warp-lang 1.8.1__py3-none-macosx_10_13_universal2.whl → 1.9.0__py3-none-macosx_10_13_universal2.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of warp-lang might be problematic. Click here for more details.

Files changed (134) hide show
  1. warp/__init__.py +282 -103
  2. warp/__init__.pyi +482 -110
  3. warp/bin/libwarp-clang.dylib +0 -0
  4. warp/bin/libwarp.dylib +0 -0
  5. warp/build.py +93 -30
  6. warp/build_dll.py +47 -67
  7. warp/builtins.py +955 -137
  8. warp/codegen.py +312 -206
  9. warp/config.py +1 -1
  10. warp/context.py +1249 -784
  11. warp/examples/core/example_marching_cubes.py +1 -0
  12. warp/examples/core/example_render_opengl.py +100 -3
  13. warp/examples/fem/example_apic_fluid.py +98 -52
  14. warp/examples/fem/example_convection_diffusion_dg.py +25 -4
  15. warp/examples/fem/example_diffusion_mgpu.py +8 -3
  16. warp/examples/fem/utils.py +68 -22
  17. warp/fabric.py +1 -1
  18. warp/fem/cache.py +27 -19
  19. warp/fem/domain.py +2 -2
  20. warp/fem/field/nodal_field.py +2 -2
  21. warp/fem/field/virtual.py +264 -166
  22. warp/fem/geometry/geometry.py +5 -5
  23. warp/fem/integrate.py +129 -51
  24. warp/fem/space/restriction.py +4 -0
  25. warp/fem/space/shape/tet_shape_function.py +3 -10
  26. warp/jax_experimental/custom_call.py +1 -1
  27. warp/jax_experimental/ffi.py +2 -1
  28. warp/marching_cubes.py +708 -0
  29. warp/native/array.h +99 -4
  30. warp/native/builtin.h +82 -5
  31. warp/native/bvh.cpp +64 -28
  32. warp/native/bvh.cu +58 -58
  33. warp/native/bvh.h +2 -2
  34. warp/native/clang/clang.cpp +7 -7
  35. warp/native/coloring.cpp +8 -2
  36. warp/native/crt.cpp +2 -2
  37. warp/native/crt.h +3 -5
  38. warp/native/cuda_util.cpp +41 -10
  39. warp/native/cuda_util.h +10 -4
  40. warp/native/exports.h +1842 -1908
  41. warp/native/fabric.h +2 -1
  42. warp/native/hashgrid.cpp +37 -37
  43. warp/native/hashgrid.cu +2 -2
  44. warp/native/initializer_array.h +1 -1
  45. warp/native/intersect.h +2 -2
  46. warp/native/mat.h +1910 -116
  47. warp/native/mathdx.cpp +43 -43
  48. warp/native/mesh.cpp +24 -24
  49. warp/native/mesh.cu +26 -26
  50. warp/native/mesh.h +4 -2
  51. warp/native/nanovdb/GridHandle.h +179 -12
  52. warp/native/nanovdb/HostBuffer.h +8 -7
  53. warp/native/nanovdb/NanoVDB.h +517 -895
  54. warp/native/nanovdb/NodeManager.h +323 -0
  55. warp/native/nanovdb/PNanoVDB.h +2 -2
  56. warp/native/quat.h +331 -14
  57. warp/native/range.h +7 -1
  58. warp/native/reduce.cpp +10 -10
  59. warp/native/reduce.cu +13 -14
  60. warp/native/runlength_encode.cpp +2 -2
  61. warp/native/runlength_encode.cu +5 -5
  62. warp/native/scan.cpp +3 -3
  63. warp/native/scan.cu +4 -4
  64. warp/native/sort.cpp +10 -10
  65. warp/native/sort.cu +22 -22
  66. warp/native/sparse.cpp +8 -8
  67. warp/native/sparse.cu +13 -13
  68. warp/native/spatial.h +366 -17
  69. warp/native/temp_buffer.h +2 -2
  70. warp/native/tile.h +283 -69
  71. warp/native/vec.h +381 -14
  72. warp/native/volume.cpp +54 -54
  73. warp/native/volume.cu +1 -1
  74. warp/native/volume.h +2 -1
  75. warp/native/volume_builder.cu +30 -37
  76. warp/native/warp.cpp +150 -149
  77. warp/native/warp.cu +323 -192
  78. warp/native/warp.h +227 -226
  79. warp/optim/linear.py +736 -271
  80. warp/render/imgui_manager.py +289 -0
  81. warp/render/render_opengl.py +85 -6
  82. warp/sim/graph_coloring.py +2 -2
  83. warp/sparse.py +558 -175
  84. warp/tests/aux_test_module_aot.py +7 -0
  85. warp/tests/cuda/test_async.py +3 -3
  86. warp/tests/cuda/test_conditional_captures.py +101 -0
  87. warp/tests/geometry/test_marching_cubes.py +233 -12
  88. warp/tests/sim/test_coloring.py +6 -6
  89. warp/tests/test_array.py +56 -5
  90. warp/tests/test_codegen.py +3 -2
  91. warp/tests/test_context.py +8 -15
  92. warp/tests/test_enum.py +136 -0
  93. warp/tests/test_examples.py +2 -2
  94. warp/tests/test_fem.py +45 -2
  95. warp/tests/test_fixedarray.py +229 -0
  96. warp/tests/test_func.py +18 -15
  97. warp/tests/test_future_annotations.py +7 -5
  98. warp/tests/test_linear_solvers.py +30 -0
  99. warp/tests/test_map.py +1 -1
  100. warp/tests/test_mat.py +1518 -378
  101. warp/tests/test_mat_assign_copy.py +178 -0
  102. warp/tests/test_mat_constructors.py +574 -0
  103. warp/tests/test_module_aot.py +287 -0
  104. warp/tests/test_print.py +69 -0
  105. warp/tests/test_quat.py +140 -34
  106. warp/tests/test_quat_assign_copy.py +145 -0
  107. warp/tests/test_reload.py +2 -1
  108. warp/tests/test_sparse.py +71 -0
  109. warp/tests/test_spatial.py +140 -34
  110. warp/tests/test_spatial_assign_copy.py +160 -0
  111. warp/tests/test_struct.py +43 -3
  112. warp/tests/test_types.py +0 -20
  113. warp/tests/test_vec.py +179 -34
  114. warp/tests/test_vec_assign_copy.py +143 -0
  115. warp/tests/tile/test_tile.py +184 -18
  116. warp/tests/tile/test_tile_cholesky.py +605 -0
  117. warp/tests/tile/test_tile_load.py +169 -0
  118. warp/tests/tile/test_tile_mathdx.py +2 -558
  119. warp/tests/tile/test_tile_matmul.py +1 -1
  120. warp/tests/tile/test_tile_mlp.py +1 -1
  121. warp/tests/tile/test_tile_shared_memory.py +5 -5
  122. warp/tests/unittest_suites.py +6 -0
  123. warp/tests/walkthrough_debug.py +1 -1
  124. warp/thirdparty/unittest_parallel.py +108 -9
  125. warp/types.py +554 -264
  126. warp/utils.py +68 -86
  127. {warp_lang-1.8.1.dist-info → warp_lang-1.9.0.dist-info}/METADATA +28 -65
  128. {warp_lang-1.8.1.dist-info → warp_lang-1.9.0.dist-info}/RECORD +131 -121
  129. warp/native/marching.cpp +0 -19
  130. warp/native/marching.cu +0 -514
  131. warp/native/marching.h +0 -19
  132. {warp_lang-1.8.1.dist-info → warp_lang-1.9.0.dist-info}/WHEEL +0 -0
  133. {warp_lang-1.8.1.dist-info → warp_lang-1.9.0.dist-info}/licenses/LICENSE.md +0 -0
  134. {warp_lang-1.8.1.dist-info → warp_lang-1.9.0.dist-info}/top_level.txt +0 -0
warp/native/warp.h CHANGED
@@ -28,157 +28,157 @@ struct timing_result_t;
28
28
  // this is the core runtime API exposed on the DLL level
29
29
  extern "C"
30
30
  {
31
- WP_API int init();
32
- //WP_API void shutdown();
31
+ WP_API int wp_init();
32
+ //WP_API void wp_shutdown();
33
33
 
