/ocp_qp_common.h

@@ -72,6 +72,7 @@ typedef struct
     void (*memory_get)(void *config_, void *mem_, const char *field, void* value);
     acados_size_t (*workspace_calculate_size)(void *config, void *dims, void *opts);
     int (*evaluate)(void *config, void *qp_in, void *qp_out, void *opts, void *mem, void *work);
+    void (*memory_reset)(void *config, void *qp_in, void *qp_out, void *opts, void *mem, void *work);
     void (*eval_sens)(void *config, void *qp_in, void *qp_out, void *opts, void *mem, void *work);
 } qp_solver_config;
 #endif

third_party/acados/include/acados/ocp_qp/ocp_qp_hpipm.h

@@ -63,6 +63,7 @@ typedef struct ocp_qp_hpipm_memory_
     struct d_ocp_qp_ipm_ws *hpipm_workspace;
     double time_qp_solver_call;
     int iter;
+    int status;
 
 } ocp_qp_hpipm_memory;
 
@@ -87,6 +88,8 @@ acados_size_t ocp_qp_hpipm_workspace_calculate_size(void *config, void *dims, vo
 //
 int ocp_qp_hpipm(void *config, void *qp_in, void *qp_out, void *opts_, void *mem_, void *work_);
 //
+void ocp_qp_hpipm_memory_reset(void *config_, void *qp_in_, void *qp_out_, void *opts_, void *mem_, void *work_);
+//
 void ocp_qp_hpipm_eval_sens(void *config, void *qp_in, void *qp_out, void *opts_, void *mem_, void *work_);
 //
 void ocp_qp_hpipm_config_initialize_default(void *config);

third_party/acados/include/acados/ocp_qp/ocp_qp_hpmpc.h

@@ -99,6 +99,7 @@ typedef struct ocp_qp_hpmpc_memory_
 
     double time_qp_solver_call;
     int iter;
+    int status;
 
 } ocp_qp_hpmpc_memory;
 

third_party/acados/include/acados/ocp_qp/ocp_qp_ooqp.h

@@ -112,6 +112,7 @@ typedef struct ocp_qp_ooqp_memory_
     int nnz;  // max(nnzQ, nnzA, nnzC)
     double time_qp_solver_call;
     int iter;
+    int status;
 
 } ocp_qp_ooqp_memory;
 

third_party/acados/include/acados/ocp_qp/ocp_qp_osqp.h

@@ -75,6 +75,7 @@ typedef struct ocp_qp_osqp_memory_
 
     double time_qp_solver_call;
     int iter;
+    int status;
 
 } ocp_qp_osqp_memory;
 

third_party/acados/include/acados/ocp_qp/ocp_qp_qpdunes.h

@@ -71,6 +71,7 @@ typedef struct ocp_qp_qpdunes_memory_
     qpData_t qpData;
     double time_qp_solver_call;
     int iter;
+    int status;
 
 } ocp_qp_qpdunes_memory;
 

third_party/acados/include/acados/ocp_qp/ocp_qp_xcond_solver.h

@@ -92,6 +92,7 @@ typedef struct
     acados_size_t (*memory_calculate_size)(void *config, ocp_qp_xcond_solver_dims *dims, void *opts);
     void *(*memory_assign)(void *config, ocp_qp_xcond_solver_dims *dims, void *opts, void *raw_memory);
     void (*memory_get)(void *config_, void *mem_, const char *field, void* value);
+    void (*memory_reset)(void *config, ocp_qp_xcond_solver_dims *dims, ocp_qp_in *qp_in, ocp_qp_out *qp_out, void *opts, void *mem, void *work);
     acados_size_t (*workspace_calculate_size)(void *config, ocp_qp_xcond_solver_dims *dims, void *opts);
     int (*evaluate)(void *config, ocp_qp_xcond_solver_dims *dims, ocp_qp_in *qp_in, ocp_qp_out *qp_out, void *opts, void *mem, void *work);
     void (*eval_sens)(void *config, ocp_qp_xcond_solver_dims *dims, ocp_qp_in *param_qp_in, ocp_qp_out *sens_qp_out, void *opts, void *mem, void *work);

third_party/acados/include/acados/sim/sim_lifted_irk_integrator.h

@@ -96,6 +96,7 @@ typedef struct
     struct blasfeo_dvec *u;         // controls (nu) -- for expansion step
 
     int update_sens;
+    // int init_K;
 
 	double time_sim;
 	double time_ad;

third_party/acados/include/acados/utils/types.h

@@ -35,6 +35,21 @@
 #ifndef ACADOS_UTILS_TYPES_H_
 #define ACADOS_UTILS_TYPES_H_
 
+/* Symbol visibility in DLLs */
+#ifndef ACADOS_SYMBOL_EXPORT
+  #if defined(_WIN32) || defined(__WIN32__) || defined(__CYGWIN__)
+    #if defined(STATIC_LINKED)
+      #define ACADOS_SYMBOL_EXPORT
+    #else
+      #define ACADOS_SYMBOL_EXPORT __declspec(dllexport)
+    #endif
+  #elif defined(__GNUC__) && ((__GNUC__ >= 4) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 4))
+    #define ACADOS_SYMBOL_EXPORT __attribute__ ((visibility ("default")))
+  #else
+    #define ACADOS_SYMBOL_EXPORT
+  #endif
+#endif
+
 #ifdef __cplusplus
 extern "C" {
 #endif

third_party/acados/include/acados_c/ocp_nlp_interface.h

@@ -46,6 +46,7 @@ extern "C" {
 #include "acados/sim/sim_irk_integrator.h"
 #include "acados/sim/sim_lifted_irk_integrator.h"
 #include "acados/sim/sim_gnsf.h"
+#include "acados/utils/types.h"
 // acados_c
 #include "acados_c/ocp_qp_interface.h"
 #include "acados_c/sim_interface.h"
@@ -151,46 +152,46 @@ typedef struct ocp_nlp_solver
 /// default/invalid state.
 ///
 /// \param N Horizon length
-ocp_nlp_plan_t *ocp_nlp_plan_create(int N);
+ACADOS_SYMBOL_EXPORT ocp_nlp_plan_t *ocp_nlp_plan_create(int N);
 
 /// Destructor for plan struct, frees memory.
 ///
 /// \param plan_ The plan struct to destroy.
-void ocp_nlp_plan_destroy(void* plan_);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_plan_destroy(void* plan_);
 
 
 /// Constructs an nlp configuration struct from a plan.
 ///
 /// \param plan The plan (user nlp configuration).
-ocp_nlp_config *ocp_nlp_config_create(ocp_nlp_plan_t plan);
+ACADOS_SYMBOL_EXPORT ocp_nlp_config *ocp_nlp_config_create(ocp_nlp_plan_t plan);
 
 /// Desctructor of the nlp configuration.
 ///
 /// \param config_ The configuration struct.
-void ocp_nlp_config_destroy(void *config_);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_config_destroy(void *config_);
 
 
 /// Constructs an struct that contains the dimensions of the variables.
 ///
 /// \param config_ The configuration struct.
-ocp_nlp_dims *ocp_nlp_dims_create(void *config_);
+ACADOS_SYMBOL_EXPORT ocp_nlp_dims *ocp_nlp_dims_create(void *config_);
 
 /// Destructor of The dimension struct.
 ///
 /// \param dims_ The dimension struct.
-void ocp_nlp_dims_destroy(void *dims_);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_dims_destroy(void *dims_);
 
 
 /// Constructs an input struct for a non-linear programs.
 ///
 /// \param config The configuration struct.
 /// \param dims The dimension struct.
-ocp_nlp_in *ocp_nlp_in_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
+ACADOS_SYMBOL_EXPORT ocp_nlp_in *ocp_nlp_in_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
 
 /// Destructor of the inputs struct.
 ///
 /// \param in The inputs struct.
-void ocp_nlp_in_destroy(void *in);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_in_destroy(void *in);
 
 
 /// Sets the sampling times for the given stage.
@@ -201,9 +202,12 @@ void ocp_nlp_in_destroy(void *in);
 /// \param stage Stage number.
 /// \param field Has to be "Ts" (TBC other options).
 /// \param value The sampling times (floating point).
-void ocp_nlp_in_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in, int stage,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_in_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in, int stage,
         const char *field, void *value);
 
+///
+// void ocp_nlp_in_get(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in, int stage,
+//         const char *field, void *value);
 
 /// Sets the function pointers to the dynamics functions for the given stage.
 ///
@@ -214,7 +218,7 @@ void ocp_nlp_in_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
 /// \param fun_type The name of the function type, either impl_ode_fun,
 ///     impl_ode_fun_jac_x_xdot, impl_ode_jac_x_xdot_u (TBC)
 /// \param fun_ptr Function pointer to the dynamics function.
-int ocp_nlp_dynamics_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
+ACADOS_SYMBOL_EXPORT int ocp_nlp_dynamics_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
         int stage, const char *fun_type, void *fun_ptr);
 
 
@@ -227,7 +231,7 @@ int ocp_nlp_dynamics_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_n
 /// \param field The name of the field, either nls_res_jac,
 ///     y_ref, W (others TBC)
 /// \param value Cost values.
-int ocp_nlp_cost_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
+ACADOS_SYMBOL_EXPORT int ocp_nlp_cost_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_in *in,
         int stage, const char *field, void *value);
 
 
@@ -239,7 +243,7 @@ int ocp_nlp_cost_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_i
 /// \param stage Stage number.
 /// \param field The name of the field, either lb, ub (others TBC)
 /// \param value Constraints function or values.
-int ocp_nlp_constraints_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims,
+ACADOS_SYMBOL_EXPORT int ocp_nlp_constraints_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims,
         ocp_nlp_in *in, int stage, const char *field, void *value);
 
