Automatically generate functions with multiple returns

Author: mdboom

cuda_bindings/cuda/bindings/_v2/nvrtc.pyx

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 #
 # This code was automatically generated across versions from 12.9.0 to 13.3.0. Do not modify it directly.
-# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=a36c7e54cf29166832dd9aebc1fa71cc3649498794a2846e707396419caebe10
+# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=81a8413522bab04b3eb89373f90a9879758fdeeeb8141425dc27c4504df84b78
 
 
 # <<<< PREAMBLE CONTENT >>>>
@@ -657,6 +657,8 @@ cpdef tuple version():
 cpdef int get_num_supported_archs() except? -1:
     """nvrtcGetNumSupportedArchs sets the output parameter ``num_archs`` with the number of architectures supported by NVRTC. This can then be used to pass an array to ``nvrtcGetSupportedArchs`` to get the supported architectures.
 
+    see ``nvrtcGetSupportedArchs``.
+
     Returns:
         int: number of supported architectures.
 
@@ -672,6 +674,8 @@ cpdef int get_num_supported_archs() except? -1:
 cpdef object get_supported_archs():
     """nvrtcGetSupportedArchs populates the array passed via the output parameter ``supported_archs`` with the architectures supported by NVRTC. The array is sorted in the ascending order. The size of the array to be passed can be determined using ``nvrtcGetNumSupportedArchs``.
 
+    see ``nvrtcGetNumSupportedArchs``.
+
     Returns:
         int: sorted array of supported architectures.
 
@@ -681,13 +685,19 @@ cpdef object get_supported_archs():
     with nogil:
         __status__ = nvrtcGetNumSupportedArchs(&numArchs)
     check_status(__status__)
+    cdef _cyb_view.array supported_archs_alloc
+    cdef int *supported_archs_ptr
     if numArchs == 0:
-        return _cyb_view.array(shape=(1,), itemsize=sizeof(int), format="i", mode="c")[:0]
-    cdef _cyb_view.array supported_archs = _cyb_view.array(shape=(numArchs,), itemsize=sizeof(int), format="i", mode="c")
-    cdef int *supported_archs_ptr = <int *>(supported_archs.data)
-    with nogil:
-        __status__ = nvrtcGetSupportedArchs(supported_archs_ptr)
-    check_status(__status__)
+        supported_archs_alloc = _cyb_view.array(shape=(1,), itemsize=sizeof(int), format="i", mode="c")
+        supported_archs = supported_archs_alloc[:0]
+    else:
+        supported_archs_alloc = _cyb_view.array(shape=(numArchs,), itemsize=sizeof(int), format="i", mode="c")
+        supported_archs = supported_archs_alloc
+    supported_archs_ptr = <int *>(supported_archs_alloc.data)
+    if numArchs != 0:
+        with nogil:
+            __status__ = nvrtcGetSupportedArchs(supported_archs_ptr)
+        check_status(__status__)
     return supported_archs
 
 
@@ -739,13 +749,12 @@ cpdef bytes get_ptx(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetPTXSize(<Program>prog, &ptxSizeRet)
     check_status(__status__)
-    if ptxSizeRet == 0:
-        return b""
     cdef bytes _ptx_ = bytes(ptxSizeRet)
     cdef char* ptx = _ptx_
-    with nogil:
-        __status__ = nvrtcGetPTX(<Program>prog, ptx)
-    check_status(__status__)
+    if ptxSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetPTX(<Program>prog, ptx)
+        check_status(__status__)
     return _ptx_
 
 
@@ -782,13 +791,12 @@ cpdef bytes get_cubin(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetCUBINSize(<Program>prog, &cubinSizeRet)
     check_status(__status__)
-    if cubinSizeRet == 0:
-        return b""
     cdef bytes _cubin_ = bytes(cubinSizeRet)
     cdef char* cubin = _cubin_
-    with nogil:
-        __status__ = nvrtcGetCUBIN(<Program>prog, cubin)
-    check_status(__status__)
+    if cubinSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetCUBIN(<Program>prog, cubin)
+        check_status(__status__)
     return _cubin_
 
 
@@ -825,13 +833,12 @@ cpdef bytes get_ltoir(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetLTOIRSize(<Program>prog, &LTOIRSizeRet)
     check_status(__status__)
-    if LTOIRSizeRet == 0:
-        return b""
     cdef bytes _ltoir_ = bytes(LTOIRSizeRet)
     cdef char* ltoir = _ltoir_
-    with nogil:
-        __status__ = nvrtcGetLTOIR(<Program>prog, ltoir)
-    check_status(__status__)
+    if LTOIRSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetLTOIR(<Program>prog, ltoir)
+        check_status(__status__)
     return _ltoir_
 