34
34
  // get error message from C++
35
- WP_API const char* get_error_string();
35
+ WP_API const char* wp_get_error_string();
36
36
 
37
37
  // allow disabling error output, which is handy during tests that expect failure
38
- WP_API void set_error_output_enabled(int enable);
39
- WP_API int is_error_output_enabled();
38
+ WP_API void wp_set_error_output_enabled(int enable);
39
+ WP_API int wp_is_error_output_enabled();
40
40
 
41
41
  // whether Warp was compiled with CUDA support
42
- WP_API int is_cuda_enabled();
42
+ WP_API int wp_is_cuda_enabled();
43
43
  // whether Warp was compiled with enhanced CUDA compatibility
44
- WP_API int is_cuda_compatibility_enabled();
44
+ WP_API int wp_is_cuda_compatibility_enabled();
45
45
  // whether Warp was compiled with MathDx support
46
- WP_API int is_mathdx_enabled();
46
+ WP_API int wp_is_mathdx_enabled();
47
47
  // whether Warp was compiled with debug support
48
- WP_API int is_debug_enabled();
49
-
50
- WP_API uint16_t float_to_half_bits(float x);
51
- WP_API float half_bits_to_float(uint16_t u);
52
-
53
- WP_API void* alloc_host(size_t s);
54
- WP_API void* alloc_pinned(size_t s);
55
- WP_API void* alloc_device(void* context, size_t s); // uses cudaMallocAsync() if supported, cudaMalloc() otherwise
56
- WP_API void* alloc_device_default(void* context, size_t s); // uses cudaMalloc()
57
- WP_API void* alloc_device_async(void* context, size_t s); // uses cudaMallocAsync()
58
-
59
- WP_API void free_host(void* ptr);
60
- WP_API void free_pinned(void* ptr);
61
- WP_API void free_device(void* context, void* ptr); // uses cudaFreeAsync() if supported, cudaFree() otherwise
62
- WP_API void free_device_default(void* context, void* ptr); // uses cudaFree()
63
- WP_API void free_device_async(void* context, void* ptr); // uses cudaFreeAsync()
64
-
65
- WP_API bool memcpy_h2h(void* dest, void* src, size_t n);
66
- WP_API bool memcpy_h2d(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
67
- WP_API bool memcpy_d2h(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
68
- WP_API bool memcpy_d2d(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
69
- WP_API bool memcpy_p2p(void* dst_context, void* dst, void* src_context, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
70
-
71
- WP_API void memset_host(void* dest, int value, size_t n);
72
- WP_API void memset_device(void* context, void* dest, int value, size_t n);
48
+ WP_API int wp_is_debug_enabled();
49
+
50
+ WP_API uint16_t wp_float_to_half_bits(float x);
51
+ WP_API float wp_half_bits_to_float(uint16_t u);
52
+
53
+ WP_API void* wp_alloc_host(size_t s);
54
+ WP_API void* wp_alloc_pinned(size_t s);
55
+ WP_API void* wp_alloc_device(void* context, size_t s); // uses cudaMallocAsync() if supported, cudaMalloc() otherwise
56
+ WP_API void* wp_alloc_device_default(void* context, size_t s); // uses cudaMalloc()
57
+ WP_API void* wp_alloc_device_async(void* context, size_t s); // uses cudaMallocAsync()
58
+
59
+ WP_API void wp_free_host(void* ptr);
60
+ WP_API void wp_free_pinned(void* ptr);
61
+ WP_API void wp_free_device(void* context, void* ptr); // uses cudaFreeAsync() if supported, cudaFree() otherwise
62
+ WP_API void wp_free_device_default(void* context, void* ptr); // uses cudaFree()
63
+ WP_API void wp_free_device_async(void* context, void* ptr); // uses cudaFreeAsync()
64
+
65
+ WP_API bool wp_memcpy_h2h(void* dest, void* src, size_t n);
66
+ WP_API bool wp_memcpy_h2d(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
67
+ WP_API bool wp_memcpy_d2h(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
68
+ WP_API bool wp_memcpy_d2d(void* context, void* dest, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
69
+ WP_API bool wp_memcpy_p2p(void* dst_context, void* dst, void* src_context, void* src, size_t n, void* stream=WP_CURRENT_STREAM);
70
+
71
+ WP_API void wp_memset_host(void* dest, int value, size_t n);
72
+ WP_API void wp_memset_device(void* context, void* dest, int value, size_t n);
73
73
 