 /* out */
@@ -248,12 +252,12 @@ int ocp_nlp_constraints_model_set(ocp_nlp_config *config, ocp_nlp_dims *dims,
 ///
 /// \param config The configuration struct.
 /// \param dims The dimension struct.
-ocp_nlp_out *ocp_nlp_out_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
+ACADOS_SYMBOL_EXPORT ocp_nlp_out *ocp_nlp_out_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
 
 /// Destructor of the output struct.
 ///
 /// \param out The output struct.
-void ocp_nlp_out_destroy(void *out);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_out_destroy(void *out);
 
 
 /// Sets fields in the output struct of an nlp solver, used to initialize the solver.
@@ -264,7 +268,7 @@ void ocp_nlp_out_destroy(void *out);
 /// \param stage Stage number.
 /// \param field The name of the field, either x, u, pi.
 /// \param value Initialization values.
-void ocp_nlp_out_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_out_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
         int stage, const char *field, void *value);
 
 
@@ -276,21 +280,21 @@ void ocp_nlp_out_set(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *ou
 /// \param stage Stage number.
 /// \param field The name of the field, either x, u, z, pi.
 /// \param value Pointer to the output memory.
-void ocp_nlp_out_get(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_out_get(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
         int stage, const char *field, void *value);
 
 //
-void ocp_nlp_get_at_stage(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_solver *solver,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_get_at_stage(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_solver *solver,
         int stage, const char *field, void *value);
 
 // TODO(andrea): remove this once/if the MATLAB interface uses the new setters below?
-int ocp_nlp_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
+ACADOS_SYMBOL_EXPORT int ocp_nlp_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
         int stage, const char *field);
 
-void ocp_nlp_constraint_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_constraint_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
         int stage, const char *field, int *dims_out);
 
-void ocp_nlp_cost_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_cost_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
         int stage, const char *field, int *dims_out);
 
 void ocp_nlp_dynamics_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *dims, ocp_nlp_out *out,
@@ -302,12 +306,12 @@ void ocp_nlp_dynamics_dims_get_from_attr(ocp_nlp_config *config, ocp_nlp_dims *d
 ///
 /// \param config The configuration struct.
 /// \param dims The dimension struct.
-void *ocp_nlp_solver_opts_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
+ACADOS_SYMBOL_EXPORT void *ocp_nlp_solver_opts_create(ocp_nlp_config *config, ocp_nlp_dims *dims);
 
 /// Destructor of the options.
 ///
 /// \param opts The options struct.
-void ocp_nlp_solver_opts_destroy(void *opts);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_opts_destroy(void *opts);
 
 /// Sets an option.
 ///
@@ -315,10 +319,13 @@ void ocp_nlp_solver_opts_destroy(void *opts);
 /// \param opts_ The options struct.
 /// \param field Name of the option.
 /// \param value Value of the option.
-void ocp_nlp_solver_opts_set(ocp_nlp_config *config, void *opts_, const char *field, void* value);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_opts_set(ocp_nlp_config *config, void *opts_, const char *field, void* value);
+
+
+// ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_opts_get(ocp_nlp_config *config, void *opts_, const char *field, void* value);
 
 
-void ocp_nlp_solver_opts_set_at_stage(ocp_nlp_config *config, void *opts_, int stage, const char *field, void* value);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_opts_set_at_stage(ocp_nlp_config *config, void *opts_, int stage, const char *field, void* value);
 
 
 /// TBC
@@ -327,7 +334,7 @@ void ocp_nlp_solver_opts_set_at_stage(ocp_nlp_config *config, void *opts_, int s
 /// \param config The configuration struct.
 /// \param dims The dimension struct.
 /// \param opts_ The options struct.
-void ocp_nlp_solver_opts_update(ocp_nlp_config *config, ocp_nlp_dims *dims, void *opts_);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_opts_update(ocp_nlp_config *config, ocp_nlp_dims *dims, void *opts_);
 
 
 /* solver */
@@ -338,12 +345,12 @@ void ocp_nlp_solver_opts_update(ocp_nlp_config *config, ocp_nlp_dims *dims, void
 /// \param dims The dimension struct.
 /// \param opts_ The options struct.
 /// \return The solver.
-ocp_nlp_solver *ocp_nlp_solver_create(ocp_nlp_config *config, ocp_nlp_dims *dims, void *opts_);
+ACADOS_SYMBOL_EXPORT ocp_nlp_solver *ocp_nlp_solver_create(ocp_nlp_config *config, ocp_nlp_dims *dims, void *opts_);
 
 /// Destructor of the solver.
 ///
 /// \param solver The solver struct.
-void ocp_nlp_solver_destroy(void *solver);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_destroy(void *solver);
 
 /// Solves the optimal control problem. Call ocp_nlp_precompute before
 /// calling this functions (TBC).
@@ -351,7 +358,17 @@ void ocp_nlp_solver_destroy(void *solver);
 /// \param solver The solver struct.
 /// \param nlp_in The inputs struct.
 /// \param nlp_out The output struct.
-int ocp_nlp_solve(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+ACADOS_SYMBOL_EXPORT int ocp_nlp_solve(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+
+
+
+/// Resets the memory of the QP solver
+///
+/// \param solver The solver struct.
+/// \param nlp_in The inputs struct.
+/// \param nlp_out The output struct.
+ACADOS_SYMBOL_EXPORT void ocp_nlp_solver_reset_qp_memory(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+
 
 /// Performs precomputations for the solver. Needs to be called before
 /// ocl_nlp_solve (TBC).
@@ -359,7 +376,7 @@ int ocp_nlp_solve(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_o
 /// \param solver The solver struct.
 /// \param nlp_in The inputs struct.
 /// \param nlp_out The output struct.
-int ocp_nlp_precompute(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+ACADOS_SYMBOL_EXPORT int ocp_nlp_precompute(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
 
 
 /// Computes cost function value.
@@ -367,7 +384,7 @@ int ocp_nlp_precompute(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *
 /// \param solver The solver struct.
 /// \param nlp_in The inputs struct.
 /// \param nlp_out The output struct.
-void ocp_nlp_eval_cost(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_eval_cost(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
 
 //
 void ocp_nlp_eval_residuals(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
@@ -377,18 +394,18 @@ void ocp_nlp_eval_residuals(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_
 /// \param solver The solver struct.
 /// \param nlp_in The inputs struct.
 /// \param nlp_out The output struct.
-void ocp_nlp_eval_residuals(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_eval_residuals(ocp_nlp_solver *solver, ocp_nlp_in *nlp_in, ocp_nlp_out *nlp_out);
 
 
 //
-void ocp_nlp_eval_param_sens(ocp_nlp_solver *solver, char *field, int stage, int index, ocp_nlp_out *sens_nlp_out);
+ACADOS_SYMBOL_EXPORT void ocp_nlp_eval_param_sens(ocp_nlp_solver *solver, char *field, int stage, int index, ocp_nlp_out *sens_nlp_out);
 
 /* get */
 /// \param config The configuration struct.
 /// \param solver The solver struct.
 /// \param field Supports "sqp_iter", "status", "nlp_res", "time_tot", ...
 /// \param return_value_ Pointer to the output memory.
-void ocp_nlp_get(ocp_nlp_config *config, ocp_nlp_solver *solver,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_get(ocp_nlp_config *config, ocp_nlp_solver *solver,
         const char *field, void *return_value_);
 
 /* set */
@@ -399,7 +416,7 @@ void ocp_nlp_get(ocp_nlp_config *config, ocp_nlp_solver *solver,
 /// \param stage Stage number.
 /// \param field Supports "z_guess", "xdot_guess" (IRK), "phi_guess" (GNSF-IRK)
 /// \param value The initial guess for the algebraic variables in the integrator (if continuous model is used).
-void ocp_nlp_set(ocp_nlp_config *config, ocp_nlp_solver *solver,
+ACADOS_SYMBOL_EXPORT void ocp_nlp_set(ocp_nlp_config *config, ocp_nlp_solver *solver,
         int stage, const char *field, void *value);
 
 

third_party/acados/include/acados_c/sim_interface.h

@@ -74,64 +74,64 @@ typedef struct
 
 /* config */
 //
-sim_config *sim_config_create(sim_solver_plan_t plan);
+ACADOS_SYMBOL_EXPORT sim_config *sim_config_create(sim_solver_plan_t plan);
 //
-void sim_config_destroy(void *config);
+ACADOS_SYMBOL_EXPORT void sim_config_destroy(void *config);
 
 /* dims */
 //
-void *sim_dims_create(void *config_);
+ACADOS_SYMBOL_EXPORT void *sim_dims_create(void *config_);
 //
-void sim_dims_destroy(void *dims);
+ACADOS_SYMBOL_EXPORT void sim_dims_destroy(void *dims);
 //
-void sim_dims_set(sim_config *config, void *dims, const char *field, const int* value);
+ACADOS_SYMBOL_EXPORT void sim_dims_set(sim_config *config, void *dims, const char *field, const int* value);
 //
-void sim_dims_get(sim_config *config, void *dims, const char *field, int* value);
+ACADOS_SYMBOL_EXPORT void sim_dims_get(sim_config *config, void *dims, const char *field, int* value);
 //
-void sim_dims_get_from_attr(sim_config *config, void *dims, const char *field, int *dims_out);
+ACADOS_SYMBOL_EXPORT void sim_dims_get_from_attr(sim_config *config, void *dims, const char *field, int *dims_out);
 