 
@@ -868,19 +875,21 @@ cpdef bytes get_optix_ir(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetOptiXIRSize(<Program>prog, &optixirSizeRet)
     check_status(__status__)
-    if optixirSizeRet == 0:
-        return b""
     cdef bytes _optixir_ = bytes(optixirSizeRet)
     cdef char* optixir = _optixir_
-    with nogil:
-        __status__ = nvrtcGetOptiXIR(<Program>prog, optixir)
-    check_status(__status__)
+    if optixirSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetOptiXIR(<Program>prog, optixir)
+        check_status(__status__)
     return _optixir_
 
 
 cpdef size_t get_program_log_size(intptr_t prog) except? 0:
     """nvrtcGetProgramLogSize sets ``log_size_ret`` with the size of the log generated by the previous compilation of ``prog`` (including the trailing ``NULL``).
 
+    Note that compilation log may be generated with warnings and informative
+    messages, even when the compilation of ``prog`` succeeds.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
 
@@ -912,19 +921,21 @@ cpdef bytes get_program_log(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetProgramLogSize(<Program>prog, &logSizeRet)
     check_status(__status__)
-    if logSizeRet == 0:
-        return b""
     cdef bytes _log_ = bytes(logSizeRet)
     cdef char* log = _log_
-    with nogil:
-        __status__ = nvrtcGetProgramLog(<Program>prog, log)
-    check_status(__status__)
+    if logSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetProgramLog(<Program>prog, log)
+        check_status(__status__)
     return _log_
 
 
 cpdef add_name_expression(intptr_t prog, name_expression):
     """nvrtcAddNameExpression notes the given name expression denoting the address of a global function or device/__constant__ variable.
 
+    The identical name expression string must be provided on a subsequent call
+    to nvrtcGetLoweredName to extract the lowered name.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
         name_expression (str): constant expression denoting the
@@ -961,6 +972,10 @@ cpdef size_t get_pch_heap_size() except? 0:
 cpdef set_pch_heap_size(size_t size):
     """set the size of the PCH Heap.
 
+    The requested size may be rounded up to a platform dependent alignment
+    (e.g. page size). If the PCH Heap has already been allocated, the heap
+    memory will be freed and a new PCH Heap will be allocated.
+
     Args:
         size (size_t): requested size of the PCH Heap, in bytes.
 
@@ -974,6 +989,20 @@ cpdef set_pch_heap_size(size_t size):
 cpdef int get_pch_create_status(intptr_t prog) except? -1:
     """returns the PCH creation status.
 
+    NVRTC_SUCCESS indicates that the PCH was successfully created.
+    NVRTC_ERROR_NO_PCH_CREATE_ATTEMPTED indicates that no PCH creation was
+    attempted, either because PCH functionality was not requested during the
+    preceding nvrtcCompileProgram call, or automatic PCH processing was
+    requested, and compiler chose not to create a PCH file.
+    NVRTC_ERROR_PCH_CREATE_HEAP_EXHAUSTED indicates that a PCH file could
+    potentially have been created, but the compiler ran out space in the PCH
+    heap. In this scenario, the :func:`get_pch_heap_size_required` can be used
+    to query the required heap size, the heap can be reallocated for this size
+    with :func:`set_pch_heap_size` and PCH creation may be reattempted again
+    invoking :func:`compile_program` with a new NVRTC program instance.
+    NVRTC_ERROR_PCH_CREATE indicates that an error condition prevented the PCH
+    file from being created.
+
     Args:
         prog (intptr_t): CUDA Runtime Compilation program.
 
@@ -1037,13 +1066,12 @@ cpdef bytes get_tile_ir(intptr_t prog):
     with nogil:
         __status__ = nvrtcGetTileIRSize(<Program>prog, &TileIRSizeRet)
     check_status(__status__)
-    if TileIRSizeRet == 0:
-        return b""
     cdef bytes _tile_ir_ = bytes(TileIRSizeRet)
     cdef char* tile_ir = _tile_ir_
-    with nogil:
-        __status__ = nvrtcGetTileIR(<Program>prog, tile_ir)
-    check_status(__status__)
+    if TileIRSizeRet != 0:
+        with nogil:
+            __status__ = nvrtcGetTileIR(<Program>prog, tile_ir)
+        check_status(__status__)
     return _tile_ir_
 