74
74
  // takes srcsize bytes starting at src and repeats them n times at dst (writes srcsize * n bytes in total):
75
- WP_API void memtile_host(void* dest, const void* src, size_t srcsize, size_t n);
76
- WP_API void memtile_device(void* context, void* dest, const void* src, size_t srcsize, size_t n);
75
+ WP_API void wp_memtile_host(void* dest, const void* src, size_t srcsize, size_t n);
76
+ WP_API void wp_memtile_device(void* context, void* dest, const void* src, size_t srcsize, size_t n);
77
77
 
78
- WP_API uint64_t bvh_create_host(wp::vec3* lowers, wp::vec3* uppers, int num_items, int constructor_type);
79
- WP_API void bvh_destroy_host(uint64_t id);
80
- WP_API void bvh_refit_host(uint64_t id);
78
+ WP_API uint64_t wp_bvh_create_host(wp::vec3* lowers, wp::vec3* uppers, int num_items, int constructor_type);
79
+ WP_API void wp_bvh_destroy_host(uint64_t id);
80
+ WP_API void wp_bvh_refit_host(uint64_t id);
81
81
 
82
- WP_API uint64_t bvh_create_device(void* context, wp::vec3* lowers, wp::vec3* uppers, int num_items, int constructor_type);
83
- WP_API void bvh_destroy_device(uint64_t id);
84
- WP_API void bvh_refit_device(uint64_t id);
82
+ WP_API uint64_t wp_bvh_create_device(void* context, wp::vec3* lowers, wp::vec3* uppers, int num_items, int constructor_type);
83
+ WP_API void wp_bvh_destroy_device(uint64_t id);
84
+ WP_API void wp_bvh_refit_device(uint64_t id);
85
85
 
86
86
  // create a user-accessible copy of the mesh, it is the
87
87
  // users responsibility to keep-alive the points/tris data for the duration of the mesh lifetime
88
- WP_API uint64_t mesh_create_host(wp::array_t<wp::vec3> points, wp::array_t<wp::vec3> velocities, wp::array_t<int> tris,
88
+ WP_API uint64_t wp_mesh_create_host(wp::array_t<wp::vec3> points, wp::array_t<wp::vec3> velocities, wp::array_t<int> tris,
89
89
  int num_points, int num_tris, int support_winding_number, int constructor_type);
90
- WP_API void mesh_destroy_host(uint64_t id);
91
- WP_API void mesh_refit_host(uint64_t id);
90
+ WP_API void wp_mesh_destroy_host(uint64_t id);
91
+ WP_API void wp_mesh_refit_host(uint64_t id);
92
92
 
93
- WP_API uint64_t mesh_create_device(void* context, wp::array_t<wp::vec3> points, wp::array_t<wp::vec3> velocities,
93
+ WP_API uint64_t wp_mesh_create_device(void* context, wp::array_t<wp::vec3> points, wp::array_t<wp::vec3> velocities,
94
94
  wp::array_t<int> tris, int num_points, int num_tris, int support_winding_number, int constructor_type);
95
- WP_API void mesh_destroy_device(uint64_t id);
96
- WP_API void mesh_refit_device(uint64_t id);
97
-
98
- WP_API void mesh_set_points_host(uint64_t id, wp::array_t<wp::vec3> points);
99
- WP_API void mesh_set_points_device(uint64_t id, wp::array_t<wp::vec3> points);
100
-
101
- WP_API void mesh_set_velocities_host(uint64_t id, wp::array_t<wp::vec3> velocities);
102
- WP_API void mesh_set_velocities_device(uint64_t id, wp::array_t<wp::vec3> velocities);
103
-
104
- WP_API uint64_t hash_grid_create_host(int dim_x, int dim_y, int dim_z);
105
- WP_API void hash_grid_reserve_host(uint64_t id, int num_points);
106
- WP_API void hash_grid_destroy_host(uint64_t id);
107
- WP_API void hash_grid_update_host(uint64_t id, float cell_width, const wp::array_t<wp::vec3>* points);
108
-
109
- WP_API uint64_t hash_grid_create_device(void* context, int dim_x, int dim_y, int dim_z);
110
- WP_API void hash_grid_reserve_device(uint64_t id, int num_points);
111
- WP_API void hash_grid_destroy_device(uint64_t id);
112
- WP_API void hash_grid_update_device(uint64_t id, float cell_width, const wp::array_t<wp::vec3>* points);
113
-
114
- WP_API uint64_t volume_create_host(void* buf, uint64_t size, bool copy, bool owner);
115
- WP_API void volume_get_tiles_host(uint64_t id, void* buf);
116
- WP_API void volume_get_voxels_host(uint64_t id, void* buf);
117
- WP_API void volume_destroy_host(uint64_t id);
118
-
119
- WP_API uint64_t volume_create_device(void* context, void* buf, uint64_t size, bool copy, bool owner);
120
- WP_API void volume_get_tiles_device(uint64_t id, void* buf);
121
- WP_API void volume_get_voxels_device(uint64_t id, void* buf);
122
- WP_API void volume_destroy_device(uint64_t id);
95
+ WP_API void wp_mesh_destroy_device(uint64_t id);
96
+ WP_API void wp_mesh_refit_device(uint64_t id);
97
+
98
+ WP_API void wp_mesh_set_points_host(uint64_t id, wp::array_t<wp::vec3> points);
99
+ WP_API void wp_mesh_set_points_device(uint64_t id, wp::array_t<wp::vec3> points);
100
+
101
+ WP_API void wp_mesh_set_velocities_host(uint64_t id, wp::array_t<wp::vec3> velocities);
102
+ WP_API void wp_mesh_set_velocities_device(uint64_t id, wp::array_t<wp::vec3> velocities);
103
+
104
+ WP_API uint64_t wp_hash_grid_create_host(int dim_x, int dim_y, int dim_z);
105
+ WP_API void wp_hash_grid_reserve_host(uint64_t id, int num_points);
106
+ WP_API void wp_hash_grid_destroy_host(uint64_t id);
107
+ WP_API void wp_hash_grid_update_host(uint64_t id, float cell_width, const wp::array_t<wp::vec3>* points);
108
+
109
+ WP_API uint64_t wp_hash_grid_create_device(void* context, int dim_x, int dim_y, int dim_z);
110
+ WP_API void wp_hash_grid_reserve_device(uint64_t id, int num_points);
111
+ WP_API void wp_hash_grid_destroy_device(uint64_t id);
112
+ WP_API void wp_hash_grid_update_device(uint64_t id, float cell_width, const wp::array_t<wp::vec3>* points);
113
+
114
+ WP_API uint64_t wp_volume_create_host(void* buf, uint64_t size, bool copy, bool owner);
115
+ WP_API void wp_volume_get_tiles_host(uint64_t id, void* buf);
116
+ WP_API void wp_volume_get_voxels_host(uint64_t id, void* buf);
117
+ WP_API void wp_volume_destroy_host(uint64_t id);
118
+
119
+ WP_API uint64_t wp_volume_create_device(void* context, void* buf, uint64_t size, bool copy, bool owner);
120
+ WP_API void wp_volume_get_tiles_device(uint64_t id, void* buf);
121
+ WP_API void wp_volume_get_voxels_device(uint64_t id, void* buf);
122
+ WP_API void wp_volume_destroy_device(uint64_t id);
123
123
 