 /* in */
 //
-sim_in *sim_in_create(sim_config *config, void *dims);
+ACADOS_SYMBOL_EXPORT sim_in *sim_in_create(sim_config *config, void *dims);
 //
-void sim_in_destroy(void *out);
+ACADOS_SYMBOL_EXPORT void sim_in_destroy(void *out);
 //
-int sim_in_set(void *config_, void *dims_, sim_in *in, const char *field, void *value);
+ACADOS_SYMBOL_EXPORT int sim_in_set(void *config_, void *dims_, sim_in *in, const char *field, void *value);
 
 
 /* out */
 //
-sim_out *sim_out_create(sim_config *config, void *dims);
+ACADOS_SYMBOL_EXPORT sim_out *sim_out_create(sim_config *config, void *dims);
 //
-void sim_out_destroy(void *out);
+ACADOS_SYMBOL_EXPORT void sim_out_destroy(void *out);
 //
-int sim_out_get(void *config, void *dims, sim_out *out, const char *field, void *value);
+ACADOS_SYMBOL_EXPORT int sim_out_get(void *config, void *dims, sim_out *out, const char *field, void *value);
 
 /* opts */
 //
-void *sim_opts_create(sim_config *config, void *dims);
+ACADOS_SYMBOL_EXPORT void *sim_opts_create(sim_config *config, void *dims);
 //
-void sim_opts_destroy(void *opts);
+ACADOS_SYMBOL_EXPORT void sim_opts_destroy(void *opts);
 //
-void sim_opts_set(sim_config *config, void *opts, const char *field, void *value);
+ACADOS_SYMBOL_EXPORT void sim_opts_set(sim_config *config, void *opts, const char *field, void *value);
 //
-void sim_opts_get(sim_config *config, void *opts, const char *field, void *value);
+ACADOS_SYMBOL_EXPORT void sim_opts_get(sim_config *config, void *opts, const char *field, void *value);
 
 /* solver */
 //
-acados_size_t sim_calculate_size(sim_config *config, void *dims, void *opts_);
+ACADOS_SYMBOL_EXPORT acados_size_t sim_calculate_size(sim_config *config, void *dims, void *opts_);
 //
-sim_solver *sim_assign(sim_config *config, void *dims, void *opts_, void *raw_memory);
+ACADOS_SYMBOL_EXPORT sim_solver *sim_assign(sim_config *config, void *dims, void *opts_, void *raw_memory);
 //
-sim_solver *sim_solver_create(sim_config *config, void *dims, void *opts_);
+ACADOS_SYMBOL_EXPORT sim_solver *sim_solver_create(sim_config *config, void *dims, void *opts_);
 //
-void sim_solver_destroy(void *solver);
+ACADOS_SYMBOL_EXPORT void sim_solver_destroy(void *solver);
 //
-int sim_solve(sim_solver *solver, sim_in *in, sim_out *out);
+ACADOS_SYMBOL_EXPORT int sim_solve(sim_solver *solver, sim_in *in, sim_out *out);
 //
-int sim_precompute(sim_solver *solver, sim_in *in, sim_out *out);
+ACADOS_SYMBOL_EXPORT int sim_precompute(sim_solver *solver, sim_in *in, sim_out *out);
 //
-int sim_solver_set(sim_solver *solver, const char *field, void *value);
+ACADOS_SYMBOL_EXPORT int sim_solver_set(sim_solver *solver, const char *field, void *value);
 
 #ifdef __cplusplus
 } /* extern "C" */

third_party/acados/include/blasfeo/include/blasfeo_block_size.h

@@ -53,6 +53,7 @@
 #define D_PS 8 // panel size
 #define D_PLD 8 // 4 // GCD of panel length
 #define D_M_KERNEL 24 // max kernel size
+#define D_N_KERNEL 8 // max kernel size
 #define D_KC 128 //256 // 192
 #define D_NC 144 //72 //96 //72 // 120 // 512
 #define D_MC 2400 // 6000
@@ -60,6 +61,7 @@
 #define S_PS 16 // panel size
 #define S_PLD 4 // GCD of panel length TODO probably 16 when writing assebly
 #define S_M_KERNEL 32 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 128 //256
 #define S_NC 128 //144
 #define S_MC 3000
@@ -74,13 +76,15 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 12 // max kernel size
+#define D_N_KERNEL 8 // max kernel size
 #define D_KC 256 // 192
-#define D_NC 72 //96 //72 // 120 // 512
+#define D_NC 64 //96 //72 // 120 // 512
-#define D_MC 1500 // 6000
+#define D_MC 1500
 // single
 #define S_PS 8 // panel size
 #define S_PLD 4 // 2 // GCD of panel length
 #define S_M_KERNEL 24 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 144
 #define S_MC 3000
@@ -95,6 +99,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 8 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256 //320 //256 //320
 #define D_NC 72 //64 //72 //60 // 120
 #define D_MC 1000 // 800
@@ -102,6 +107,7 @@
 #define S_PS 8 // panel size
 #define S_PLD 4 // 2 // GCD of panel length
 #define S_M_KERNEL 16 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 144
 #define S_MC 2000
@@ -114,6 +120,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -121,6 +128,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -133,6 +141,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -140,6 +149,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -153,6 +163,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -160,6 +171,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -173,6 +185,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -180,11 +193,35 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
 
 
+#elif defined(TARGET_ARMV8A_APPLE_M1)
+// common
+#define CACHE_LINE_SIZE 64
+#define L1_CACHE_SIZE (128*1024) // L1 data cache size (big cores): 64 kB, ?-way ; DTLB1 ?
+#define LLC_CACHE_SIZE (12*1024*1024) // LLC (L2) cache size (big cores): 12 MB
+// double
+#define D_PS 4 // panel size
+#define D_PLD 4 // 2 // GCD of panel length
+#define D_M_KERNEL 8 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
+#define D_KC 512 //256
+#define D_NC 128 //256
+#define D_MC 6000
+// single
+#define S_PS 4
+#define S_PLD 4 //2
+#define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
+#define S_KC 512
+#define S_NC 256
+#define S_MC 6000
+
+
 #elif defined(TARGET_ARMV8A_ARM_CORTEX_A76)
 // common
 #define CACHE_LINE_SIZE 64
@@ -194,6 +231,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 8 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 512 //256
 #define D_NC 128 //256
 #define D_MC 6000
@@ -201,6 +239,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 512
 #define S_NC 256
 #define S_MC 6000
@@ -215,6 +254,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 8 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 320
 #define D_NC 256
 #define D_MC 6000
@@ -222,6 +262,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -236,13 +277,15 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 8 // max kernel size
-#define D_KC 224 //256 //192
+#define D_N_KERNEL 4 // max kernel size
-#define D_NC 40 //36 //48
+#define D_KC 128 //224 //256 //192
-#define D_MC 512 //488 //600
+#define D_NC 72 //40 //36 //48
+#define D_MC (4*192) //512 //488 //600
 // single
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -257,6 +300,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 12 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 224
 #define D_NC 160
 #define D_MC 6000
@@ -264,6 +308,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -278,6 +323,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 12 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 160
 #define D_NC 128
 #define D_MC 6000
@@ -285,6 +331,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 8 // max kernel size
+#define S_N_KERNEL 8 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -298,6 +345,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -305,6 +353,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -318,6 +367,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -325,6 +375,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -338,6 +389,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -345,6 +397,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy
@@ -358,6 +411,7 @@
 #define D_PS 4 // panel size
 #define D_PLD 4 // 2 // GCD of panel length
 #define D_M_KERNEL 4 // max kernel size
+#define D_N_KERNEL 4 // max kernel size
 #define D_KC 256
 #define D_NC 128 // TODO these are just dummy
 #define D_MC 3000 // TODO these are just dummy
@@ -366,6 +420,7 @@
 #define S_PS 4
 #define S_PLD 4 //2
 #define S_M_KERNEL 4 // max kernel size
+#define S_N_KERNEL 4 // max kernel size
 #define S_KC 256
 #define S_NC 128 // TODO these are just dummy
 #define S_MC 3000 // TODO these are just dummy

third_party/acados/include/blasfeo/include/blasfeo_common.h

@@ -198,6 +198,24 @@ struct blasfeo_pm_smat
 	int memsize; // size of needed memory
 	};
 
+struct blasfeo_pm_dvec
+	{
+	double *mem; // pointer to passed chunk of memory
+	double *pa; // pointer to a pm array of doubles, the first is aligned to cache line size
+	int m; // size
+	int pm; // packed size
+	int memsize; // size of needed memory
+	};
+
+struct blasfeo_pm_svec
+	{
+	float *mem; // pointer to passed chunk of memory
+	float *pa; // pointer to a pm array of floats, the first is aligned to cache line size
+	int m; // size
+	int pm; // packed size
+	int memsize; // size of needed memory
+	};
+
 // Explicitly column-major matrix structure
 struct blasfeo_cm_dmat
 	{
@@ -221,11 +239,31 @@ struct blasfeo_cm_smat
 	int memsize; // size of needed memory
 	};
 
+struct blasfeo_cm_dvec
+	{
+	double *mem; // pointer to passed chunk of memory
+	double *pa; // pointer to a m array of doubles, the first is aligned to cache line size
+	int m; // size
+	int memsize; // size of needed memory
+	};
+
+struct blasfeo_cm_svec
+	{
+	float *mem; // pointer to passed chunk of memory
+	float *pa; // pointer to a m array of floats, the first is aligned to cache line size
+	int m; // size
+	int memsize; // size of needed memory
+	};
+
 
