triton-windows 3.3.1.post19__cp311-cp311-win_amd64.whl → 3.5.0.post21__cp311-cp311-win_amd64.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 triton-windows might be problematic. Click here for more details.

Files changed (225) hide show
  1. triton/_C/libtriton.pyd +0 -0
  2. triton/__init__.py +11 -2
  3. triton/_filecheck.py +97 -0
  4. triton/_internal_testing.py +95 -18
  5. triton/_utils.py +112 -21
  6. triton/backends/__init__.py +20 -23
  7. triton/backends/amd/__init__.py +0 -0
  8. triton/backends/amd/compiler.py +161 -119
  9. triton/backends/amd/driver.c +118 -46
  10. triton/backends/amd/driver.py +274 -96
  11. triton/backends/compiler.py +7 -21
  12. triton/backends/driver.py +13 -0
  13. triton/backends/nvidia/bin/ptxas.exe +0 -0
  14. triton/backends/nvidia/compiler.py +163 -106
  15. triton/backends/nvidia/driver.c +166 -101
  16. triton/backends/nvidia/driver.py +384 -202
  17. triton/compiler/__init__.py +5 -2
  18. triton/compiler/code_generator.py +439 -231
  19. triton/compiler/compiler.py +152 -84
  20. triton/experimental/__init__.py +0 -0
  21. triton/experimental/gluon/__init__.py +5 -0
  22. triton/experimental/gluon/_compiler.py +0 -0
  23. triton/experimental/gluon/_runtime.py +102 -0
  24. triton/experimental/gluon/language/__init__.py +119 -0
  25. triton/experimental/gluon/language/_core.py +490 -0
  26. triton/experimental/gluon/language/_layouts.py +583 -0
  27. triton/experimental/gluon/language/_math.py +20 -0
  28. triton/experimental/gluon/language/_semantic.py +380 -0
  29. triton/experimental/gluon/language/_standard.py +80 -0
  30. triton/experimental/gluon/language/amd/__init__.py +4 -0
  31. triton/experimental/gluon/language/amd/_layouts.py +96 -0
  32. triton/experimental/gluon/language/amd/cdna3/__init__.py +100 -0
  33. triton/experimental/gluon/language/amd/cdna4/__init__.py +48 -0
  34. triton/experimental/gluon/language/amd/cdna4/async_copy.py +151 -0
  35. triton/experimental/gluon/language/extra/__init__.py +3 -0
  36. triton/experimental/gluon/language/nvidia/__init__.py +4 -0
  37. triton/experimental/gluon/language/nvidia/ampere/__init__.py +3 -0
  38. triton/experimental/gluon/language/nvidia/ampere/async_copy.py +74 -0
  39. triton/experimental/gluon/language/nvidia/ampere/mbarrier.py +80 -0
  40. triton/experimental/gluon/language/nvidia/blackwell/__init__.py +387 -0
  41. triton/experimental/gluon/language/nvidia/blackwell/tma.py +52 -0
  42. triton/experimental/gluon/language/nvidia/hopper/__init__.py +132 -0
  43. triton/experimental/gluon/language/nvidia/hopper/mbarrier.py +34 -0
  44. triton/experimental/gluon/language/nvidia/hopper/tma.py +97 -0
  45. triton/experimental/gluon/nvidia/__init__.py +4 -0
  46. triton/experimental/gluon/nvidia/blackwell.py +3 -0
  47. triton/experimental/gluon/nvidia/hopper.py +45 -0
  48. triton/knobs.py +546 -0
  49. triton/language/__init__.py +50 -19
  50. triton/language/core.py +909 -572
  51. triton/language/extra/cuda/__init__.py +10 -7
  52. triton/language/extra/cuda/gdc.py +42 -0
  53. triton/language/extra/cuda/libdevice.py +394 -394
  54. triton/language/extra/cuda/utils.py +21 -21
  55. triton/language/extra/hip/__init__.py +3 -1
  56. triton/language/extra/hip/libdevice.py +120 -104
  57. triton/language/extra/hip/utils.py +35 -0
  58. triton/language/extra/libdevice.py +4 -0
  59. triton/language/math.py +65 -66
  60. triton/language/random.py +12 -2
  61. triton/language/semantic.py +1757 -1768
  62. triton/language/standard.py +127 -62
  63. triton/language/target_info.py +54 -0
  64. triton/runtime/_allocation.py +15 -3
  65. triton/runtime/_async_compile.py +55 -0
  66. triton/runtime/autotuner.py +117 -60
  67. triton/runtime/build.py +83 -17
  68. triton/runtime/cache.py +61 -47
  69. triton/runtime/driver.py +25 -47
  70. triton/runtime/interpreter.py +95 -50
  71. triton/runtime/jit.py +445 -248
  72. triton/runtime/tcc/include/_mingw.h +8 -10
  73. triton/runtime/tcc/include/assert.h +5 -0
  74. triton/runtime/tcc/include/errno.h +1 -1
  75. triton/runtime/tcc/include/float.h +21 -3
  76. triton/runtime/tcc/include/iso646.h +36 -0
  77. triton/runtime/tcc/include/limits.h +5 -0
  78. triton/runtime/tcc/include/malloc.h +2 -2
  79. triton/runtime/tcc/include/math.h +21 -261
  80. triton/runtime/tcc/include/stdalign.h +16 -0
  81. triton/runtime/tcc/include/stdarg.h +5 -70
  82. triton/runtime/tcc/include/stdatomic.h +171 -0
  83. triton/runtime/tcc/include/stddef.h +7 -19
  84. triton/runtime/tcc/include/stdlib.h +15 -4
  85. triton/runtime/tcc/include/stdnoreturn.h +7 -0
  86. triton/runtime/tcc/include/sys/stat.h +2 -2
  87. triton/runtime/tcc/include/sys/types.h +5 -0
  88. triton/runtime/tcc/include/tcc/tcc_libm.h +444 -27
  89. triton/runtime/tcc/include/tccdefs.h +342 -0
  90. triton/runtime/tcc/include/tgmath.h +89 -0
  91. triton/runtime/tcc/include/uchar.h +33 -0
  92. triton/runtime/tcc/include/unistd.h +1 -0
  93. triton/runtime/tcc/include/winapi/qos.h +72 -0
  94. triton/runtime/tcc/include/winapi/shellapi.h +59 -0
  95. triton/runtime/tcc/include/winapi/winbase.h +9 -2
  96. triton/runtime/tcc/include/winapi/wincon.h +8 -0
  97. triton/runtime/tcc/include/winapi/windows.h +1 -1
  98. triton/runtime/tcc/include/winapi/winnls.h +778 -0
  99. triton/runtime/tcc/include/winapi/winnt.h +9 -7
  100. triton/runtime/tcc/include/winapi/winsock2.h +1474 -0
  101. triton/runtime/tcc/include/winapi/ws2ipdef.h +21 -0
  102. triton/runtime/tcc/include/winapi/ws2tcpip.h +391 -0
  103. triton/runtime/tcc/lib/libtcc1.a +0 -0
  104. triton/runtime/tcc/lib/python314.def +1800 -0
  105. triton/runtime/tcc/lib/python314t.def +1809 -0
  106. triton/runtime/tcc/libtcc.dll +0 -0
  107. triton/runtime/tcc/tcc.exe +0 -0
  108. triton/testing.py +16 -12
  109. triton/tools/compile.py +62 -14
  110. triton/tools/disasm.py +3 -4
  111. triton/tools/extra/cuda/compile.c +1 -0
  112. triton/tools/extra/hip/compile.cpp +66 -0
  113. triton/tools/extra/hip/compile.h +13 -0
  114. triton/tools/ragged_tma.py +92 -0
  115. triton/tools/tensor_descriptor.py +34 -0
  116. triton/windows_utils.py +52 -81
  117. {triton_windows-3.3.1.post19.dist-info → triton_windows-3.5.0.post21.dist-info}/METADATA +8 -4
  118. triton_windows-3.5.0.post21.dist-info/RECORD +217 -0
  119. triton_windows-3.5.0.post21.dist-info/entry_points.txt +3 -0
  120. triton_windows-3.5.0.post21.dist-info/licenses/LICENSE +23 -0
  121. triton_windows-3.5.0.post21.dist-info/top_level.txt +1 -0
  122. triton/backends/amd/include/hip/amd_detail/amd_channel_descriptor.h +0 -358
  123. triton/backends/amd/include/hip/amd_detail/amd_device_functions.h +0 -1010
  124. triton/backends/amd/include/hip/amd_detail/amd_hip_atomic.h +0 -1638
  125. triton/backends/amd/include/hip/amd_detail/amd_hip_bf16.h +0 -1814
  126. triton/backends/amd/include/hip/amd_detail/amd_hip_bfloat16.h +0 -293