124
- WP_API uint64_t volume_from_tiles_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space, const void* bg_value, uint32_t bg_value_size, const char* bg_value_type);
125
- WP_API uint64_t volume_index_from_tiles_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space);
126
- WP_API uint64_t volume_from_active_voxels_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space);
127
-
128
- WP_API void volume_get_buffer_info(uint64_t id, void** buf, uint64_t* size);
129
- WP_API void volume_get_voxel_size(uint64_t id, float* dx, float* dy, float* dz);
130
- WP_API void volume_get_tile_and_voxel_count(uint64_t id, uint32_t& tile_count, uint64_t& voxel_count);
131
- WP_API const char* volume_get_grid_info(uint64_t id, uint64_t *grid_size, uint32_t *grid_index, uint32_t *grid_count, float translation[3], float transform[9], char type_str[16]);
132
- WP_API uint32_t volume_get_blind_data_count(uint64_t id);
133
- WP_API const char* volume_get_blind_data_info(uint64_t id, uint32_t data_index, void** buf, uint64_t* value_count, uint32_t* value_size, char type_str[16]);
124
+ WP_API uint64_t wp_volume_from_tiles_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space, const void* bg_value, uint32_t bg_value_size, const char* bg_value_type);
125
+ WP_API uint64_t wp_volume_index_from_tiles_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space);
126
+ WP_API uint64_t wp_volume_from_active_voxels_device(void* context, void* points, int num_points, float transform[9], float translation[3], bool points_in_world_space);
127
+
128
+ WP_API void wp_volume_get_buffer_info(uint64_t id, void** buf, uint64_t* size);
129
+ WP_API void wp_volume_get_voxel_size(uint64_t id, float* dx, float* dy, float* dz);
130
+ WP_API void wp_volume_get_tile_and_voxel_count(uint64_t id, uint32_t& tile_count, uint64_t& voxel_count);
131
+ WP_API const char* wp_volume_get_grid_info(uint64_t id, uint64_t *grid_size, uint32_t *grid_index, uint32_t *grid_count, float translation[3], float transform[9], char type_str[16]);
132
+ WP_API uint32_t wp_volume_get_blind_data_count(uint64_t id);
133
+ WP_API const char* wp_volume_get_blind_data_info(uint64_t id, uint32_t data_index, void** buf, uint64_t* value_count, uint32_t* value_size, char type_str[16]);
134
134
 
135
- WP_API uint64_t marching_cubes_create_device(void* context);
136
- WP_API void marching_cubes_destroy_device(uint64_t id);
137
- WP_API int marching_cubes_surface_device(uint64_t id, const float* field, int nx, int ny, int nz, float threshold, wp::vec3* verts, int* triangles, int max_verts, int max_tris, int* out_num_verts, int* out_num_tris);
135
+ WP_API uint64_t wp_marching_cubes_create_device(void* context);
136
+ WP_API void wp_marching_cubes_destroy_device(uint64_t id);
137
+ WP_API int wp_marching_cubes_surface_device(uint64_t id, const float* field, int nx, int ny, int nz, float threshold, wp::vec3* verts, int* triangles, int max_verts, int max_tris, int* out_num_verts, int* out_num_tris);
138
138
 
139
139
  // generic copy supporting non-contiguous arrays
140
- WP_API bool array_copy_host(void* dst, void* src, int dst_type, int src_type, int elem_size);
141
- WP_API bool array_copy_device(void* context, void* dst, void* src, int dst_type, int src_type, int elem_size);
140
+ WP_API bool wp_array_copy_host(void* dst, void* src, int dst_type, int src_type, int elem_size);
141
+ WP_API bool wp_array_copy_device(void* context, void* dst, void* src, int dst_type, int src_type, int elem_size);
142
142
 
143
143
  // generic fill for non-contiguous arrays
144
- WP_API void array_fill_host(void* arr, int arr_type, const void* value, int value_size);
145
- WP_API void array_fill_device(void* context, void* arr, int arr_type, const void* value, int value_size);
144
+ WP_API void wp_array_fill_host(void* arr, int arr_type, const void* value, int value_size);
145
+ WP_API void wp_array_fill_device(void* context, void* arr, int arr_type, const void* value, int value_size);
146
146
 
147
- WP_API void array_inner_float_host(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
148
- WP_API void array_inner_double_host(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
149
- WP_API void array_inner_float_device(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
150
- WP_API void array_inner_double_device(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
147
+ WP_API void wp_array_inner_float_host(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
148
+ WP_API void wp_array_inner_double_host(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
149
+ WP_API void wp_array_inner_float_device(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
150
+ WP_API void wp_array_inner_double_device(uint64_t a, uint64_t b, uint64_t out, int count, int stride_a, int stride_b, int type_len);
151
151
 