 #define BLASFEO_PM_DMATEL(sA,ai,aj) ((sA)->pA[((ai)-((ai)&((sA)->ps-1)))*(sA)->cn+(aj)*((sA)->ps)+((ai)&((sA)->ps-1))])
 #define BLASFEO_PM_SMATEL(sA,ai,aj) ((sA)->pA[((ai)-((ai)&((sA)->ps-1)))*(sA)->cn+(aj)*((sA)->ps)+((ai)&((sA)->ps-1))])
+#define BLASFEO_PM_DVECEL(sa,ai) ((sa)->pa[ai])
+#define BLASFEO_PM_SVECEL(sa,ai) ((sa)->pa[ai])
 #define BLASFEO_CM_DMATEL(sA,ai,aj) ((sA)->pA[(ai)+(aj)*(sA)->m])
 #define BLASFEO_CM_SMATEL(sA,ai,aj) ((sA)->pA[(ai)+(aj)*(sA)->m])
+#define BLASFEO_CM_DVECEL(sa,ai) ((sa)->pa[ai])
+#define BLASFEO_CM_SVECEL(sa,ai) ((sa)->pa[ai])
 
 
 

third_party/acados/include/blasfeo/include/blasfeo_d_aux.h

@@ -237,6 +237,17 @@ void blasfeo_cm_create_dmat(int m, int n, struct blasfeo_pm_dmat *sA, void *memo
 
 
 
+//
+// BLAS API helper functions
+//
+
+#if ( defined(BLAS_API) & defined(MF_PANELMAJ) )
+// aux
+void blasfeo_cm_dgetr(int m, int n, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj);
+#endif
+
+
+
 #ifdef __cplusplus
 }
 #endif

third_party/acados/include/blasfeo/include/blasfeo_d_blas_api.h

@@ -44,6 +44,30 @@
 
 
 
+#ifdef BLAS_API
+#ifdef CBLAS_API
+#ifndef BLASFEO_CBLAS_ENUM
+#define BLASFEO_CBLAS_ENUM
+#ifdef FORTRAN_BLAS_API
+#ifndef CBLAS_H
+enum CBLAS_ORDER {CblasRowMajor=101, CblasColMajor=102};
+enum CBLAS_TRANSPOSE {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113};
+enum CBLAS_UPLO {CblasUpper=121, CblasLower=122};
+enum CBLAS_DIAG {CblasNonUnit=131, CblasUnit=132};
+enum CBLAS_SIDE {CblasLeft=141, CblasRight=142};
+#endif // CBLAS_H
+#else // FORTRAN_BLAS_API
+enum BLASFEO_CBLAS_ORDER {BlasfeoCblasRowMajor=101, BlasfeoCblasColMajor=102};
+enum BLASFEO_CBLAS_TRANSPOSE {BlasfeoCblasNoTrans=111, BlasfeoCblasTrans=112, BlasfeoCblasConjTrans=113};
+enum BLASFEO_CBLAS_UPLO {BlasfeoCblasUpper=121, BlasfeoCblasLower=122};
+enum BLASFEO_CBLAS_DIAG {BlasfeoCblasNonUnit=131, BlasfeoCblasUnit=132};
+enum BLASFEO_CBLAS_SIDE {BlasfeoCblasLeft=141, BlasfeoCblasRight=142};
+#endif // FORTRAN_BLAS_API
+#endif // BLASFEO_CBLAS_ENUM
+#endif // CBLAS_API
+#endif // BLAS_API
+
+
 #ifdef __cplusplus
 extern "C" {
 #endif
@@ -66,6 +90,14 @@ void dcopy_(int *n, double *x, int *incx, double *y, int *incy);
 //
 double ddot_(int *n, double *x, int *incx, double *y, int *incy);
 
+// BLAS 2
+//
+void dgemv_(char *tran, int *m, int *n, double *alpha, double *A, int *lda, double *x, int *incx, double *beta, double *y, int *incy);
+//
+void dsymv_(char *uplo, int *n, double *alpha, double *A, int *lda, double *x, int *incx, double *beta, double *y, int *incy);
+//
+void dger_(int *m, int *n, double *alpha, double *x, int *incx, double *y, int *incy, double *A, int *lda);
+
 // BLAS 3
 //
 void dgemm_(char *ta, char *tb, int *m, int *n, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
@@ -75,6 +107,8 @@ void dsyrk_(char *uplo, char *ta, int *m, int *k, double *alpha, double *A, int
 void dtrmm_(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
 //
 void dtrsm_(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
+//
+void dsyr2k_(char *uplo, char *ta, int *m, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
 
 
 
@@ -100,49 +134,125 @@ void dtrtrs_(char *uplo, char *trans, char *diag, int *m, int *n, double *A, int
 
 
 
+// aux
+void dgetr_(int *m, int *n, double *A, int *lda, double *B, int *ldb);
+
+
+
+#ifdef CBLAS_API
+
+
+
+// CBLAS 1
+//
+void cblas_daxpy(const int N, const double alpha, const double *X, const int incX, double *Y, const int incY);
+//
+void cblas_dswap(const int N, double *X, const int incX, double *Y, const int incY);
+//
+void cblas_dcopy(const int N, const double *X, const int incX, double *Y, const int incY);
+
+// CBLAS 3
+//
+void cblas_dgemm(const enum CBLAS_ORDER Order, const enum CBLAS_TRANSPOSE TransA, const enum CBLAS_TRANSPOSE TransB, const int M, const int N, const int K, const double alpha, const double *A, const int lda, const double *B, const int ldb, const double beta, double *C, const int ldc);
+//
+void cblas_dsyrk(const enum CBLAS_ORDER Order, const enum CBLAS_UPLO Uplo, const enum CBLAS_TRANSPOSE Trans, const int N, const int K, const double alpha, const double *A, const int lda, const double beta, double *C, const int ldc);
+//
+void cblas_dtrmm(const enum CBLAS_ORDER Order, const enum CBLAS_SIDE Side, const enum CBLAS_UPLO Uplo, const enum CBLAS_TRANSPOSE TransA, const enum CBLAS_DIAG Diag, const int M, const int N, const double alpha, const double *A, const int lda, double *B, const int ldb);
+//
+void cblas_dtrsm(const enum CBLAS_ORDER Order, const enum CBLAS_SIDE Side, const enum CBLAS_UPLO Uplo, const enum CBLAS_TRANSPOSE TransA, const enum CBLAS_DIAG Diag, const int M, const int N, const double alpha, const double *A, const int lda, double *B, const int ldb);
+
+
+
+#endif // CBLAS_API
+
+
+
 #else // BLASFEO_API
 
 
 
 // BLAS 1
 //
-void blas_daxpy(int *n, double *alpha, double *x, int *incx, double *y, int *incy);
+void blasfeo_blas_daxpy(int *n, double *alpha, double *x, int *incx, double *y, int *incy);
 //
-double blas_ddot(int *n, double *x, int *incx, double *y, int *incy);
+double blasfeo_blas_ddot(int *n, double *x, int *incx, double *y, int *incy);
 //
-void blas_dcopy(int *n, double *x, int *incx, double *y, int *incy);
+void blasfeo_blas_dcopy(int *n, double *x, int *incx, double *y, int *incy);
+
+// BLAS 2
+//
+void blasfeo_blas_dgemv(char *trans, int *m, int *n, double *alpha, double *A, int *lda, double *x, int *incx, double *beta, double *y, int *incy);
+//
+void blasfeo_blas_dsymv(char *uplo, int *n, double *alpha, double *A, int *lda, double *x, int *incx, double *beta, double *y, int *incy);
+//
+void blasfeo_blas_dger(int *m, int *n, double *alpha, double *x, int *incx, double *y, int *incy, double *A, int *lda);
 
 // BLAS 3
 //
-void blas_dgemm(char *ta, char *tb, int *m, int *n, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
+void blasfeo_blas_dgemm(char *ta, char *tb, int *m, int *n, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
+//
+void blasfeo_blas_dsyrk(char *uplo, char *ta, int *m, int *k, double *alpha, double *A, int *lda, double *beta, double *C, int *ldc);
 //
-void blas_dsyrk(char *uplo, char *ta, int *m, int *k, double *alpha, double *A, int *lda, double *beta, double *C, int *ldc);
+void blasfeo_blas_dtrmm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
 //
-void blas_dtrmm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
+void blasfeo_blas_dtrsm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
 //
-void blas_dtrsm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
+void blasfeo_blas_dsyr2k(char *uplo, char *ta, int *m, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
 
 
 