  127. triton/backends/amd/include/hip/amd_detail/amd_hip_common.h +0 -32
  128. triton/backends/amd/include/hip/amd_detail/amd_hip_complex.h +0 -174
  129. triton/backends/amd/include/hip/amd_detail/amd_hip_cooperative_groups.h +0 -835
  130. triton/backends/amd/include/hip/amd_detail/amd_hip_fp16.h +0 -1809
  131. triton/backends/amd/include/hip/amd_detail/amd_hip_fp8.h +0 -1391
  132. triton/backends/amd/include/hip/amd_detail/amd_hip_gl_interop.h +0 -108
  133. triton/backends/amd/include/hip/amd_detail/amd_hip_math_constants.h +0 -124
  134. triton/backends/amd/include/hip/amd_detail/amd_hip_runtime.h +0 -405
  135. triton/backends/amd/include/hip/amd_detail/amd_hip_runtime_pt_api.h +0 -196
  136. triton/backends/amd/include/hip/amd_detail/amd_hip_unsafe_atomics.h +0 -565
  137. triton/backends/amd/include/hip/amd_detail/amd_hip_vector_types.h +0 -2226
  138. triton/backends/amd/include/hip/amd_detail/amd_math_functions.h +0 -104
  139. triton/backends/amd/include/hip/amd_detail/amd_surface_functions.h +0 -244
  140. triton/backends/amd/include/hip/amd_detail/amd_warp_functions.h +0 -538
  141. triton/backends/amd/include/hip/amd_detail/amd_warp_sync_functions.h +0 -288
  142. triton/backends/amd/include/hip/amd_detail/concepts.hpp +0 -30
  143. triton/backends/amd/include/hip/amd_detail/device_library_decls.h +0 -133
  144. triton/backends/amd/include/hip/amd_detail/functional_grid_launch.hpp +0 -218
  145. triton/backends/amd/include/hip/amd_detail/grid_launch.h +0 -67
  146. triton/backends/amd/include/hip/amd_detail/grid_launch.hpp +0 -50
  147. triton/backends/amd/include/hip/amd_detail/grid_launch_GGL.hpp +0 -26
  148. triton/backends/amd/include/hip/amd_detail/helpers.hpp +0 -137
  149. triton/backends/amd/include/hip/amd_detail/hip_api_trace.hpp +0 -1446
  150. triton/backends/amd/include/hip/amd_detail/hip_assert.h +0 -101
  151. triton/backends/amd/include/hip/amd_detail/hip_cooperative_groups_helper.h +0 -242
  152. triton/backends/amd/include/hip/amd_detail/hip_fp16_gcc.h +0 -254
  153. triton/backends/amd/include/hip/amd_detail/hip_fp16_math_fwd.h +0 -96
  154. triton/backends/amd/include/hip/amd_detail/hip_ldg.h +0 -100
  155. triton/backends/amd/include/hip/amd_detail/hip_prof_str.h +0 -10570
  156. triton/backends/amd/include/hip/amd_detail/hip_runtime_prof.h +0 -78
  157. triton/backends/amd/include/hip/amd_detail/host_defines.h +0 -184
  158. triton/backends/amd/include/hip/amd_detail/hsa_helpers.hpp +0 -102
  159. triton/backends/amd/include/hip/amd_detail/macro_based_grid_launch.hpp +0 -798
  160. triton/backends/amd/include/hip/amd_detail/math_fwd.h +0 -698
  161. triton/backends/amd/include/hip/amd_detail/ockl_image.h +0 -177
  162. triton/backends/amd/include/hip/amd_detail/program_state.hpp +0 -107
  163. triton/backends/amd/include/hip/amd_detail/texture_fetch_functions.h +0 -491
  164. triton/backends/amd/include/hip/amd_detail/texture_indirect_functions.h +0 -478
  165. triton/backends/amd/include/hip/channel_descriptor.h +0 -39
  166. triton/backends/amd/include/hip/device_functions.h +0 -38
  167. triton/backends/amd/include/hip/driver_types.h +0 -468
  168. triton/backends/amd/include/hip/hip_bf16.h +0 -36
  169. triton/backends/amd/include/hip/hip_bfloat16.h +0 -44
  170. triton/backends/amd/include/hip/hip_common.h +0 -100
  171. triton/backends/amd/include/hip/hip_complex.h +0 -38
  172. triton/backends/amd/include/hip/hip_cooperative_groups.h +0 -46
  173. triton/backends/amd/include/hip/hip_deprecated.h +0 -95
  174. triton/backends/amd/include/hip/hip_ext.h +0 -161
  175. triton/backends/amd/include/hip/hip_fp16.h +0 -36
  176. triton/backends/amd/include/hip/hip_fp8.h +0 -33
  177. triton/backends/amd/include/hip/hip_gl_interop.h +0 -32
  178. triton/backends/amd/include/hip/hip_hcc.h +0 -24
  179. triton/backends/amd/include/hip/hip_math_constants.h +0 -36
  180. triton/backends/amd/include/hip/hip_profile.h +0 -27
  181. triton/backends/amd/include/hip/hip_runtime.h +0 -75
  182. triton/backends/amd/include/hip/hip_runtime_api.h +0 -9261
  183. triton/backends/amd/include/hip/hip_texture_types.h +0 -29
  184. triton/backends/amd/include/hip/hip_vector_types.h +0 -41
  185. triton/backends/amd/include/hip/hip_version.h +0 -17
  186. triton/backends/amd/include/hip/hiprtc.h +0 -421
  187. triton/backends/amd/include/hip/library_types.h +0 -78
  188. triton/backends/amd/include/hip/math_functions.h +0 -42
  189. triton/backends/amd/include/hip/surface_types.h +0 -63
  190. triton/backends/amd/include/hip/texture_types.h +0 -194
  191. triton/backends/amd/include/hsa/Brig.h +0 -1131
  192. triton/backends/amd/include/hsa/amd_hsa_common.h +0 -91
  193. triton/backends/amd/include/hsa/amd_hsa_elf.h +0 -462
  194. triton/backends/amd/include/hsa/amd_hsa_kernel_code.h +0 -269
  195. triton/backends/amd/include/hsa/amd_hsa_queue.h +0 -109
  196. triton/backends/amd/include/hsa/amd_hsa_signal.h +0 -80
  197. triton/backends/amd/include/hsa/hsa.h +0 -5738
  198. triton/backends/amd/include/hsa/hsa_amd_tool.h +0 -91
  199. triton/backends/amd/include/hsa/hsa_api_trace.h +0 -579
  200. triton/backends/amd/include/hsa/hsa_api_trace_version.h +0 -68
  201. triton/backends/amd/include/hsa/hsa_ext_amd.h +0 -3146
  202. triton/backends/amd/include/hsa/hsa_ext_finalize.h +0 -531
  203. triton/backends/amd/include/hsa/hsa_ext_image.h +0 -1454
  204. triton/backends/amd/include/hsa/hsa_ven_amd_aqlprofile.h +0 -488
  205. triton/backends/amd/include/hsa/hsa_ven_amd_loader.h +0 -667
  206. triton/backends/amd/include/hsa/hsa_ven_amd_pc_sampling.h +0 -416
  207. triton/backends/amd/include/roctracer/ext/prof_protocol.h +0 -107
  208. triton/backends/amd/include/roctracer/hip_ostream_ops.h +0 -4515
  209. triton/backends/amd/include/roctracer/hsa_ostream_ops.h +0 -1727
  210. triton/backends/amd/include/roctracer/hsa_prof_str.h +0 -3059
  211. triton/backends/amd/include/roctracer/roctracer.h +0 -779
  212. triton/backends/amd/include/roctracer/roctracer_ext.h +0 -81
  213. triton/backends/amd/include/roctracer/roctracer_hcc.h +0 -24
  214. triton/backends/amd/include/roctracer/roctracer_hip.h +0 -37
  215. triton/backends/amd/include/roctracer/roctracer_hsa.h +0 -112
  216. triton/backends/amd/include/roctracer/roctracer_plugin.h +0 -137
  217. triton/backends/amd/include/roctracer/roctracer_roctx.h +0 -67
  218. triton/backends/amd/include/roctracer/roctx.h +0 -229
  219. triton/language/_utils.py +0 -21
  220. triton/language/extra/cuda/_experimental_tma.py +0 -106
  221. triton/runtime/tcc/lib/libtcc1-64.a +0 -0
  222. triton/tools/experimental_descriptor.py +0 -32
  223. triton_windows-3.3.1.post19.dist-info/RECORD +0 -260
  224. triton_windows-3.3.1.post19.dist-info/top_level.txt +0 -14
  225. {triton_windows-3.3.1.post19.dist-info → triton_windows-3.5.0.post21.dist-info}/WHEEL +0 -0
triton/windows_utils.py CHANGED
@@ -54,14 +54,11 @@ def max_version(
54
54
 