 

cuda_bindings/cuda/bindings/nvml.pxd

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 #
 # This code was automatically generated across versions from 12.9.1 to 13.3.0. Do not modify it directly.
-# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=b6fe9a4efd0077f8c09ef4f826880ad0a54100455d4465953c4127d3de8c4d91
+# CYTHON-BINDINGS-GENERATED-DO-NOT-MODIFY-THIS-FILE: format=1; content-sha256=c45bc4c9772d2e6fe0163353f54a44655db866a025cc1adab756e44df3b39731
 
 
 
@@ -62,7 +62,6 @@ ctypedef nvmlUnrepairableMemoryStatus_v1_t UnrepairableMemoryStatus_v1
 ctypedef nvmlRusdSettings_v1_t RusdSettings_v1
 ctypedef nvmlPowerValue_v2_t PowerValue_v2
 ctypedef nvmlVgpuTypeMaxInstance_v1_t VgpuTypeMaxInstance_v1
-ctypedef nvmlVgpuProcessUtilizationSample_t VgpuProcessUtilizationSample
 ctypedef nvmlGpuFabricInfo_t GpuFabricInfo
 ctypedef nvmlSystemEventSetCreateRequest_v1_t SystemEventSetCreateRequest_v1
 ctypedef nvmlSystemEventSetFreeRequest_v1_t SystemEventSetFreeRequest_v1
@@ -163,13 +162,15 @@ cpdef int system_get_cuda_driver_version() except *
 cpdef int system_get_cuda_driver_version_v2() except 0
 cpdef str system_get_process_name(unsigned int pid)
 cpdef object system_get_hic_version()
+cpdef object system_get_topology_gpu_set(unsigned int cpu_number)
 cpdef unsigned int unit_get_count() except? 0
 cpdef intptr_t unit_get_handle_by_index(unsigned int index) except? 0
 cpdef object unit_get_unit_info(intptr_t unit)
 cpdef object unit_get_led_state(intptr_t unit)
 cpdef object unit_get_psu_info(intptr_t unit)
 cpdef unsigned int unit_get_temperature(intptr_t unit, unsigned int type) except? 0
 cpdef object unit_get_fan_speed_info(intptr_t unit)
+cpdef object unit_get_devices(intptr_t unit)
 cpdef unsigned int device_get_count_v2() except? 0
 cpdef object device_get_attributes_v2(intptr_t device)
 cpdef intptr_t device_get_handle_by_index_v2(unsigned int index) except? 0
@@ -189,6 +190,7 @@ cpdef device_set_cpu_affinity(intptr_t device)
 cpdef device_clear_cpu_affinity(intptr_t device)
 cpdef unsigned int device_get_numa_node_id(intptr_t device) except? 0
 cpdef int device_get_topology_common_ancestor(intptr_t device1, intptr_t device2) except? -1
+cpdef object device_get_topology_nearest_gpus(intptr_t device, int level)
 cpdef int device_get_p2p_status(intptr_t device1, intptr_t device2, int p2p_index) except? -1
 cpdef str device_get_uuid(intptr_t device)
 cpdef unsigned int device_get_minor_number(intptr_t device) except? 0
@@ -271,6 +273,7 @@ cpdef object device_get_graphics_running_processes_v3(intptr_t device)
 cpdef object device_get_mps_compute_running_processes_v3(intptr_t device)
 cpdef int device_on_same_board(intptr_t device1, intptr_t device2) except? 0
 cpdef int device_get_api_restriction(intptr_t device, int api_type) except? -1
+cpdef tuple device_get_samples(intptr_t device, int type, unsigned long long last_seen_time_stamp)
 cpdef object device_get_bar1_memory_info(intptr_t device)
 cpdef unsigned int device_get_irq_num(intptr_t device) except? 0
 cpdef unsigned int device_get_num_gpu_cores(intptr_t device) except? 0
@@ -297,6 +300,7 @@ cpdef object device_get_accounting_stats(intptr_t device, unsigned int pid)
 cpdef object device_get_accounting_pids(intptr_t device)
 cpdef unsigned int device_get_accounting_buffer_size(intptr_t device) except? 0
 cpdef object device_get_retired_pages(intptr_t device, int cause)
+cpdef tuple device_get_retired_pages_v2(intptr_t device, int cause)
 cpdef int device_get_retired_pages_pending_status(intptr_t device) except? -1
 cpdef tuple device_get_remapped_rows(intptr_t device)
 cpdef object device_get_row_remapper_histogram(intptr_t device)
@@ -352,6 +356,8 @@ cpdef device_set_vgpu_capabilities(intptr_t device, int capability, int state)