 // LAPACK
 //
-void blas_dgesv(int *m, int *n, double *A, int *lda, int *ipiv, double *B, int *ldb, int *info);
+void blasfeo_lapack_dgesv(int *m, int *n, double *A, int *lda, int *ipiv, double *B, int *ldb, int *info);
 //
-void blas_dgetrf(int *m, int *n, double *A, int *lda, int *ipiv, int *info);
+void blasfeo_lapack_dgetrf(int *m, int *n, double *A, int *lda, int *ipiv, int *info);
 //
-void blas_dgetrf_np(int *m, int *n, double *A, int *lda, int *info);
+void blasfeo_lapack_dgetrf_np(int *m, int *n, double *A, int *lda, int *info);
 //
-void blas_dgetrs(char *trans, int *m, int *n, double *A, int *lda, int *ipiv, double *B, int *ldb, int *info);
+void blasfeo_lapack_dgetrs(char *trans, int *m, int *n, double *A, int *lda, int *ipiv, double *B, int *ldb, int *info);
 //
-void blas_dlaswp(int *n, double *A, int *lda, int *k1, int *k2, int *ipiv, int *incx);
+void blasfeo_lapack_dlaswp(int *n, double *A, int *lda, int *k1, int *k2, int *ipiv, int *incx);
 //
-void blas_dposv(char *uplo, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
+void blasfeo_lapack_dposv(char *uplo, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
 //
-void blas_dpotrf(char *uplo, int *m, double *A, int *lda, int *info);
+void blasfeo_lapack_dpotrf(char *uplo, int *m, double *A, int *lda, int *info);
 //
-void blas_dpotrs(char *uplo, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
+void blasfeo_lapack_dpotrs(char *uplo, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
 //
-void blas_dtrtrs(char *uplo, char *trans, char *diag, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
+void blasfeo_lapack_dtrtrs(char *uplo, char *trans, char *diag, int *m, int *n, double *A, int *lda, double *B, int *ldb, int *info);
+
+
+
+// aux
+void blasfeo_blas_dgetr(int *m, int *n, double *A, int *lda, double *B, int *ldb);
+
+
+
+#ifdef CBLAS_API
+
+
+
+// CBLAS 1
+//
+void blasfeo_cblas_daxpy(const int N, const double alpha, const double *X, const int incX, double *Y, const int incY);
+//
+void blasfeo_cblas_dswap(const int N, double *X, const int incX, double *Y, const int incY);
+//
+void blasfeo_cblas_dcopy(const int N, const double *X, const int incX, double *Y, const int incY);
+
+// CBLAS 3
+//
+void blasfeo_cblas_dgemm(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_TRANSPOSE TransA, const enum BLASFEO_CBLAS_TRANSPOSE TransB, const int M, const int N, const int K, const double alpha, const double *A, const int lda, const double *B, const int ldb, const double beta, double *C, const int ldc);
+//
+void blasfeo_cblas_dsyrk(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_UPLO Uplo, const enum BLASFEO_CBLAS_TRANSPOSE Trans, const int N, const int K, const double alpha, const double *A, const int lda, const double beta, double *C, const int ldc);
+//
+void blasfeo_cblas_dtrmm(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_SIDE Side, const enum BLASFEO_CBLAS_UPLO Uplo, const enum BLASFEO_CBLAS_TRANSPOSE TransA, const enum BLASFEO_CBLAS_DIAG Diag, const int M, const int N, const double alpha, const double *A, const int lda, double *B, const int ldb);
+//
+void blasfeo_cblas_dtrsm(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_SIDE Side, const enum BLASFEO_CBLAS_UPLO Uplo, const enum BLASFEO_CBLAS_TRANSPOSE TransA, const enum BLASFEO_CBLAS_DIAG Diag, const int M, const int N, const double alpha, const double *A, const int lda, double *B, const int ldb);
+
+
+
+#endif // CBLAS_API
 
 
 

third_party/acados/include/blasfeo/include/blasfeo_d_blasfeo_api.h

@@ -119,6 +119,10 @@ void blasfeo_dgemv_nt(int m, int n, double alpha_n, double alpha_t, struct blasf
 // z <= beta * y + alpha * A * x, where A is symmetric and only the lower triangular patr of A is accessed
 void blasfeo_dsymv_l(int m, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
 void blasfeo_dsymv_l_mn(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
+// z <= beta * y + alpha * A * x, where A is symmetric and only the upper triangular patr of A is accessed
+void blasfeo_dsymv_u(int m, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
+// D = C + alpha * x * y^T
+void blasfeo_dger(int m, int n, double alpha, struct blasfeo_dvec *sx, int xi, struct blasfeo_dvec *sy, int yi, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
 
 // diagonal
 
@@ -227,6 +231,14 @@ void blasfeo_dtrsm_runu(int m, int n, double alpha, struct blasfeo_dmat *sA, int
 void blasfeo_dtrsm_rutn(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, struct blasfeo_dmat *sD, int di, int dj);
 // D <= alpha * B * A^{-T} , with A upper triangular with unit diagonal
 void blasfeo_dtrsm_rutu(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T; C, D lower triangular
+void blasfeo_dsyr2k_ln(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A; C, D lower triangular
+void blasfeo_dsyr2k_lt(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T; C, D upper triangular
+void blasfeo_dsyr2k_un(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A; C, D upper triangular
+void blasfeo_dsyr2k_ut(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
 
 // diagonal
 
@@ -327,6 +339,16 @@ void blasfeo_cm_dtrmm_runn(int m, int n, double alpha, struct blasfeo_cm_dmat *s
 void blasfeo_cm_dtrmm_runu(int m, int n, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, struct blasfeo_cm_dmat *sD, int di, int dj);
 void blasfeo_cm_dtrmm_rutn(int m, int n, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, struct blasfeo_cm_dmat *sD, int di, int dj);
 void blasfeo_cm_dtrmm_rutu(int m, int n, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, struct blasfeo_cm_dmat *sD, int di, int dj);
+void blasfeo_cm_dsyr2k_ln(int m, int k, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, double beta, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
+void blasfeo_cm_dsyr2k_lt(int m, int k, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, double beta, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
+void blasfeo_cm_dsyr2k_un(int m, int k, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, double beta, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
+void blasfeo_cm_dsyr2k_ut(int m, int k, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dmat *sB, int bi, int bj, double beta, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
+// BLAS 2
+void blasfeo_cm_dgemv_n(int m, int n, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dvec *sx, int xi, double beta, struct blasfeo_cm_dvec *sy, int yi, struct blasfeo_cm_dvec *sz, int zi);
+void blasfeo_cm_dgemv_t(int m, int n, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dvec *sx, int xi, double beta, struct blasfeo_cm_dvec *sy, int yi, struct blasfeo_cm_dvec *sz, int zi);
+void blasfeo_cm_dsymv_l(int m, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dvec *sx, int xi, double beta, struct blasfeo_cm_dvec *sy, int yi, struct blasfeo_cm_dvec *sz, int zi);
+void blasfeo_cm_dsymv_u(int m, double alpha, struct blasfeo_cm_dmat *sA, int ai, int aj, struct blasfeo_cm_dvec *sx, int xi, double beta, struct blasfeo_cm_dvec *sy, int yi, struct blasfeo_cm_dvec *sz, int zi);
+void blasfeo_cm_dger(int m, int n, double alpha, struct blasfeo_cm_dvec *sx, int xi, struct blasfeo_cm_dvec *sy, int yi, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
 // LAPACK
 void blasfeo_cm_dpotrf_l(int m, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);
 void blasfeo_cm_dpotrf_u(int m, struct blasfeo_cm_dmat *sC, int ci, int cj, struct blasfeo_cm_dmat *sD, int di, int dj);

third_party/acados/include/blasfeo/include/blasfeo_d_blasfeo_hp_api.h

@@ -0,0 +1,84 @@
+/**************************************************************************************************
+*                                                                                                 *
+* This file is part of BLASFEO.                                                                   *
+*                                                                                                 *
+* BLASFEO -- BLAS For Embedded Optimization.                                                      *
+* Copyright (C) 2019 by Gianluca Frison.                                                          *
+* Developed at IMTEK (University of Freiburg) under the supervision of Moritz Diehl.              *
+* All rights reserved.                                                                            *
+*                                                                                                 *
+* The 2-Clause BSD License                                                                        *
+*                                                                                                 *
+* Redistribution and use in source and binary forms, with or without                              *
+* modification, are permitted provided that the following conditions are met:                     *
+*                                                                                                 *
+* 1. Redistributions of source code must retain the above copyright notice, this                  *
+*    list of conditions and the following disclaimer.                                             *
+* 2. Redistributions in binary form must reproduce the above copyright notice,                    *
+*    this list of conditions and the following disclaimer in the documentation                    *
+*    and/or other materials provided with the distribution.                                       *
+*                                                                                                 *
+* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND                 *
+* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED                   *
+* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE                          *
+* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR                 *
+* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES                  *
+* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;                    *
+* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND                     *
+* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT                      *
+* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS                   *
+* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.                                    *
+*                                                                                                 *
+* Author: Gianluca Frison, gianluca.frison (at) imtek.uni-freiburg.de                             *
+*                                                                                                 *
+**************************************************************************************************/
+
+#ifndef BLASFEO_D_BLASFEO_HP_API_H_
+#define BLASFEO_D_BLASFEO_HP_API_H_
+
+
+
+#include "blasfeo_common.h"
+
+
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+
+
+//
+// level 3 BLAS
+//
+
+// dense
+
+
+// D <= beta * C + alpha * A^T * B
+void blasfeo_hp_dgemm_tn(int m, int n, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T; C, D lower triangular
+void blasfeo_hp_dsyrk_ln(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * A^T ; C, D lower triangular
+void blasfeo_hp_dsyrk3_ln(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= alpha * B * A^{-T} , with A lower triangular
+void blasfeo_hp_dtrsm_rltn(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, struct blasfeo_dmat *sD, int di, int dj);
+
+
+//
+// level 2 BLAS
+//
+
+// dense
+
+// z <= beta * y + alpha * A * x
+void blasfeo_hp_dgemv_n(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
+
+
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif  // BLASFEO_D_BLASFEO_HP_API_H_
+

third_party/acados/include/blasfeo/include/blasfeo_d_blasfeo_ref_api.h

@@ -119,6 +119,10 @@ void blasfeo_ref_dgemv_nt(int m, int n, double alpha_n, double alpha_t, struct b
 // z <= beta * y + alpha * A * x, where A is symmetric and only the lower triangular patr of A is accessed
 void blasfeo_ref_dsymv_l(int m, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
 void blasfeo_ref_dsymv_l_mn(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
+// z <= beta * y + alpha * A * x, where A is symmetric and only the upper triangular patr of A is accessed
+void blasfeo_ref_dsymv_u(int m, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dvec *sx, int xi, double beta, struct blasfeo_dvec *sy, int yi, struct blasfeo_dvec *sz, int zi);
+// D = C + alpha * x * y^T
+void blasfeo_ref_dger(int m, int n, double alpha, struct blasfeo_dvec *sx, int xi, struct blasfeo_dvec *sy, int yi, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
 