55
55
 
56
56
  def check_msvc(msvc_base_path: Path, version: str) -> bool:
57
- return all(
58
- x.exists()
59
- for x in [
60
- msvc_base_path / version / "bin" / "Hostx64" / "x64" / "cl.exe",
61
- msvc_base_path / version / "include" / "vcruntime.h",
62
- msvc_base_path / version / "lib" / "x64" / "vcruntime.lib",
63
- ]
64
- )
57
+ return all(x.exists() for x in [
58
+ msvc_base_path / version / "bin" / "Hostx64" / "x64" / "cl.exe",
59
+ msvc_base_path / version / "include" / "vcruntime.h",
60
+ msvc_base_path / version / "lib" / "x64" / "vcruntime.lib",
61
+ ])
65
62
 
66
63
 
67
64
  def find_msvc_env() -> tuple[Optional[Path], Optional[str]]:
@@ -72,20 +69,16 @@ def find_msvc_env() -> tuple[Optional[Path], Optional[str]]:
72
69
 
73
70
  version = os.getenv("VCToolsVersion")
74
71
  if not check_msvc(msvc_base_path, version):
75
- warnings.warn(
76
- f"Environment variables VCINSTALLDIR = {os.getenv('VCINSTALLDIR')}, "
77
- f"VCToolsVersion = {os.getenv('VCToolsVersion')} are set, "
78
- "but this MSVC installation is incomplete."
79
- )
72
+ warnings.warn(f"Environment variables VCINSTALLDIR = {os.getenv('VCINSTALLDIR')}, "
73
+ f"VCToolsVersion = {os.getenv('VCToolsVersion')} are set, "
74
+ "but this MSVC installation is incomplete.")
80
75
  return None, None
81
76
 
82
77
  return msvc_base_path, version
83
78
 
84
79
 
85
80
  def find_msvc_vswhere() -> tuple[Optional[Path], Optional[str]]:
86
- vswhere_path = find_in_program_files(
87
- r"Microsoft Visual Studio\Installer\vswhere.exe"
88
- )
81
+ vswhere_path = find_in_program_files(r"Microsoft Visual Studio\Installer\vswhere.exe")
89
82
  if vswhere_path is None:
90
83
  return None, None
91
84
 
@@ -111,9 +104,7 @@ def find_msvc_vswhere() -> tuple[Optional[Path], Optional[str]]:
111
104
  if not msvc_base_path.exists():
112
105
  return None, None
113
106
 
114
- version = max_version(
115
- os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path)
116
- )
107
+ version = max_version(os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path))
117
108
  if version is None:
118
109
  return None, None
119
110
 
@@ -132,9 +123,7 @@ def find_msvc_envpath() -> tuple[Optional[Path], Optional[str]]:
132
123
  if not msvc_base_path.exists():
133
124
  continue
134
125
 
135
- version = max_version(
136
- os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path)
137
- )
126
+ version = max_version(os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path))
138
127
  if version is None:
139
128
  continue
140
129
 
@@ -153,9 +142,7 @@ def find_msvc_hardcoded() -> tuple[Optional[Path], Optional[str]]:
153
142
  paths = sorted(paths)[::-1]
154
143
  for msvc_base_path in paths:
155
144
  msvc_base_path = Path(msvc_base_path)
156
- version = max_version(
157
- os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path)
158
- )
145
+ version = max_version(os.listdir(msvc_base_path), check=partial(check_msvc, msvc_base_path))
159
146
  if version is None:
160
147
  continue
161
148
  return msvc_base_path, version
@@ -188,13 +175,10 @@ def find_msvc(env_only: bool) -> tuple[Optional[str], list[str], list[str]]:
188
175
 