152
- WP_API void array_sum_float_device(uint64_t a, uint64_t out, int count, int stride, int type_len);
153
- WP_API void array_sum_float_host(uint64_t a, uint64_t out, int count, int stride, int type_len);
154
- WP_API void array_sum_double_host(uint64_t a, uint64_t out, int count, int stride, int type_len);
155
- WP_API void array_sum_double_device(uint64_t a, uint64_t out, int count, int stride, int type_len);
152
+ WP_API void wp_array_sum_float_device(uint64_t a, uint64_t out, int count, int stride, int type_len);
153
+ WP_API void wp_array_sum_float_host(uint64_t a, uint64_t out, int count, int stride, int type_len);
154
+ WP_API void wp_array_sum_double_host(uint64_t a, uint64_t out, int count, int stride, int type_len);
155
+ WP_API void wp_array_sum_double_device(uint64_t a, uint64_t out, int count, int stride, int type_len);
156
156
 
157
- WP_API void array_scan_int_host(uint64_t in, uint64_t out, int len, bool inclusive);
158
- WP_API void array_scan_float_host(uint64_t in, uint64_t out, int len, bool inclusive);
157
+ WP_API void wp_array_scan_int_host(uint64_t in, uint64_t out, int len, bool inclusive);
158
+ WP_API void wp_array_scan_float_host(uint64_t in, uint64_t out, int len, bool inclusive);
159
159
 
160
- WP_API void array_scan_int_device(uint64_t in, uint64_t out, int len, bool inclusive);
161
- WP_API void array_scan_float_device(uint64_t in, uint64_t out, int len, bool inclusive);
160
+ WP_API void wp_array_scan_int_device(uint64_t in, uint64_t out, int len, bool inclusive);
161
+ WP_API void wp_array_scan_float_device(uint64_t in, uint64_t out, int len, bool inclusive);
162
162
 
163
- WP_API void radix_sort_pairs_int_host(uint64_t keys, uint64_t values, int n);
164
- WP_API void radix_sort_pairs_int_device(uint64_t keys, uint64_t values, int n);
163
+ WP_API void wp_radix_sort_pairs_int_host(uint64_t keys, uint64_t values, int n);
164
+ WP_API void wp_radix_sort_pairs_int_device(uint64_t keys, uint64_t values, int n);
165
165
 
166
- WP_API void radix_sort_pairs_float_host(uint64_t keys, uint64_t values, int n);
167
- WP_API void radix_sort_pairs_float_device(uint64_t keys, uint64_t values, int n);
166
+ WP_API void wp_radix_sort_pairs_float_host(uint64_t keys, uint64_t values, int n);
167
+ WP_API void wp_radix_sort_pairs_float_device(uint64_t keys, uint64_t values, int n);
168
168
 
169
- WP_API void radix_sort_pairs_int64_host(uint64_t keys, uint64_t values, int n);
170
- WP_API void radix_sort_pairs_int64_device(uint64_t keys, uint64_t values, int n);
169
+ WP_API void wp_radix_sort_pairs_int64_host(uint64_t keys, uint64_t values, int n);
170
+ WP_API void wp_radix_sort_pairs_int64_device(uint64_t keys, uint64_t values, int n);
171
171
 
172
- WP_API void segmented_sort_pairs_float_host(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
173
- WP_API void segmented_sort_pairs_float_device(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
172
+ WP_API void wp_segmented_sort_pairs_float_host(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
173
+ WP_API void wp_segmented_sort_pairs_float_device(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
174
174
 
175
- WP_API void segmented_sort_pairs_int_host(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
176
- WP_API void segmented_sort_pairs_int_device(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
175
+ WP_API void wp_segmented_sort_pairs_int_host(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
176
+ WP_API void wp_segmented_sort_pairs_int_device(uint64_t keys, uint64_t values, int n, uint64_t segment_start_indices, uint64_t segment_end_indices, int num_segments);
177
177
 
178
- WP_API void runlength_encode_int_host(uint64_t values, uint64_t run_values, uint64_t run_lengths, uint64_t run_count, int n);
179
- WP_API void runlength_encode_int_device(uint64_t values, uint64_t run_values, uint64_t run_lengths, uint64_t run_count, int n);
178
+ WP_API void wp_runlength_encode_int_host(uint64_t values, uint64_t run_values, uint64_t run_lengths, uint64_t run_count, int n);
179
+ WP_API void wp_runlength_encode_int_device(uint64_t values, uint64_t run_values, uint64_t run_lengths, uint64_t run_count, int n);
180
180
 
181
- WP_API void bsr_matrix_from_triplets_host(
181
+ WP_API void wp_bsr_matrix_from_triplets_host(
182
182
  int block_size,
183
183
  int scalar_size_in_bytes,
184
184
  int row_count,
@@ -196,7 +196,7 @@ extern "C"
196
196
  int* bsr_columns,
197
197
  int* bsr_nnz,
198
198
  void* bsr_nnz_event);
199
- WP_API void bsr_matrix_from_triplets_device(
199
+ WP_API void wp_bsr_matrix_from_triplets_device(
200
200
  int block_size,
201
201
  int scalar_size_in_bytes,
202
202
  int row_count,
@@ -215,14 +215,14 @@ extern "C"
215
215
  int* bsr_nnz,
216
216
  void* bsr_nnz_event);
217
217
 
218
- WP_API void bsr_transpose_host(
218
+ WP_API void wp_bsr_transpose_host(
219
219
  int row_count, int col_count, int nnz,
220
220
  const int* bsr_offsets, const int* bsr_columns,
221
221
  int* transposed_bsr_offsets,
222
222
  int* transposed_bsr_columns,
223
223
  int* src_block_indices
224
224
  );
225
- WP_API void bsr_transpose_device(
225
+ WP_API void wp_bsr_transpose_device(
226
226
  int row_count, int col_count, int nnz,
227
227
  const int* bsr_offsets, const int* bsr_columns,
228
228
  int* transposed_bsr_offsets,
@@ -230,125 +230,126 @@ extern "C"
230
230
  int* src_block_indices);
231
231
 