 cpdef object device_get_grid_licensable_features_v4(intptr_t device)
 cpdef unsigned int get_vgpu_driver_capabilities(int capability) except? 0
 cpdef unsigned int device_get_vgpu_capabilities(intptr_t device, int capability) except? 0
+cpdef object device_get_supported_vgpus(intptr_t device)
+cpdef object device_get_creatable_vgpus(intptr_t device)
 cpdef str vgpu_type_get_class(unsigned int vgpu_type_id)
 cpdef unsigned int vgpu_type_get_gpu_instance_profile_id(unsigned int vgpu_type_id) except? 0
 cpdef tuple vgpu_type_get_device_id(unsigned int vgpu_type_id)
@@ -363,6 +369,8 @@ cpdef unsigned int vgpu_type_get_frame_rate_limit(unsigned int vgpu_type_id) exc
 cpdef unsigned int vgpu_type_get_max_instances(intptr_t device, unsigned int vgpu_type_id) except? 0
 cpdef unsigned int vgpu_type_get_max_instances_per_vm(unsigned int vgpu_type_id) except? 0
 cpdef object vgpu_type_get_bar1_info(unsigned int vgpu_type_id)
+cpdef object device_get_active_vgpus(intptr_t device)
+cpdef tuple vgpu_instance_get_vm_id(unsigned int vgpu_instance)
 cpdef str vgpu_instance_get_uuid(unsigned int vgpu_instance)
 cpdef str vgpu_instance_get_vm_driver_version(unsigned int vgpu_instance)
 cpdef unsigned long long vgpu_instance_get_fb_usage(unsigned int vgpu_instance) except? 0
@@ -388,8 +396,10 @@ cpdef object device_get_vgpu_scheduler_log(intptr_t device)
 cpdef object device_get_vgpu_scheduler_state(intptr_t device)
 cpdef object device_get_vgpu_scheduler_capabilities(intptr_t device)
 cpdef device_set_vgpu_scheduler_state(intptr_t device, intptr_t p_scheduler_state)
+cpdef tuple get_vgpu_version()
 cpdef set_vgpu_version(intptr_t vgpu_version)
-cpdef tuple device_get_vgpu_process_utilization(intptr_t device, unsigned long long last_seen_time_stamp)
+cpdef tuple device_get_vgpu_utilization(intptr_t device, unsigned long long last_seen_time_stamp)
+cpdef object device_get_vgpu_process_utilization(intptr_t device, unsigned long long last_seen_time_stamp)
 cpdef int vgpu_instance_get_accounting_mode(unsigned int vgpu_instance) except? -1
 cpdef object vgpu_instance_get_accounting_pids(unsigned int vgpu_instance)
 cpdef object vgpu_instance_get_accounting_stats(unsigned int vgpu_instance, unsigned int pid)
@@ -404,6 +414,7 @@ cpdef unsigned int device_get_gpu_instance_remaining_capacity(intptr_t device, u
 cpdef intptr_t device_create_gpu_instance(intptr_t device, unsigned int profile_id) except? 0
 cpdef intptr_t device_create_gpu_instance_with_placement(intptr_t device, unsigned int profile_id, intptr_t placement) except? 0
 cpdef gpu_instance_destroy(intptr_t gpu_instance)
+cpdef object device_get_gpu_instances(intptr_t device, unsigned int profile_id)
 cpdef intptr_t device_get_gpu_instance_by_id(intptr_t device, unsigned int id) except? 0
 cpdef object gpu_instance_get_info(intptr_t gpu_instance)
 cpdef object gpu_instance_get_compute_instance_profile_info_v(intptr_t gpu_instance, unsigned int profile, unsigned int eng_profile)
@@ -412,6 +423,7 @@ cpdef object gpu_instance_get_compute_instance_possible_placements(intptr_t gpu_
 cpdef intptr_t gpu_instance_create_compute_instance(intptr_t gpu_instance, unsigned int profile_id) except? 0
 cpdef intptr_t gpu_instance_create_compute_instance_with_placement(intptr_t gpu_instance, unsigned int profile_id, intptr_t placement) except? 0
 cpdef compute_instance_destroy(intptr_t compute_instance)
+cpdef object gpu_instance_get_compute_instances(intptr_t gpu_instance, unsigned int profile_id)
 cpdef intptr_t gpu_instance_get_compute_instance_by_id(intptr_t gpu_instance, unsigned int id) except? 0
 cpdef object compute_instance_get_info_v2(intptr_t compute_instance)
 cpdef unsigned int device_is_mig_device_handle(intptr_t device) except? 0