189
176
 
190
177
  def check_winsdk(winsdk_base_path: Path, version: str) -> bool:
191
- return all(
192
- x.exists()
193
- for x in [
194
- winsdk_base_path / "Include" / version / "ucrt" / "stdlib.h",
195
- winsdk_base_path / "Lib" / version / "ucrt" / "x64" / "ucrt.lib",
196
- ]
197
- )
178
+ return all(x.exists() for x in [
179
+ winsdk_base_path / "Include" / version / "ucrt" / "stdlib.h",
180
+ winsdk_base_path / "Lib" / version / "ucrt" / "x64" / "ucrt.lib",
181
+ ])
198
182
 
199
183
 
200
184
  def find_winsdk_env() -> tuple[Optional[Path], Optional[str]]:
@@ -204,14 +188,17 @@ def find_winsdk_env() -> tuple[Optional[Path], Optional[str]]:
204
188
  winsdk_base_path = Path(winsdk_base_path)
205
189
 
206
190
  version = os.getenv("WindowsSDKVersion")
207
- if version:
208
- version = version.rstrip("\\")
191
+ if version is None:
192
+ version = os.getenv("WindowsSDKVer")
193
+ if version is None:
194
+ warnings.warn(f"Environment variable WindowsSdkDir = {winsdk_base_path}, "
195
+ "but WindowsSDKVersion (or WindowsSDKVer) is not set.")
196
+ return None, None
197
+ version = version.rstrip("\\")
209
198
  if not check_winsdk(winsdk_base_path, version):
210
- warnings.warn(
211
- f"Environment variables WindowsSdkDir = {os.getenv('WindowsSdkDir')}, "
212
- f"WindowsSDKVersion = {os.getenv('WindowsSDKVersion')} are set, "
213
- "but this Windows SDK installation is incomplete."
214
- )
199
+ warnings.warn(f"Environment variables WindowsSdkDir = {winsdk_base_path}, "
200
+ f"WindowsSDKVersion (or WindowsSDKVer) = {version} are set, "
201
+ "but this Windows SDK installation is incomplete.")
215
202
  return None, None
216
203
 
217
204
  return winsdk_base_path, version
@@ -220,9 +207,7 @@ def find_winsdk_env() -> tuple[Optional[Path], Optional[str]]:
220
207
  def find_winsdk_registry() -> tuple[Optional[Path], Optional[str]]:
221
208
  try:
222
209
  reg = winreg.ConnectRegistry(None, winreg.HKEY_LOCAL_MACHINE)
223
- key = winreg.OpenKeyEx(
224
- reg, r"SOFTWARE\WOW6432Node\Microsoft\Microsoft SDKs\Windows\v10.0"
225
- )
210
+ key = winreg.OpenKeyEx(reg, r"SOFTWARE\WOW6432Node\Microsoft\Microsoft SDKs\Windows\v10.0")
226
211
  folder = winreg.QueryValueEx(key, "InstallationFolder")[0]
227
212
  winreg.CloseKey(key)
228
213
  except OSError:
@@ -288,10 +273,8 @@ def find_winsdk(env_only: bool) -> tuple[list[str], list[str]]:
288
273
  return [], []
289
274
 
290
275
 
291
- @functools.cache
292
- def find_msvc_winsdk(
293
- env_only: bool = False,
294
- ) -> tuple[Optional[str], list[str], list[str]]:
276
+ @functools.lru_cache
277
+ def find_msvc_winsdk(env_only: bool = False, ) -> tuple[Optional[str], list[str], list[str]]:
295
278
  msvc_bin_path, msvc_inc_dirs, msvc_lib_dirs = find_msvc(env_only)
296
279
  winsdk_inc_dirs, winsdk_lib_dirs = find_winsdk(env_only)
297
280
  return (
@@ -301,15 +284,18 @@ def find_msvc_winsdk(
301
284
  )
302
285
 
303
286
 
304
- @functools.cache
287
+ @functools.lru_cache
305
288
  def find_python() -> list[str]:
289
+ version = sysconfig.get_python_version().replace(".", "")
290
+ if sysconfig.get_config_var("Py_GIL_DISABLED"):
291
+ version += "t"
306
292
  for python_base_path in [
307
- sys.exec_prefix,
308
- sys.base_exec_prefix,
309
- os.path.dirname(sys.executable),
293
+ sys.exec_prefix,
294
+ sys.base_exec_prefix,
295
+ os.path.dirname(sys.executable),
310
296
  ]:
311
297
  python_lib_dir = Path(python_base_path) / "libs"
312
- if (python_lib_dir / "python3.lib").exists():
298
+ if (python_lib_dir / f"python{version}.lib").exists():
313
299
  return [str(python_lib_dir)]
314
300
 
315
301
  warnings.warn("Failed to find Python libs.")
@@ -318,14 +304,11 @@ def find_python() -> list[str]:
318
304
 
319
305
  def check_and_find_cuda(base_path: Path) -> tuple[Optional[str], list[str], list[str]]:
320
306
  # pip
321
- if all(
322
- x.exists()
323
- for x in [
307
+ if all(x.exists() for x in [
324
308
  base_path / "cuda_nvcc" / "bin" / "ptxas.exe",
325
309
  base_path / "cuda_runtime" / "include" / "cuda.h",
326
310
  base_path / "cuda_runtime" / "lib" / "x64" / "cuda.lib",
327
- ]
328
- ):
311
+ ]):
329
312
  return (
330
313
  str(base_path / "cuda_nvcc" / "bin"),
331
314
  [str(base_path / "cuda_runtime" / "include")],
@@ -333,14 +316,11 @@ def check_and_find_cuda(base_path: Path) -> tuple[Optional[str], list[str], list
333
316
  )
334
317
 
335
318
  # conda
336
- if all(
337
- x.exists()
338
- for x in [
319
+ if all(x.exists() for x in [
339
320
  base_path / "bin" / "ptxas.exe",
340
321
  base_path / "include" / "cuda.h",
341
322
  base_path / "lib" / "cuda.lib",
342
- ]
343
- ):
323
+ ]):
344
324
  return (
345
325
  str(base_path / "bin"),
346
326
  [str(base_path / "include")],
@@ -348,14 +328,11 @@ def check_and_find_cuda(base_path: Path) -> tuple[Optional[str], list[str], list
348
328
  )
349
329
 
350
330
  # bundled or system-wide
351
- if all(
352
- x.exists()
353
- for x in [
331
+ if all(x.exists() for x in [
354
332
  base_path / "bin" / "ptxas.exe",
355
333
  base_path / "include" / "cuda.h",
356
334
  base_path / "lib" / "x64" / "cuda.lib",
357
- ]
358
- ):
335
+ ]):
359
336
  return (
360
337
  str(base_path / "bin"),
361
338
  [str(base_path / "include")],
@@ -372,9 +349,7 @@ def find_cuda_env() -> tuple[Optional[str], list[str], list[str]]:
372
349
  continue
373
350
 
374
351
  cuda_base_path = Path(cuda_base_path)
375
- cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs = check_and_find_cuda(
376
- cuda_base_path
377
- )
352
+ cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs = check_and_find_cuda(cuda_base_path)
378
353
  if cuda_bin_path:
379
354
  return cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs
380
355
 