232
232
 
233
- WP_API int cuda_driver_version(); // CUDA driver version
234
- WP_API int cuda_toolkit_version(); // CUDA Toolkit version used to build Warp
235
- WP_API bool cuda_driver_is_initialized();
236
-
237
- WP_API int nvrtc_supported_arch_count();
238
- WP_API void nvrtc_supported_archs(int* archs);
239
-
240
- WP_API int cuda_device_get_count();
241
- WP_API void* cuda_device_get_primary_context(int ordinal);
242
- WP_API const char* cuda_device_get_name(int ordinal);
243
- WP_API int cuda_device_get_arch(int ordinal);
244
- WP_API int cuda_device_get_sm_count(int ordinal);
245
- WP_API void cuda_device_get_uuid(int ordinal, char uuid[16]);
246
- WP_API int cuda_device_get_pci_domain_id(int ordinal);
247
- WP_API int cuda_device_get_pci_bus_id(int ordinal);
248
- WP_API int cuda_device_get_pci_device_id(int ordinal);
249
- WP_API int cuda_device_is_uva(int ordinal);
250
- WP_API int cuda_device_is_mempool_supported(int ordinal);
251
- WP_API int cuda_device_is_ipc_supported(int ordinal);
252
- WP_API int cuda_device_set_mempool_release_threshold(int ordinal, uint64_t threshold);
253
- WP_API uint64_t cuda_device_get_mempool_release_threshold(int ordinal);
254
- WP_API uint64_t cuda_device_get_mempool_used_mem_current(int ordinal);
255
- WP_API uint64_t cuda_device_get_mempool_used_mem_high(int ordinal);
256
- WP_API void cuda_device_get_memory_info(int ordinal, size_t* free_mem, size_t* total_mem);
257
-
258
- WP_API void* cuda_context_get_current();
259
- WP_API void cuda_context_set_current(void* context);
260
- WP_API void cuda_context_push_current(void* context);
261
- WP_API void cuda_context_pop_current();
262
- WP_API void* cuda_context_create(int device_ordinal);
263
- WP_API void cuda_context_destroy(void* context);
264
- WP_API int cuda_context_get_device_ordinal(void* context);
265
- WP_API int cuda_context_is_primary(void* context);
266
- WP_API void* cuda_context_get_stream(void* context);
267
- WP_API void cuda_context_set_stream(void* context, void* stream, int sync);
233
+ WP_API int wp_cuda_driver_version(); // CUDA driver version
234
+ WP_API int wp_cuda_toolkit_version(); // CUDA Toolkit version used to build Warp
235
+ WP_API bool wp_cuda_driver_is_initialized();
236
+
237
+ WP_API int wp_nvrtc_supported_arch_count();
238
+ WP_API void wp_nvrtc_supported_archs(int* archs);
239
+
240
+ WP_API int wp_cuda_device_get_count();
241
+ WP_API void* wp_cuda_device_get_primary_context(int ordinal);
242
+ WP_API const char* wp_cuda_device_get_name(int ordinal);
243
+ WP_API int wp_cuda_device_get_arch(int ordinal);
244
+ WP_API int wp_cuda_device_get_sm_count(int ordinal);
245
+ WP_API void wp_cuda_device_get_uuid(int ordinal, char uuid[16]);
246
+ WP_API int wp_cuda_device_get_pci_domain_id(int ordinal);
247
+ WP_API int wp_cuda_device_get_pci_bus_id(int ordinal);
248
+ WP_API int wp_cuda_device_get_pci_device_id(int ordinal);
249
+ WP_API int wp_cuda_device_is_uva(int ordinal);
250
+ WP_API int wp_cuda_device_is_mempool_supported(int ordinal);
251
+ WP_API int wp_cuda_device_is_ipc_supported(int ordinal);
252
+ WP_API int wp_cuda_device_set_mempool_release_threshold(int ordinal, uint64_t threshold);
253
+ WP_API uint64_t wp_cuda_device_get_mempool_release_threshold(int ordinal);
254
+ WP_API uint64_t wp_cuda_device_get_mempool_used_mem_current(int ordinal);
255
+ WP_API uint64_t wp_cuda_device_get_mempool_used_mem_high(int ordinal);
256
+ WP_API void wp_cuda_device_get_memory_info(int ordinal, size_t* free_mem, size_t* total_mem);
257
+
258
+ WP_API void* wp_cuda_context_get_current();
259
+ WP_API void wp_cuda_context_set_current(void* context);
260
+ WP_API void wp_cuda_context_push_current(void* context);
261
+ WP_API void wp_cuda_context_pop_current();
262
+ WP_API void* wp_cuda_context_create(int device_ordinal);
263
+ WP_API void wp_cuda_context_destroy(void* context);
264
+ WP_API int wp_cuda_context_get_device_ordinal(void* context);
265
+ WP_API int wp_cuda_context_is_primary(void* context);
266
+ WP_API void* wp_cuda_context_get_stream(void* context);
267
+ WP_API void wp_cuda_context_set_stream(void* context, void* stream, int sync);
268
268
 
269
269
  // ensures all device side operations have completed in the current context
270
- WP_API void cuda_context_synchronize(void* context);
270
+ WP_API void wp_cuda_context_synchronize(void* context);
271
271
 
272
272
  // return cudaError_t code
273
- WP_API uint64_t cuda_context_check(void* context);
273
+ WP_API uint64_t wp_cuda_context_check(void* context);
274
274
 
275
275
  // peer access
276
- WP_API int cuda_is_peer_access_supported(int target_ordinal, int peer_ordinal);
277
- WP_API int cuda_is_peer_access_enabled(void* target_context, void* peer_context);
278
- WP_API int cuda_set_peer_access_enabled(void* target_context, void* peer_context, int enable);
279
- WP_API int cuda_is_mempool_access_enabled(int target_ordinal, int peer_ordinal);
280
- WP_API int cuda_set_mempool_access_enabled(int target_ordinal, int peer_ordinal, int enable);
276
+ WP_API int wp_cuda_is_peer_access_supported(int target_ordinal, int peer_ordinal);
277
+ WP_API int wp_cuda_is_peer_access_enabled(void* target_context, void* peer_context);
278
+ WP_API int wp_cuda_set_peer_access_enabled(void* target_context, void* peer_context, int enable);
279
+ WP_API int wp_cuda_is_mempool_access_enabled(int target_ordinal, int peer_ordinal);
280
+ WP_API int wp_cuda_set_mempool_access_enabled(int target_ordinal, int peer_ordinal, int enable);
281
281
 