 // diagonal
 
@@ -214,6 +218,14 @@ void blasfeo_ref_dtrsm_runu(int m, int n, double alpha, struct blasfeo_dmat *sA,
 void blasfeo_ref_dtrsm_rutn(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, struct blasfeo_dmat *sD, int di, int dj);
 // D <= alpha * B * A^{-T} , with A upper triangular with unit diagonal
 void blasfeo_ref_dtrsm_rutu(int m, int n, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T ; C, D lower triangular
+void blasfeo_ref_dsyr2k_ln(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A ; C, D lower triangular
+void blasfeo_ref_dsyr2k_lt(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T ; C, D upper triangular
+void blasfeo_ref_dsyr2k_un(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A ; C, D upper triangular
+void blasfeo_ref_dsyr2k_ut(int m, int k, double alpha, struct blasfeo_dmat *sA, int ai, int aj, struct blasfeo_dmat *sB, int bi, int bj, double beta, struct blasfeo_dmat *sC, int ci, int cj, struct blasfeo_dmat *sD, int di, int dj);
 
 // diagonal
 

third_party/acados/include/blasfeo/include/blasfeo_d_kernel.h

@@ -728,14 +728,22 @@ void kernel_dsyrk_nt_l_12x4_lib44cc(int kmax, double *alpha, double *A, int sda,
 void kernel_dsyrk_nt_l_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_u_12x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_u_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_12x4_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_12x4_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dsyr2k_nt_l_12x4_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dsyr2k_nt_l_12x4_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_12x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_one_12x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_one_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrmm_nt_ru_12x4_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd);
+void kernel_dtrmm_nt_ru_12x4_vs_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_12x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_12x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_12x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nt_ll_inv_12x4_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E);
+void kernel_dtrsm_nt_ll_inv_12x4_vs_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nt_rl_inv_12x4_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE);
 void kernel_dtrsm_nt_rl_inv_12x4_vs_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE, int m1, int n1);
 void kernel_dtrsm_nt_rl_inv_12x4_lib44ccc(int kmax, double *A, int sda, double *B, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -751,6 +759,7 @@ void kernel_dpotrf_nt_l_12x4_vs_lib44cc(int kmax, double *A, int sda, double *B,
 // 4x12
 void kernel_dgemm_nt_4x12_lib44cc(int kmax, double *alpha, double *A, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dgemm_nt_4x12_vs_lib44cc(int kmax, double *alpha, double *A, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dgemm_nt_4x12_p0_vs_lib44cc(int kmax, double *alpha, double *A, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1, double *A_p, double *B_p);
 void kernel_dtrmm_nt_rl_4x12_tran_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_4x12_tran_vs_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_one_4x12_tran_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd);
@@ -764,6 +773,10 @@ void kernel_dsyrk_nt_l_8x8_lib44cc(int kmax, double *alpha, double *A, int sda,
 void kernel_dsyrk_nt_l_8x8_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_u_8x8_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_u_8x8_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_8x8_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, int sdb0, double *A1, int sda1, double *B1, int sdb1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_8x8_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, int sdb0, double *A1, int sda1, double *B1, int sdb1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dsyr2k_nt_l_8x8_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, int sdb0, double *A1, int sda1, double *B1, int sdb1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dsyr2k_nt_l_8x8_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, int sdb0, double *A1, int sda1, double *B1, int sdb1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dpotrf_nt_l_8x8_lib44cc(int kmax, double *A, int sda, double *B, int sdb, double *C, int ldc, double *D, int ldd, double *dD);
 void kernel_dpotrf_nt_l_8x8_vs_lib44cc(int kmax, double *A, int sda, double *B, int sdb, double *C, int ldc, double *D, int ldd, double *dD, int m1, int n1);
 // 8x4
@@ -775,14 +788,22 @@ void kernel_dsyrk_nt_l_8x4_lib44cc(int kmax, double *alpha, double *A, int sda,
 void kernel_dsyrk_nt_l_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_u_8x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_u_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_8x4_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_8x4_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dsyr2k_nt_l_8x4_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dsyr2k_nt_l_8x4_vs_lib44cc(int kmax, double *alpha, double *A0, int sda0, double *B0, double *A1, int sda1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_8x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_one_8x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_one_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrmm_nt_ru_8x4_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd);
+void kernel_dtrmm_nt_ru_8x4_vs_lib444c(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int sdc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_8x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_8x4_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_8x4_vs_lib44cc(int kmax, double *alpha, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nt_ll_inv_8x4_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E);
+void kernel_dtrsm_nt_ll_inv_8x4_vs_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nt_rl_inv_8x4_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE);
 void kernel_dtrsm_nt_rl_inv_8x4_vs_lib44cc4(int kmax, double *A, int sda, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE, int m1, int n1);
 void kernel_dtrsm_nt_rl_inv_8x4_lib44ccc(int kmax, double *A, int sda, double *B, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -813,6 +834,10 @@ void kernel_dsyrk_nt_l_4x4_lib44cc(int kmax, double *alpha, double *A, double *B
 void kernel_dsyrk_nt_l_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_u_4x4_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_u_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_4x4_lib44cc(int kmax, double *alpha, double *A0, double *B0, double *A1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_4x4_vs_lib44cc(int kmax, double *alpha, double *A0, double *B0, double *A1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dsyr2k_nt_l_4x4_lib44cc(int kmax, double *alpha, double *A0, double *B0, double *A1, double *B1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dsyr2k_nt_l_4x4_vs_lib44cc(int kmax, double *alpha, double *A0, double *B0, double *A1, double *B1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_4x4_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_rl_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_4x4_tran_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd);
@@ -821,6 +846,8 @@ void kernel_dtrmm_nt_rl_one_4x4_lib44cc(int kmax, double *alpha, double *A, doub
 void kernel_dtrmm_nt_rl_one_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_rl_one_4x4_tran_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd);
 void kernel_dtrmm_nt_rl_one_4x4_tran_vs_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd, int m1, int n1);
+void kernel_dtrmm_nt_ru_4x4_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd);
+void kernel_dtrmm_nt_ru_4x4_vs_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_4x4_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_4x4_tran_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd);
@@ -829,6 +856,8 @@ void kernel_dtrmm_nt_ru_one_4x4_lib44cc(int kmax, double *alpha, double *A, doub
 void kernel_dtrmm_nt_ru_one_4x4_vs_lib44cc(int kmax, double *alpha, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_4x4_tran_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_4x4_tran_vs_lib444c(int kmax, double *alpha, double *A, double *B, double *beta, double *C, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nt_ll_inv_4x4_lib44cc4(int kmax, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *inv_diag_E);
+void kernel_dtrsm_nt_ll_inv_4x4_vs_lib44cc4(int kmax, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nt_rl_inv_4x4_lib44cc4(int kmax, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE);
 void kernel_dtrsm_nt_rl_inv_4x4_vs_lib44cc4(int kmax, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *dE, int m1, int n1);
 void kernel_dtrsm_nt_rl_inv_4x4_lib44ccc(int kmax, double *A, double *B, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -852,6 +881,8 @@ void kernel_dgemm_nt_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda, d
 void kernel_dgemm_nt_12x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_l_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_l_12x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_12x4_lib4ccc(int kmax, double *alpha, double *A0, int sda0, double *B0, int ldb0, double *A1, int sda1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_12x4_vs_lib4ccc(int kmax, double *alpha, double *A0, int sda0, double *B0, int ldb0, double *A1, int sda1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nn_rl_12x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_one_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
@@ -868,6 +899,8 @@ void kernel_dtrmm_nt_ru_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda
 void kernel_dtrmm_nt_ru_12x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_12x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_12x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nn_ll_inv_12x4_lib4ccc4(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E);
+void kernel_dtrsm_nn_ll_inv_12x4_vs_lib4ccc4(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nn_ll_one_12x4_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde);
 void kernel_dtrsm_nn_ll_one_12x4_vs_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, int m1, int n1);
 void kernel_dtrsm_nn_rl_inv_12x4_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -920,6 +953,8 @@ void kernel_dgemm_nt_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda, do
 void kernel_dgemm_nt_8x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_l_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_l_8x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_8x4_lib4ccc(int kmax, double *alpha, double *A0, int sda0, double *B0, int ldb0, double *A1, int sda1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_8x4_vs_lib4ccc(int kmax, double *alpha, double *A0, int sda0, double *B0, int ldb0, double *A1, int sda1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nn_rl_8x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_one_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
@@ -936,6 +971,8 @@ void kernel_dtrmm_nt_ru_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda,
 void kernel_dtrmm_nt_ru_8x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_8x4_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_8x4_vs_lib4ccc(int kmax, double *alpha, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nn_ll_inv_8x4_lib4ccc4(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E);
+void kernel_dtrsm_nn_ll_inv_8x4_vs_lib4ccc4(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int sde, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nn_ll_one_8x4_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde);
 void kernel_dtrsm_nn_ll_one_8x4_vs_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, int m1, int n1);
 void kernel_dtrsm_nn_rl_inv_8x4_lib4cccc(int kmax, double *A, int sda, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -986,6 +1023,8 @@ void kernel_dgemm_nt_4x4_lib4ccc(int kmax, double *alpha, double *A, double *B,
 void kernel_dgemm_nt_4x4_vs_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dsyrk_nt_l_4x4_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dsyrk_nt_l_4x4_vs_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
+void kernel_dger2k_nt_4x4_lib4ccc(int kmax, double *alpha, double *A0, double *B0, int ldb0, double *A1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd);
+void kernel_dger2k_nt_4x4_vs_lib4ccc(int kmax, double *alpha, double *A0, double *B0, int ldb0, double *A1, double *B1, int ldb1, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_4x4_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd);
 void kernel_dtrmm_nn_rl_4x4_vs_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nn_rl_4x4_tran_lib4c4c(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, double *D, int ldd);
@@ -1018,6 +1057,8 @@ void kernel_dtrmm_nt_ru_one_4x4_lib4ccc(int kmax, double *alpha, double *A, doub
 void kernel_dtrmm_nt_ru_one_4x4_vs_lib4ccc(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 void kernel_dtrmm_nt_ru_one_4x4_tran_lib4c4c(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, double *D, int ldd);
 void kernel_dtrmm_nt_ru_one_4x4_tran_vs_lib4c4c(int kmax, double *alpha, double *A, double *B, int ldb, double *beta, double *C, double *D, int ldd, int m1, int n1);
+void kernel_dtrsm_nn_ll_inv_4x4_lib4ccc4(int kmax, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *inv_diag_E);
+void kernel_dtrsm_nn_ll_inv_4x4_vs_lib4ccc4(int kmax, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, double *inv_diag_E, int km, int kn);
 void kernel_dtrsm_nn_ll_one_4x4_lib4cccc(int kmax, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde);
 void kernel_dtrsm_nn_ll_one_4x4_vs_lib4cccc(int kmax, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, int m1, int n1);
 void kernel_dtrsm_nn_rl_inv_4x4_lib4cccc(int kmax, double *A, double *B, int ldb, double *beta, double *C, int ldc, double *D, int ldd, double *E, int lde, double *dE);
@@ -1181,6 +1222,7 @@ void kernel_drowsw_lib(int kmax, double *pA, int lda, double *pC, int ldc);
 // 12
 void kernel_dgemm_nt_12xn_p0_lib44cc(int n, int k, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, double *A_p, double *B_p);
 void kernel_dgemm_nt_12xn_pl_lib44cc(int n, int k, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, double *A_p, double *B_p);
+void kernel_dgemm_nt_mx12_p0_lib44cc(int m, int k, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, double *A_p, double *B_p);
 // 8
 void kernel_dgemm_nt_8xn_p0_lib44cc(int n, int k, double *alpha, double *A, int sda, double *B, int sdb, double *beta, double *C, int ldc, double *D, int ldd, double *A_p, double *B_p);
 