@@ -382,9 +357,7 @@ def find_cuda_env() -> tuple[Optional[str], list[str], list[str]]:
382
357
 
383
358
 
384
359
  def find_cuda_bundled() -> tuple[Optional[str], list[str], list[str]]:
385
- cuda_base_path = (
386
- Path(sysconfig.get_paths()["platlib"]) / "triton" / "backends" / "nvidia"
387
- )
360
+ cuda_base_path = (Path(sysconfig.get_paths()["platlib"]) / "triton" / "backends" / "nvidia")
388
361
  return check_and_find_cuda(cuda_base_path)
389
362
 
390
363
 
@@ -408,23 +381,21 @@ def find_cuda_hardcoded() -> tuple[Optional[str], list[str], list[str]]:
408
381
  paths = sorted(paths)[::-1]
409
382
  for cuda_base_path in paths:
410
383
  cuda_base_path = Path(cuda_base_path)
411
- cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs = check_and_find_cuda(
412
- cuda_base_path
413
- )
384
+ cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs = check_and_find_cuda(cuda_base_path)
414
385
  if cuda_bin_path:
415
386
  return cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs
416
387
 
417
388
  return None, [], []
418
389
 
419
390
 
420
- @functools.cache
391
+ @functools.lru_cache
421
392
  def find_cuda() -> tuple[Optional[str], list[str], list[str]]:
422
393
  for f in [
423
- find_cuda_env,
424
- find_cuda_bundled,
425
- find_cuda_pip,
426
- find_cuda_conda,
427
- find_cuda_hardcoded,
394
+ find_cuda_env,
395
+ find_cuda_bundled,
396
+ find_cuda_pip,
397
+ find_cuda_conda,
398
+ find_cuda_hardcoded,
428
399
  ]:
429
400
  cuda_bin_path, cuda_inc_dirs, cuda_lib_dirs = f()
430
401
  if cuda_bin_path:
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: triton-windows
3
- Version: 3.3.1.post19
3
+ Version: 3.5.0.post21
4
4
  Summary: A language and compiler for custom Deep Learning operations
5
5
  Home-page: https://github.com/woct0rdho/triton-windows
6
6
  Author: Philippe Tillet, Dian Wu
@@ -10,14 +10,16 @@ Classifier: Development Status :: 4 - Beta
10
10
  Classifier: Intended Audience :: Developers
11
11
  Classifier: Topic :: Software Development :: Build Tools
12
12
  Classifier: License :: OSI Approved :: MIT License
13
- Classifier: Programming Language :: Python :: 3.9
14
13
  Classifier: Programming Language :: Python :: 3.10
15
14
  Classifier: Programming Language :: Python :: 3.11
16
15
  Classifier: Programming Language :: Python :: 3.12
17
16
  Classifier: Programming Language :: Python :: 3.13
18
- Requires-Dist: setuptools>=40.8.0
17
+ Classifier: Programming Language :: Python :: 3.14
18
+ Requires-Python: >=3.10,<3.15
19
+ License-File: LICENSE
20
+ Requires-Dist: importlib-metadata; python_version < "3.10"
19
21
  Provides-Extra: build
20
- Requires-Dist: cmake>=3.20; extra == "build"
22
+ Requires-Dist: cmake<4.0,>=3.20; extra == "build"
21
23
  Requires-Dist: lit; extra == "build"
22
24
  Provides-Extra: tests
23
25
  Requires-Dist: autopep8; extra == "tests"
@@ -37,6 +39,8 @@ Dynamic: author-email
37
39
  Dynamic: classifier
38
40
  Dynamic: home-page
39
41
  Dynamic: keywords
42
+ Dynamic: license-file
40
43
  Dynamic: provides-extra
41
44
  Dynamic: requires-dist
45
+ Dynamic: requires-python
42
46
  Dynamic: summary
@@ -0,0 +1,217 @@
1
+ triton/__init__.py,sha256=3k-6i-O8Fi4127sAdQdTs0yN6rNV_blAuYY2lIUQzq4,1661
2
+ triton/_filecheck.py,sha256=2i88UORAasMECwAcW5NlCfJbYae7cVd2zOhAoGjI_uc,3164
3
+ triton/_internal_testing.py,sha256=LMdOSxV1JUDQQSqV__YsfcDab0E5yrUm0KSwPLghCp8,8405
4
+ triton/_utils.py,sha256=JjqxiYSZ4GSNTkOHWS6tJ6ME114bGahRpQfuH79Dv08,3595
5
+ triton/errors.py,sha256=8WfnuRKLG578mgY6cBA3ECruVMf9ULEKFNgRcJ6IhWM,89
6
+ triton/knobs.py,sha256=QYsHVVRBItf4XMCwY-xjpjnrVbmRP1ysk6X2phK3xJo,16713
7
+ triton/testing.py,sha256=vbEQRNrOnnzRQvVVSaiZrUo8AC0XPV40GJxfvkKYLh0,20276
8
+ triton/windows_utils.py,sha256=JMi6mjOApzh2-cw3Wl_nl6ji7JkwexYI7xgo2Et3ihU,12903
9
+ triton/_C/libtriton.pyd,sha256=-fZkCNJfdVoXLY3gaX04zp0P1OnXPjF90vHRo0FeSnA,100470784
10
+ triton/backends/__init__.py,sha256=X7290kf96Fk9QnfLScsX4UDG3zPyH_-31E4A7pVOijM,1612
11
+ triton/backends/compiler.py,sha256=MY2_cQG26p68z8VwRv2Nlj_h2DfEhwBbN-30caMgep0,2840
12
+ triton/backends/driver.py,sha256=JcsL7pVHSIM13i1JWvc4KYum1MIPHTeUHcF3UcV6IR4,1802
13
+ triton/backends/amd/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
14
+ triton/backends/amd/compiler.py,sha256=A7_8wWyda-Xqj95kXK86bjwKyJLuxJBRp51MyZtmucc,20874