282
282
  // inter-process communication
283
- WP_API void cuda_ipc_get_mem_handle(void* ptr, char* out_buffer);
284
- WP_API void* cuda_ipc_open_mem_handle(void* context, char* handle);
285
- WP_API void cuda_ipc_close_mem_handle(void* ptr);
286
- WP_API void cuda_ipc_get_event_handle(void* context, void* event, char* out_buffer);
287
- WP_API void* cuda_ipc_open_event_handle(void* context, char* handle);
288
-
289
- WP_API void* cuda_stream_create(void* context, int priority);
290
- WP_API void cuda_stream_destroy(void* context, void* stream);
291
- WP_API int cuda_stream_query(void* stream);
292
- WP_API void cuda_stream_register(void* context, void* stream);
293
- WP_API void cuda_stream_unregister(void* context, void* stream);
294
- WP_API void* cuda_stream_get_current();
295
- WP_API void cuda_stream_synchronize(void* stream);
296
- WP_API void cuda_stream_wait_event(void* stream, void* event);
297
- WP_API void cuda_stream_wait_stream(void* stream, void* other_stream, void* event);
298
- WP_API int cuda_stream_is_capturing(void* stream);
299
- WP_API uint64_t cuda_stream_get_capture_id(void* stream);
300
- WP_API int cuda_stream_get_priority(void* stream);
301
-
302
- WP_API void* cuda_event_create(void* context, unsigned flags);
303
- WP_API void cuda_event_destroy(void* event);
304
- WP_API int cuda_event_query(void* event);
305
- WP_API void cuda_event_record(void* event, void* stream, bool timing=false);
306
- WP_API void cuda_event_synchronize(void* event);
307
- WP_API float cuda_event_elapsed_time(void* start_event, void* end_event);
308
-
309
- WP_API bool cuda_graph_begin_capture(void* context, void* stream, int external);
310
- WP_API bool cuda_graph_end_capture(void* context, void* stream, void** graph_ret);
311
- WP_API bool cuda_graph_create_exec(void* context, void* stream, void* graph, void** graph_exec_ret);
312
- WP_API bool cuda_graph_launch(void* graph, void* stream);
313
- WP_API bool cuda_graph_destroy(void* context, void* graph);
314
- WP_API bool cuda_graph_exec_destroy(void* context, void* graph_exec);
315
- WP_API bool capture_debug_dot_print(void* graph, const char *path, uint32_t flags);
316
-
317
- WP_API bool cuda_graph_insert_if_else(void* context, void* stream, int* condition, void** if_graph_ret, void** else_graph_ret);
318
- WP_API bool cuda_graph_insert_while(void* context, void* stream, int* condition, void** body_graph_ret, uint64_t* handle_ret);
319
- WP_API bool cuda_graph_set_condition(void* context, void* stream, int* condition, uint64_t handle);
320
- WP_API bool cuda_graph_pause_capture(void* context, void* stream, void** graph_ret);
321
- WP_API bool cuda_graph_resume_capture(void* context, void* stream, void* graph);
322
- WP_API bool cuda_graph_insert_child_graph(void* context, void* stream, void* child_graph);
323
-
324
- WP_API size_t cuda_compile_program(const char* cuda_src, const char* program_name, int arch, const char* include_dir, int num_cuda_include_dirs, const char** cuda_include_dirs, bool debug, bool verbose, bool verify_fp, bool fast_math, bool fuse_fp, bool lineinfo, bool compile_time_trace, const char* output_path, size_t num_ltoirs, char** ltoirs, size_t* ltoir_sizes, int* ltoir_input_types);
325
- WP_API bool cuda_compile_fft(const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int size, int elements_per_thread, int direction, int precision, int* shared_memory_size);
326
- WP_API bool cuda_compile_dot(const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int M, int N, int K, int precision_A, int precision_B, int precision_C, int type, int arrangement_A, int arrangement_B, int arrangement_C, int num_threads);
327
- WP_API bool cuda_compile_solver(const char* fatbin_output_path, const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int M, int N, int NRHS, int function, int side, int diag, int precision, int arrangement_A, int arrangement_B, int fill_mode, int num_threads);
328
-
329
- WP_API void* cuda_load_module(void* context, const char* ptx);
330
- WP_API void cuda_unload_module(void* context, void* module);
331
- WP_API void* cuda_get_kernel(void* context, void* module, const char* name);
332
- WP_API size_t cuda_launch_kernel(void* context, void* kernel, size_t dim, int max_blocks, int block_dim, int shared_memory_bytes, void** args, void* stream);
333
- WP_API int cuda_get_max_shared_memory(void* context);
334
- WP_API bool cuda_configure_kernel_shared_memory(void* kernel, int size);
335
-
336
- WP_API void cuda_set_context_restore_policy(bool always_restore);
337
- WP_API int cuda_get_context_restore_policy();
338
-
339
- WP_API void cuda_graphics_map(void* context, void* resource);
340
- WP_API void cuda_graphics_unmap(void* context, void* resource);
341
- WP_API void cuda_graphics_device_ptr_and_size(void* context, void* resource, uint64_t* ptr, size_t* size);
342
- WP_API void* cuda_graphics_register_gl_buffer(void* context, uint32_t gl_buffer, unsigned int flags);
343
- WP_API void cuda_graphics_unregister_resource(void* context, void* resource);
283
+ WP_API void wp_cuda_ipc_get_mem_handle(void* ptr, char* out_buffer);
284
+ WP_API void* wp_cuda_ipc_open_mem_handle(void* context, char* handle);
285
+ WP_API void wp_cuda_ipc_close_mem_handle(void* ptr);
286
+ WP_API void wp_cuda_ipc_get_event_handle(void* context, void* event, char* out_buffer);
287
+ WP_API void* wp_cuda_ipc_open_event_handle(void* context, char* handle);
288
+
289
+ WP_API void* wp_cuda_stream_create(void* context, int priority);
290
+ WP_API void wp_cuda_stream_destroy(void* context, void* stream);
291
+ WP_API int wp_cuda_stream_query(void* stream);
292
+ WP_API void wp_cuda_stream_register(void* context, void* stream);