@@ -1232,11 +1274,43 @@ void kernel_dgemm_tt_8x8_libc8cc(int kmax, double *alpha, double *A, int lda, do
 void kernel_dgemm_tt_8x8_vs_libc8cc(int kmax, double *alpha, double *A, int lda, double *B, double *beta, double *C, int ldc, double *D, int ldd, int m1, int n1);
 
 
+// level 2 BLAS
+void kernel_dgemv_n_4_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z);
+void kernel_dgemv_n_4_vs_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z, int km);
+void kernel_dgemv_t_4_libc(int kmax, double *alpha, double *A, int lda, double *x, double *beta, double *y, double *z);
+void kernel_dgemv_t_4_vs_libc(int kmax, double *alpha, double *A, int lda, double *x, double *beta, double *y, double *z, int km);
+void kernel_dsymv_l_4_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z);
+void kernel_dsymv_l_4_vs_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z, int km);
+void kernel_dsymv_u_4_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z);
+void kernel_dsymv_u_4_vs_libc(int kmax, double *alpha, double *A, int lda, double *x, double *z, int km);
+void kernel_dger_4_libc(int kmax, double *alpha, double *x, double *y, double *C, int ldc, double *D, int ldd);
+void kernel_dger_4_vs_libc(int kmax, double *alpha, double *x, double *y, double *C, int ldc, double *D, int ldd, int km);
+
+
 
 // aux
 void kernel_dvecld_inc1(int kmax, double *x);
 void kernel_dveccp_inc1(int kmax, double *x, double *y);
 
+//void kernel_dgetr_nt_8_p0_lib(int kmax, double *A, int lda, double *C, int ldc, double *Ap, double *Bp);
+//void kernel_dgetr_nt_8_lib(int kmax, double *A, int lda, double *C, int ldc);
+//void kernel_dgetr_nt_4_lib(int kmax, double *A, int lda, double *C, int ldc);
+void kernel_dgetr_tn_8_p0_lib(int kmax, double *A, int lda, double *C, int ldc, double *Ap, double *Bp);
+void kernel_dgetr_tn_8_lib(int kmax, double *A, int lda, double *C, int ldc);
+void kernel_dgetr_tn_4_lib(int kmax, double *A, int lda, double *C, int ldc);
+void kernel_dgetr_tn_4_vs_lib(int kmax, double *A, int lda, double *C, int ldc, int m1);
+
+
+// building blocks for blocked algorithms
+//
+void blasfeo_hp_dgemm_nt_m2(int m, int n, int k, double alpha, double *pA, int sda, double *pB, int sdb, double beta, double *C, int ldc, double *D, int ldd);
+void blasfeo_hp_dgemm_nt_n2(int m, int n, int k, double alpha, double *pA, int sda, double *pB, int sdb, double beta, double *C, int ldc, double *D, int ldd);
+//
+void kernel_dpack_buffer_fn(int m, int n, double *A, int lda, double *pA, int sda);
+void kernel_dpack_buffer_ft(int m, int n, double *A, int lda, double *pA, int sda);
+void kernel_dpack_buffer_ln(int m, double *A, int lda, double *pA, int sda);
+void kernel_dpack_buffer_lt(int m, double *A, int lda, double *pA, int sda);
+void kernel_dpack_buffer_ut(int m, double *A, int lda, double *pA, int sda);
 
 
 

third_party/acados/include/blasfeo/include/blasfeo_processor_features.h

@@ -39,20 +39,20 @@
 /**
  * Flags to indicate the different processor features
  */
-enum
+//enum
-{
+//{
-    // x86-64 CPU features
+//    // x86-64 CPU features
-    BLASFEO_PROCESSOR_FEATURE_AVX  = 0x0001,    /// AVX instruction set
+//    BLASFEO_PROCESSOR_FEATURE_AVX  = 0x0001,    /// AVX instruction set
-    BLASFEO_PROCESSOR_FEATURE_AVX2 = 0x0002,    /// AVX2 instruction set
+//    BLASFEO_PROCESSOR_FEATURE_AVX2 = 0x0002,    /// AVX2 instruction set
-    BLASFEO_PROCESSOR_FEATURE_FMA  = 0x0004,    /// FMA instruction set
+//    BLASFEO_PROCESSOR_FEATURE_FMA  = 0x0004,    /// FMA instruction set
-    BLASFEO_PROCESSOR_FEATURE_SSE3 = 0x0008,    /// SSE3 instruction set
+//    BLASFEO_PROCESSOR_FEATURE_SSE3 = 0x0008,    /// SSE3 instruction set
-
+//
-    // ARM CPU features
+//    // ARM CPU features
-    BLASFEO_PROCESSOR_FEATURE_VFPv3  = 0x0100,  /// VFPv3 instruction set
+//    BLASFEO_PROCESSOR_FEATURE_VFPv3  = 0x0100,  /// VFPv3 instruction set
-    BLASFEO_PROCESSOR_FEATURE_NEON   = 0x0100,  /// NEON instruction set
+//    BLASFEO_PROCESSOR_FEATURE_NEON   = 0x0100,  /// NEON instruction set
-    BLASFEO_PROCESSOR_FEATURE_VFPv4  = 0x0100,  /// VFPv4 instruction set
+//    BLASFEO_PROCESSOR_FEATURE_VFPv4  = 0x0100,  /// VFPv4 instruction set
-    BLASFEO_PROCESSOR_FEATURE_NEONv2 = 0x0100,  /// NEONv2 instruction set
+//    BLASFEO_PROCESSOR_FEATURE_NEONv2 = 0x0100,  /// NEONv2 instruction set
-} BLASFEO_PROCESSOR_FEATURES;
+//} BLASFEO_PROCESSOR_FEATURES;
 
 /**
  * Test the features that this processor provides against what the library was compiled with.

third_party/acados/include/blasfeo/include/blasfeo_s_blas_api.h

@@ -44,6 +44,31 @@
 
 
 
+#ifdef BLAS_API
+#ifdef CBLAS_API
+#ifndef BLASFEO_CBLAS_ENUM
+#define BLASFEO_CBLAS_ENUM
+#ifdef FORTRAN_BLAS_API
+#ifndef CBLAS_H
+enum CBLAS_ORDER {CblasRowMajor=101, CblasColMajor=102};
+enum CBLAS_TRANSPOSE {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113};
+enum CBLAS_UPLO {CblasUpper=121, CblasLower=122};
+enum CBLAS_DIAG {CblasNonUnit=131, CblasUnit=132};
+enum CBLAS_SIDE {CblasLeft=141, CblasRight=142};
+#endif // CBLAS_H
+#else // FORTRAN_BLAS_API
+enum BLASFEO_CBLAS_ORDER {CblasRowMajor=101, CblasColMajor=102};
+enum BLASFEO_CBLAS_TRANSPOSE {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113};
+enum BLASFEO_CBLAS_UPLO {CblasUpper=121, CblasLower=122};
+enum BLASFEO_CBLAS_DIAG {CblasNonUnit=131, CblasUnit=132};
+enum BLASFEO_CBLAS_SIDE {CblasLeft=141, CblasRight=142};
+#endif // FORTRAN_BLAS_API
+#endif // BLASFEO_CBLAS_ENUM
+#endif // CBLAS_API
+#endif // BLAS_API
+
+
+
 #ifdef __cplusplus
 extern "C" {
 #endif
@@ -78,27 +103,67 @@ void spotrf_(char *uplo, int *m, float *A, int *lda, int *info);
 
 
 