15
+ triton/backends/amd/driver.c,sha256=dEpw00lFUSMSvQ5AnNc_4XDAZIHVXvUlJ5K-LmAxkvU,11149
16
+ triton/backends/amd/driver.py,sha256=5oUFCY7vPnwLBQy5HaJYerbBDUVcoLlB_wb6gGjhYwU,26951
17
+ triton/backends/amd/lib/asanrtl.bc,sha256=1xv2RlU3WvbdsghHlmhwiHewGM2B5dKts5bERM6S89o,24508
18
+ triton/backends/amd/lib/ockl.bc,sha256=wQKCzkKukIHbu0lyjKUYlhndc7S27xto6L54J0Bn-C0,246124
19
+ triton/backends/amd/lib/ocml.bc,sha256=UPNTXW0gCXUNB-c6orSYwb-mz9_mjUc7zny_vfFza44,205964
20
+ triton/backends/nvidia/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
21
+ triton/backends/nvidia/compiler.py,sha256=PyWopfTQPdkOyEU1CuNcPTzWPpbOrxmj0l8kiJNZ3uU,22064
22
+ triton/backends/nvidia/driver.c,sha256=G-Xggg8ICQoWXIBA0y_ehRUpR40Vjde5RfkKguoTNl0,19546
23
+ triton/backends/nvidia/driver.py,sha256=biC-1tsQ8MbXT2eIGLY15SLN5siLYrCaTuKpy5qHoEI,26706
24
+ triton/backends/nvidia/bin/ptxas.exe,sha256=f28E0l5aerLAfBXk7yagfOwIEE6_6_NkMx-vqPPEQ9Y,24753152
25
+ triton/backends/nvidia/include/cuda.h,sha256=Fn44OjeRImxegJ39apYUspseEfTWNGwpqSGUOnHj5WY,1183268
26
+ triton/backends/nvidia/lib/libdevice.10.bc,sha256=XC-uN8huaMOjhgWpX1EtfRLV89uYYxC-R_VzBKpype4,473728
27
+ triton/backends/nvidia/lib/x64/cuda.lib,sha256=I5DZfR8aQ9wodYo3trskSbJpJd9lHvZXsnEZ3NV30LQ,160840
28
+ triton/compiler/__init__.py,sha256=S0iIXHTRJL8MUqThupDtPyJVT0PA882c2dN5VwqOGxE,284
29
+ triton/compiler/code_generator.py,sha256=xNeZwy62o7kq02rX6lZF78Wy_GzVxz2tw6Z41X3SzPQ,71445
30
+ triton/compiler/compiler.py,sha256=Ts8gmfq527KMKPin8IRBtuZVlrUHTpL1gZQMfZ6f3Gs,21147
31
+ triton/compiler/errors.py,sha256=I9Y15pDWcL9heY4SWWdLeMDtW6Iiq2pFXzKfJ6dY_C0,1732
32
+ triton/compiler/make_launcher.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
33
+ triton/experimental/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
34
+ triton/experimental/gluon/__init__.py,sha256=GO2iamrLyIbBszA44Z3Aa-V2v4S0kq4TOKPWrRx0FYM,186
35
+ triton/experimental/gluon/_compiler.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
36
+ triton/experimental/gluon/_runtime.py,sha256=AVoWQuzXSKD02gZLAzbZ6iHw4uccf0rH-KpWVH-7ceI,3461
37
+ triton/experimental/gluon/language/__init__.py,sha256=HR1Tp5k-GhrKx_-7Jt0JCoTqlhQloIFMrUiHoYJIQNY,1765
38
+ triton/experimental/gluon/language/_core.py,sha256=1ACDrOT_X7MJZEoYAyOJ7GFANUs3FKdwO4guC9JUyKs,15695
39
+ triton/experimental/gluon/language/_layouts.py,sha256=_PzL_sL8Db6VF5qD5s6T08CxEKplCCySyYX9JWt6rhs,22740
40
+ triton/experimental/gluon/language/_math.py,sha256=MN9Fvt8FaQTpxEUCKhw83IGbkwBD2Whdp686av4vv04,564
41
+ triton/experimental/gluon/language/_semantic.py,sha256=imNOS16WVaniSo20RT_fVMYk7PjZ-lWpNM1bn1YikU8,18062
42
+ triton/experimental/gluon/language/_standard.py,sha256=9z0ZCxxa2jQjsk9pCaofNOuQT8o3WWFoW5xJGtxbio4,2718
43
+ triton/experimental/gluon/language/amd/__init__.py,sha256=x7o_QGewXOYfxv1IiCjteHLWP_7975a-EX11naLcGcI,110
44
+ triton/experimental/gluon/language/amd/_layouts.py,sha256=DLSoxygfOGuoaICPjF3GmRzEYeqAFIzFgjQuUTlFknw,4489
45
+ triton/experimental/gluon/language/amd/cdna3/__init__.py,sha256=4Yl_KgH5hE35dOQdEE0M9DLUivGxRgVend4KU1E1SV0,4235
46
+ triton/experimental/gluon/language/amd/cdna4/__init__.py,sha256=qumwcZI7tvee7qLRg-Ft_J78tgU9Wzth1OdIQsYXtDs,2157
47
+ triton/experimental/gluon/language/amd/cdna4/async_copy.py,sha256=N_jqFSMieyVhrJEKx8TkrF5R1ELn0ghRXmv1Wt4adnw,7222
48
+ triton/experimental/gluon/language/extra/__init__.py,sha256=FaDqvKRb8M8K47C3YceyfF3ZiZayLC8xZN248jNlxeQ,69
49
+ triton/experimental/gluon/language/nvidia/__init__.py,sha256=SFBuACK5P2XoYcutHEnKjqgRTboU4CPDmJz0hT6dFRQ,80
50
+ triton/experimental/gluon/language/nvidia/ampere/__init__.py,sha256=4hNkRbEZ4LhblmwMqyriZdI_a8SEEaNNRl__fv5p-I0,73
51
+ triton/experimental/gluon/language/nvidia/ampere/async_copy.py,sha256=YfB0AN1hcEKWU2WTyef4ggWtp-HBnoSHJGoWhUTMmis,2817
52
+ triton/experimental/gluon/language/nvidia/ampere/mbarrier.py,sha256=CraFTt4ykqhz8AaWLt6x9AQbGX5RjLGGGzym7LV3jaY,2682
53
+ triton/experimental/gluon/language/nvidia/blackwell/__init__.py,sha256=ni_x_NTkwEukCjH6iaDiEYowYDllXv29uLDIpfj7Okk,14705
54
+ triton/experimental/gluon/language/nvidia/blackwell/tma.py,sha256=f6u_qBCYBtwls7FVQDdROymUed_NefREQwAyF4k4XSs,2007
55
+ triton/experimental/gluon/language/nvidia/hopper/__init__.py,sha256=gJpoqWWiJ2h4uCTFh5dYub90CamRsi4_NZQi47rspNw,5351
56