293
+ WP_API void wp_cuda_stream_unregister(void* context, void* stream);
294
+ WP_API void* wp_cuda_stream_get_current();
295
+ WP_API void wp_cuda_stream_synchronize(void* stream);
296
+ WP_API void wp_cuda_stream_wait_event(void* stream, void* event);
297
+ WP_API void wp_cuda_stream_wait_stream(void* stream, void* other_stream, void* event);
298
+ WP_API int wp_cuda_stream_is_capturing(void* stream);
299
+ WP_API uint64_t wp_cuda_stream_get_capture_id(void* stream);
300
+ WP_API int wp_cuda_stream_get_priority(void* stream);
301
+
302
+ WP_API void* wp_cuda_event_create(void* context, unsigned flags);
303
+ WP_API void wp_cuda_event_destroy(void* event);
304
+ WP_API int wp_cuda_event_query(void* event);
305
+ WP_API void wp_cuda_event_record(void* event, void* stream, bool timing=false);
306
+ WP_API void wp_cuda_event_synchronize(void* event);
307
+ WP_API float wp_cuda_event_elapsed_time(void* start_event, void* end_event);
308
+
309
+ WP_API bool wp_cuda_graph_begin_capture(void* context, void* stream, int external);
310
+ WP_API bool wp_cuda_graph_end_capture(void* context, void* stream, void** graph_ret);
311
+ WP_API bool wp_cuda_graph_create_exec(void* context, void* stream, void* graph, void** graph_exec_ret);
312
+ WP_API bool wp_cuda_graph_launch(void* graph, void* stream);
313
+ WP_API bool wp_cuda_graph_destroy(void* context, void* graph);
314
+ WP_API bool wp_cuda_graph_exec_destroy(void* context, void* graph_exec);
315
+ WP_API bool wp_capture_debug_dot_print(void* graph, const char *path, uint32_t flags);
316
+
317
+ WP_API bool wp_cuda_graph_insert_if_else(void* context, void* stream, int* condition, void** if_graph_ret, void** else_graph_ret);
318
+ WP_API bool wp_cuda_graph_insert_while(void* context, void* stream, int* condition, void** body_graph_ret, uint64_t* handle_ret);
319
+ WP_API bool wp_cuda_graph_set_condition(void* context, void* stream, int* condition, uint64_t handle);
320
+ WP_API bool wp_cuda_graph_pause_capture(void* context, void* stream, void** graph_ret);
321
+ WP_API bool wp_cuda_graph_resume_capture(void* context, void* stream, void* graph);
322
+ WP_API bool wp_cuda_graph_insert_child_graph(void* context, void* stream, void* child_graph);
323
+ WP_API bool wp_cuda_graph_check_conditional_body(void* body_graph);
324
+
325
+ WP_API size_t wp_cuda_compile_program(const char* cuda_src, const char* program_name, int arch, const char* include_dir, int num_cuda_include_dirs, const char** cuda_include_dirs, bool debug, bool verbose, bool verify_fp, bool fast_math, bool fuse_fp, bool lineinfo, bool compile_time_trace, const char* output_path, size_t num_ltoirs, char** ltoirs, size_t* ltoir_sizes, int* ltoir_input_types);
326
+ WP_API bool wp_cuda_compile_fft(const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int size, int elements_per_thread, int direction, int precision, int* shared_memory_size);
327
+ WP_API bool wp_cuda_compile_dot(const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int M, int N, int K, int precision_A, int precision_B, int precision_C, int type, int arrangement_A, int arrangement_B, int arrangement_C, int num_threads);
328
+ WP_API bool wp_cuda_compile_solver(const char* fatbin_output_path, const char* ltoir_output_path, const char* symbol_name, int num_include_dirs, const char** include_dirs, const char* mathdx_include_dir, int arch, int M, int N, int NRHS, int function, int side, int diag, int precision, int arrangement_A, int arrangement_B, int fill_mode, int num_threads);
329
+
330
+ WP_API void* wp_cuda_load_module(void* context, const char* ptx);
331
+ WP_API void wp_cuda_unload_module(void* context, void* module);
332
+ WP_API void* wp_cuda_get_kernel(void* context, void* module, const char* name);
333
+ WP_API size_t wp_cuda_launch_kernel(void* context, void* kernel, size_t dim, int max_blocks, int block_dim, int shared_memory_bytes, void** args, void* stream);
334
+ WP_API int wp_cuda_get_max_shared_memory(void* context);
335
+ WP_API bool wp_cuda_configure_kernel_shared_memory(void* kernel, int size);
336
+
337
+ WP_API void wp_cuda_set_context_restore_policy(bool always_restore);
338
+ WP_API int wp_cuda_get_context_restore_policy();
339
+
340
+ WP_API void wp_cuda_graphics_map(void* context, void* resource);
341
+ WP_API void wp_cuda_graphics_unmap(void* context, void* resource);
342
+ WP_API void wp_cuda_graphics_device_ptr_and_size(void* context, void* resource, uint64_t* ptr, size_t* size);
343
+ WP_API void* wp_cuda_graphics_register_gl_buffer(void* context, uint32_t gl_buffer, unsigned int flags);
344
+ WP_API void wp_cuda_graphics_unregister_resource(void* context, void* resource);
344
345
 
345
346
  // CUDA timing
346
- WP_API void cuda_timing_begin(int flags);
347
- WP_API int cuda_timing_get_result_count();
348
- WP_API void cuda_timing_end(timing_result_t* results, int size);
347
+ WP_API void wp_cuda_timing_begin(int flags);
348
+ WP_API int wp_cuda_timing_get_result_count();
349
+ WP_API void wp_cuda_timing_end(timing_result_t* results, int size);
349
350
 
350
351
  // graph coloring
351
- WP_API int graph_coloring(int num_nodes, wp::array_t<int> edges, int algorithm, wp::array_t<int> node_colors);
352
- WP_API float balance_coloring(int num_nodes, wp::array_t<int> edges, int num_colors, float target_max_min_ratio, wp::array_t<int> node_colors);
352
+ WP_API int wp_graph_coloring(int num_nodes, wp::array_t<int> edges, int algorithm, wp::array_t<int> node_colors);
353
+ WP_API float wp_balance_coloring(int num_nodes, wp::array_t<int> edges, int num_colors, float target_max_min_ratio, wp::array_t<int> node_colors);
353
354
 
354
355
  } // extern "C"