+#ifdef CBLAS_API
+
+
+
+// CBLAS 1
+//
+void cblas_saxpy(const int N, const float alpha, const float *X, const int incX, float *Y, const int incY);
+
+// CBLAS 3
+//
+void cblas_sgemm(const enum CBLAS_ORDER Order, const enum CBLAS_TRANSPOSE TransA, const enum CBLAS_TRANSPOSE TransB, const int M, const int N, const int K, const float alpha, const float *A, const int lda, const float *B, const int ldb, const float beta, float *C, const int ldc);
+//
+void cblas_strsm(const enum CBLAS_ORDER Order, const enum CBLAS_SIDE Side, const enum CBLAS_UPLO Uplo, const enum CBLAS_TRANSPOSE TransA, const enum CBLAS_DIAG Diag, const int M, const int N, const float alpha, const float *A, const int lda, float *B, const int ldb);
+
+
+
+#endif // CBLAS_API
+
+
+
 #else // BLASFEO_API
 
 
 
 // BLAS 1
 //
-void blas_saxpy(int *n, float *alpha, float *x, int *incx, float *y, int *incy);
+void blasfeo_blas_saxpy(int *n, float *alpha, float *x, int *incx, float *y, int *incy);
 //
-float blas_sdot(int *n, float *x, int *incx, float *y, int *incy);
+float blasfeo_blas_sdot(int *n, float *x, int *incx, float *y, int *incy);
 
 // BLAS 3
 //
-void blas_sgemm(char *ta, char *tb, int *m, int *n, int *k, float *alpha, float *A, int *lda, float *B, int *ldb, float *beta, float *C, int *ldc);
+void blasfeo_blas_sgemm(char *ta, char *tb, int *m, int *n, int *k, float *alpha, float *A, int *lda, float *B, int *ldb, float *beta, float *C, int *ldc);
 //
-void blas_strsm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, float *alpha, float *A, int *lda, float *B, int *ldb);
+void blasfeo_blas_strsm(char *side, char *uplo, char *transa, char *diag, int *m, int *n, float *alpha, float *A, int *lda, float *B, int *ldb);
 
 
 
 // LAPACK
 //
-void blas_spotrf(char *uplo, int *m, float *A, int *lda, int *info);
+void blasfeo_lapack_spotrf(char *uplo, int *m, float *A, int *lda, int *info);
+
+
+
+#ifdef CBLAS_API
+
+
+
+// CBLAS 1
+//
+void blasfeo_cblas_saxpy(const int N, const float alpha, const float *X, const int incX, float *Y, const int incY);
+
+// CBLAS 3
+//
+void blasfeo_cblas_sgemm(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_TRANSPOSE TransA, const enum BLASFEO_CBLAS_TRANSPOSE TransB, const int M, const int N, const int K, const float alpha, const float *A, const int lda, const float *B, const int ldb, const float beta, float *C, const int ldc);
+//
+void blasfeo_cblas_strsm(const enum BLASFEO_CBLAS_ORDER Order, const enum BLASFEO_CBLAS_SIDE Side, const enum BLASFEO_CBLAS_UPLO Uplo, const enum BLASFEO_CBLAS_TRANSPOSE TransA, const enum BLASFEO_CBLAS_DIAG Diag, const int M, const int N, const float alpha, const float *A, const int lda, float *B, const int ldb);
+
+
+
+#endif // CBLAS_API
 
 
 

third_party/acados/include/blasfeo/include/blasfeo_s_blasfeo_api.h

@@ -115,6 +115,10 @@ void blasfeo_sgemv_nt(int m, int n, float alpha_n, float alpha_t, struct blasfeo
 // z <= beta * y + alpha * A * x, where A is symmetric and only the lower triangular patr of A is accessed
 void blasfeo_ssymv_l(int m, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
 void blasfeo_ssymv_l_mn(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
+// z <= beta * y + alpha * A * x, where A is symmetric and only the upper triangular patr of A is accessed
+void blasfeo_ssymv_u(int m, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
+// D = C + alpha * x * y^T
+void blasfeo_sger(int m, int n, float alpha, struct blasfeo_svec *sx, int xi, struct blasfeo_svec *sy, int yi, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
 
 // diagonal
 
@@ -182,6 +186,14 @@ void blasfeo_strsm_runu(int m, int n, float alpha, struct blasfeo_smat *sA, int
 void blasfeo_strsm_rutn(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, struct blasfeo_smat *sD, int di, int dj);
 // D <= alpha * B * A^{-T} , with A upper triangular with unit diagonal
 void blasfeo_strsm_rutu(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T; C, D lower triangular
+void blasfeo_ssyr2k_ln(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A; C, D lower triangular
+void blasfeo_ssyr2k_lt(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T; C, D upper triangular
+void blasfeo_ssyr2k_un(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A; C, D upper triangular
+void blasfeo_ssyr2k_ut(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
 
 // diagonal
 

third_party/acados/include/blasfeo/include/blasfeo_s_blasfeo_ref_api.h

@@ -115,6 +115,10 @@ void blasfeo_ref_sgemv_nt(int m, int n, float alpha_n, float alpha_t, struct bla
 // z <= beta * y + alpha * A * x, where A is symmetric and only the lower triangular patr of A is accessed
 void blasfeo_ref_ssymv_l(int m, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
 void blasfeo_ref_ssymv_l_mn(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
+// z <= beta * y + alpha * A * x, where A is symmetric and only the upper triangular patr of A is accessed
+void blasfeo_ref_ssymv_u(int m, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_svec *sx, int xi, float beta, struct blasfeo_svec *sy, int yi, struct blasfeo_svec *sz, int zi);
+// D = C + alpha * x * y^T
+void blasfeo_ref_sger(int m, int n, float alpha, struct blasfeo_svec *sx, int xi, struct blasfeo_svec *sy, int yi, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
 
 // diagonal
 
@@ -182,6 +186,14 @@ void blasfeo_ref_strsm_runu(int m, int n, float alpha, struct blasfeo_smat *sA,
 void blasfeo_ref_strsm_rutn(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, struct blasfeo_smat *sD, int di, int dj);
 // D <= alpha * B * A^{-T} , with A upper triangular with unit diagonal
 void blasfeo_ref_strsm_rutu(int m, int n, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T ; C, D lower triangular
+void blasfeo_ref_ssyr2k_ln(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A ; C, D lower triangular
+void blasfeo_ref_ssyr2k_lt(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A * B^T + alpha * B * A^T ; C, D upper triangular
+void blasfeo_ref_ssyr2k_un(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
+// D <= beta * C + alpha * A^T * B + alpha * B^T * A ; C, D upper triangular
+void blasfeo_ref_ssyr2k_ut(int m, int k, float alpha, struct blasfeo_smat *sA, int ai, int aj, struct blasfeo_smat *sB, int bi, int bj, float beta, struct blasfeo_smat *sC, int ci, int cj, struct blasfeo_smat *sD, int di, int dj);
 
 // diagonal
 

third_party/acados/include/blasfeo/include/blasfeo_target.h

@@ -1,13 +1,13 @@
-#ifndef TARGET_X64_INTEL_HASWELL
+#ifndef TARGET_ARMV8A_ARM_CORTEX_A57
-#define TARGET_X64_INTEL_HASWELL
+#define TARGET_ARMV8A_ARM_CORTEX_A57
 #endif
 
 #ifndef TARGET_NEED_FEATURE_AVX2
-#define TARGET_NEED_FEATURE_AVX2 1
+/* #undef TARGET_NEED_FEATURE_AVX2 */
 #endif
 
 #ifndef TARGET_NEED_FEATURE_FMA
-#define TARGET_NEED_FEATURE_FMA 1
+/* #undef TARGET_NEED_FEATURE_FMA */
 #endif
 
 #ifndef TARGET_NEED_FEATURE_SSE3
@@ -27,11 +27,11 @@
 #endif
 
 #ifndef TARGET_NEED_FEATURE_VFPv4
-/* #undef TARGET_NEED_FEATURE_VFPv4 */
+#define TARGET_NEED_FEATURE_VFPv4 1
 #endif
 
 #ifndef TARGET_NEED_FEATURE_NEONv2
-/* #undef TARGET_NEED_FEATURE_NEONv2 */
+#define TARGET_NEED_FEATURE_NEONv2 1
 #endif
 
 #ifndef LA_HIGH_PERFORMANCE

third_party/acados/include/blasfeo/include/d_blas.h

@@ -61,6 +61,7 @@ void dgemm_(char *ta, char *tb, int *m, int *n, int *k, double *alpha, double *A
 void dsyrk_(char *uplo, char *trans, int *n, int *k, double *alpha, double *A, int *lda, double *beta, double *C, int *ldc);
 void dtrmm_(char *side, char *uplo, char *trans, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
 void dtrsm_(char *side, char *uplo, char *trans, char *diag, int *m, int *n, double *alpha, double *A, int *lda, double *B, int *ldb);
+void dsyr2k_(char *uplo, char *trans, int *n, int *k, double *alpha, double *A, int *lda, double *B, int *ldb, double *beta, double *C, int *ldc);
 
 // lapack
 void dpotrf_(char *uplo, int *m, double *A, int *lda, int *info);

third_party/acados/include/blasfeo/include/s_blas.h

@@ -61,6 +61,7 @@ void sgemm_(char *ta, char *tb, int *m, int *n, int *k, float *alpha, float *A,
 void ssyrk_(char *uplo, char *trans, int *n, int *k, float *alpha, float *A, int *lda, float *beta, float *C, int *ldc);
 void strmm_(char *side, char *uplo, char *transa, char *diag, int *m, int *n, float *alpha, float *A, int *lda, float *B, int *ldb);
 void strsm_(char *side, char *uplo, char *transa, char *diag, int *m, int *n, float *alpha, float *A, int *lda, float *B, int *ldb);
+void ssyr2k_(char *uplo, char *trans, int *n, int *k, float *alpha, float *A, int *lda, float *B, int *ldb, float *beta, float *C, int *ldc);
 
 // lapack
 void spotrf_(char *uplo, int *m, float *A, int *lda, int *info);