+ triton/experimental/gluon/language/nvidia/hopper/mbarrier.py,sha256=NF2dvXMKWzQ0mtWZQpCWKJKZzkINCZ24d8cKdqDT60Q,1339
57
+ triton/experimental/gluon/language/nvidia/hopper/tma.py,sha256=UK2DeQqKgHaKtTWBXUAmlv2a5MS2jrd3okPsUSrheXQ,3588
58
+ triton/experimental/gluon/nvidia/__init__.py,sha256=ISXB4RV7RcCLsU-JhcRFeA29gCBDVk8cTwO2j99ivLc,80
59
+ triton/experimental/gluon/nvidia/blackwell.py,sha256=cllwlUCE5_YKWqySQZk7wt7Fierz345E5VwztxNRGMs,69
60
+ triton/experimental/gluon/nvidia/hopper.py,sha256=GkNztKn3bJWx-WnLv08Dkmchu1Nh3okTpzSEsCl6_Ew,1829
61
+ triton/language/__init__.py,sha256=qlEvCv0KojmEFCNiUwv6kLnijJTRm3OnBPsHScueu74,6563
62
+ triton/language/core.py,sha256=XBKBtnwNlJ2jTM8cO2iASiyqJmWVOYTYN9dT6Rr2Ef4,119331
63
+ triton/language/math.py,sha256=CKvuIc5iMKhz7Qgx9w-VcLfOOZadv5svKK4aGZLuHMc,7399
64
+ triton/language/random.py,sha256=jkuFmfgZ8yvKuub9EY27zPvsC6nhkJIk05xf4y-7SR8,7102
65
+ triton/language/semantic.py,sha256=PbS-6UJ8frDiEFr_SEvazIGTV9KY6PbonDcq6cNiW98,99927
66
+ triton/language/standard.py,sha256=Wc8pXzqrT6YrlqYnqp_sEwKVATN1DeWeVCx7TJ5bgLY,15971
67
+ triton/language/target_info.py,sha256=UwWJDDsTPe9E4vPTQF4Psf1s540KKLISdnChlbvhHGY,1364
68
+ triton/language/extra/__init__.py,sha256=XRXFvr7416pRsh_Rh-X6qV66SiEyVDVbxp4GSAE1mfc,655
69
+ triton/language/extra/libdevice.py,sha256=9gwrstjuvhwZH0uLgZhEQU3RvOJDZFY1oJT5-SZlpY4,6350
70
+ triton/language/extra/cuda/__init__.py,sha256=MBBu2EUYxsp6ygjiwO4Yh1X1EswMstfaiRTMSMGtbcw,407
71
+ triton/language/extra/cuda/gdc.py,sha256=rbVOcdD_w72sQ3-8l5KXafu-Tr0D5jYcY1fyok98WVQ,2193
72
+ triton/language/extra/cuda/libdevice.py,sha256=J7Kl0ejbAIus7-YBn2OSK71lkm3pC7G1J-5ZdHfS82U,56764
73
+ triton/language/extra/cuda/utils.py,sha256=phDcXCFViaq3p4ThwHrO8-FtU-8A8I3nk4mZZJVvTio,4426
74
+ triton/language/extra/hip/__init__.py,sha256=GFcuM-R0qCB5kDbysvjD5U2KUmOGNWcWzM8r8yqDI3Y,96
75
+ triton/language/extra/hip/libdevice.py,sha256=LNynTEKXgE3PSZp-nUMVq4ft_i1y9Tao_cAvKKPn2s0,17542
76
+ triton/language/extra/hip/utils.py,sha256=tTq-k0c-s32IvDeGkgZcmosb9UtkF8sZEaE9EJGMbSA,877
77
+ triton/runtime/__init__.py,sha256=mKL5cqIBDUw2WO80NRCh4s1G8KYaqgM59TTAbTkPPjQ,621
78
+ triton/runtime/_allocation.py,sha256=jJdUXuwt3XzAcJQibYx6gB-w2Nka2kTgW1iaQ6qfkXQ,1220
79
+ triton/runtime/_async_compile.py,sha256=57arJTL6OkUf6RHQ9SKOKHdgqWTlkNrnyT2fYR4sdjA,1703
80
+ triton/runtime/autotuner.py,sha256=1tAq_HXytm0Sxruabs70H6AwaK4FvvJvVPny0E2r6jA,20176
81
+ triton/runtime/build.py,sha256=dOOLG_9nkndF2Jp0lTPWth7GNpe4l7vgaZqMz9hc-Hs,6394
82
+ triton/runtime/cache.py,sha256=UXc6Z6M3tC033znY_xmyPg2-fh4kVyBPKEe0kVDxDqg,11364
83
+ triton/runtime/driver.py,sha256=2qxsJH3SFxYDSIffj2wXMGSviEatUQLkHFcmVboeK-A,1025
84
+ triton/runtime/errors.py,sha256=CwfJXciwel_-K3BfQfKUpLPDWrSyTnGsfJkqJojrdfQ,1052
85
+ triton/runtime/interpreter.py,sha256=KqlylwYSzky8QEOQvZhIEd6AGhzOgyPFPLD_uTlXIcM,61789
86
+ triton/runtime/jit.py,sha256=6oMeA-CGdWVMlKIeLmkXraS3cul_Jmws4UHnpmRe4W0,41822
87
+ triton/runtime/tcc/libtcc.dll,sha256=Du_3YgojwHcLb5W5BXKhDjsseNfE-WyiB8Bowv-kEiw,372224
88
+ triton/runtime/tcc/tcc.exe,sha256=Q27YjX2NXQvOYLvrO7OA4KWzSOwOJMyV5mLLOs-sxnA,25088
89
+ triton/runtime/tcc/include/_mingw.h,sha256=tu2thJzkHosjeXB6z57csya59I89-753VKoVVpeNUfA,3901
90
+ triton/runtime/tcc/include/assert.h,sha256=oLcm1_gr6sDP-1UMM-rY8jGG_ZQd5SFvpI2peiWZVlA,1641
91
+ triton/runtime/tcc/include/conio.h,sha256=3nFh-Fg12Ys4_moZ74lz3K9Y7CN7HJHPBaxTWy_zhF8,11130
92
+ triton/runtime/tcc/include/ctype.h,sha256=TPqqQ7P3QUmEEm6LHN9l-drA72jZozlr4LiCg3anSms,9755
93
+ triton/runtime/tcc/include/dir.h,sha256=hAZLF-UB1pHEPUfkWxEsKITbRnQXkQtfoUgrcjQrrfs,952
94
+ triton/runtime/tcc/include/direct.h,sha256=F5wyBDEtfPgDIQJ3Nim8s-X_95LR2AiTHLZhmkMdJDU,1979
95
+ triton/runtime/tcc/include/dirent.h,sha256=iMH3Z_3NbVG5ke4yNHktpIyFdvX4gW8XpCNE-ci7scE,3339
96
+ triton/runtime/tcc/include/dos.h,sha256=Ov6kroXGiYf-WfQFkqxeo-8QSbT7cmErsYU1jWKOLew,1090
97
+ triton/runtime/tcc/include/errno.h,sha256=3ay4jDJbCdbHSC5EbId8TQEyiijYAzMso4xUtCjRuLA,1403
98
+ triton/runtime/tcc/include/excpt.h,sha256=AJjlFgLJT4qXAvS3dtNjD1buwn7We5_DbZIEkztYrE0,3796
99
+ triton/runtime/tcc/include/fcntl.h,sha256=-9lPlFpXFlrIl726zSqGGxNR54UPp2dScDwKYi4GRvo,1269
100
+ triton/runtime/tcc/include/fenv.h,sha256=Z6gnrPTgllOvtdGPLsql_N-3Rx2KW4GXwvM9BuhGL4Q,3146
101
+ triton/runtime/tcc/include/float.h,sha256=SE_-6V_NvADNwVS53C_gP-1lYQvxpzthCHHMOWmOb44,1930
102
+ triton/runtime/tcc/include/inttypes.h,sha256=QZTAQIzbozC3z6HSCR1yoM-_IHf_H-sZ9Dbz86oq3xg,6072
103
+ triton/runtime/tcc/include/io.h,sha256=Xal8hQ6OKrYIxClHozQR9Vb211uCZOHlzynKe6e5YlY,13067
104
+ triton/runtime/tcc/include/iso646.h,sha256=IPXpPO9ynQKTDRQcI62b-UqPeMQsjTbrBIdNjRlYxsE,548
105
+ triton/runtime/tcc/include/limits.h,sha256=OV2L9y7ZG4PVEiNAia6KltiiHnL1_cvVavSu9uERDGI,2771
106
+ triton/runtime/tcc/include/locale.h,sha256=iCYm-iXbwbWQPm_ZjMhRbx5UxOBpRQJmU_BbOBJd_yw,2135
107
+ triton/runtime/tcc/include/malloc.h,sha256=t5_DmoCe3pppkOXWGuIeEkc3gKsdlYDnTFO4c4KlJ_Y,5208
108
+ triton/runtime/tcc/include/math.h,sha256=h9meUyH0mSmbtAXP3RYrb7kFx6L1rAN_Ki1B6m59JkI,15285
109
+ triton/runtime/tcc/include/mem.h,sha256=Q5VpRq7-5Q4B_dTVSmxZdBirywIlH512le1wOf16X_Y,345
110
+ triton/runtime/tcc/include/memory.h,sha256=tDQeGIkTqBn6O_EBB4qVygd3gCGTc_Qkw5rYbJTgS28,1233
111
+ triton/runtime/tcc/include/process.h,sha256=aUfJVPKvZ25mzDjWSxoWVChzQADi4nL4g8LXSoW4ICA,9091
112
+ triton/runtime/tcc/include/setjmp.h,sha256=thZOt_rkoSFjJRSS9_TlbMUNFG7HovVkDYbspNCVBG8,3867
113
+ triton/runtime/tcc/include/share.h,sha256=bekiwb1-7cMzCDBHhcISlFBk12Pu37NzwJy7tcuTPd4,639
114
+ triton/runtime/tcc/include/signal.h,sha256=BFoDGzdnM-16aFvAFwn1KBQDcp_3xgG5E7Ksov4Uk7s,1583
115
+ triton/runtime/tcc/include/stdalign.h,sha256=dhLUZXEJnMqnYWpRD3jxgMsekWcebl6SI6Lil7hLeJ0,354
116
+ triton/runtime/tcc/include/stdarg.h,sha256=o4-NNPXglljMOoiSs6foD_Vm6u7cGU5ahezgtnWZMTc,335
117
+ triton/runtime/tcc/include/stdatomic.h,sha256=6rIO2NohI5Y2lH3GF4pNqTt4jJ-F0mNtcA_kf3nX77E,7882
118
+ triton/runtime/tcc/include/stdbool.h,sha256=UlKCQiXdxIawRgZ392XkFXr10-16zWWzEKQEXq-1avc,176
119
+ triton/runtime/tcc/include/stddef.h,sha256=zGcsJknuyWbrBOspycnyf_ef1lcnaKGVPALVyvGFFSk,1144
120
+ triton/runtime/tcc/include/stdint.h,sha256=M8bI2n1WS1cCr4xv9FwAoWhCuj_-P5X39iMnUvY8Wv0,6333
121
+ triton/runtime/tcc/include/stdio.h,sha256=dJBZg0FDvNW9zqE_yGPItlh6idbfyEzVAXqY3xkN770,14903
122
+ triton/runtime/tcc/include/stdlib.h,sha256=iNeirVc03IVWk5FIs4SePXYYhkcBwH9YInrJt1uc8Ko,20730
123
+ triton/runtime/tcc/include/stdnoreturn.h,sha256=7GTY-_f7LoAN-XhLLvygqZ_CWj616NpW5N8JiTf3h_I,125
124
+ triton/runtime/tcc/include/string.h,sha256=1cAsImU3hHku7_BMxFNGe6IsIU2azodhJ-q1_My8p2I,8590
125
+ triton/runtime/tcc/include/tccdefs.h,sha256=VU3qfoZAm9k9_Y7UclxmoIoTnwhAMcCnjbcEz6mZrSA,11792
126
+ triton/runtime/tcc/include/tcclib.h,sha256=FR8_fIZXa9SS0lT4QOI02XmTxG1NmoJhTX0JU8AFdeE,2611
127
+ triton/runtime/tcc/include/tchar.h,sha256=f-X94Cj_j2nSvakQZk4sFp57ksbn8s95FetyBUqXRv8,31364
128
+ triton/runtime/tcc/include/tgmath.h,sha256=5gxm_5AX8TqxjuOIJHksp3G5I1vLTfV2iP7FgnOVJKU,3954
129
+ triton/runtime/tcc/include/time.h,sha256=RTeTotbG_HctHN1g5wH7PTk9dSk3wdayymTV8c7J_TY,8405
130
+ triton/runtime/tcc/include/uchar.h,sha256=JzBUqGpmEgZViyuFlVggixRTh2Knl5PRz13jfQgJt9g,954
131
+ triton/runtime/tcc/include/unistd.h,sha256=Bo_XR9pEwPjqcqEpnFYRyXnPf0wdNveEQ5tdpmMtJzw,24
132
+ triton/runtime/tcc/include/vadefs.h,sha256=LmqzWVWTGaEagPj1KqBHLNCxQRN_Oh6qGMQNiCfcUdQ,304
133
+ triton/runtime/tcc/include/values.h,sha256=-jdYhHsz9Zq-mbAjvgDYoCfDkezQWAof51VJfBHgxyM,126
134
+ triton/runtime/tcc/include/varargs.h,sha256=B4WIV_Tu0KYd-UvrGp1ni1P8PWegsOiTYVX4XdvNHcw,355
135
+ triton/runtime/tcc/include/wchar.h,sha256=yb8S4CoqsHg-0cZt_kPeQ8QCszkGytqbEVdQKoLHw-Q,34132
136
+ triton/runtime/tcc/include/wctype.h,sha256=4S2cW8vk37lupsdUEOooeRezwkv_nNLnFtNeAMHUkGw,4782
137
+ triton/runtime/tcc/include/sec_api/conio_s.h,sha256=9zdegWc5SR-6s5UxwdYKd7eP-aFiq6F_gXx3O_dfZQg,1535
138
+ triton/runtime/tcc/include/sec_api/crtdbg_s.h,sha256=A2MO6D58khRGoHkIU_yt61owhVPdPE7N3VaM2jFnwPE,461
139
+ triton/runtime/tcc/include/sec_api/io_s.h,sha256=bgL0rlDTBimvffNHhbbDJkKxLZSt3VZgb2_Eq2aCUP8,881
140
+ triton/runtime/tcc/include/sec_api/mbstring_s.h,sha256=2WKrgHCViVP0iyTJ6gaLNFsVgjeCb7cbmnbTbPLooys,3615
141
+ triton/runtime/tcc/include/sec_api/search_s.h,sha256=lhdOCfHFc8f66oWm1WgiWhuUbhM8bASnvWqoZcWIlqI,796
142
+ triton/runtime/tcc/include/sec_api/stdio_s.h,sha256=UQe-10DGJ0_8dnrULe1s5aj1HLDHMjnQTVpkfWLt8vE,11508
143
+ triton/runtime/tcc/include/sec_api/stdlib_s.h,sha256=hlDjS-JBx9g3QzEmh462ow7nHAt1nCNnH9jwcVx83mU,4113
144
+ triton/runtime/tcc/include/sec_api/stralign_s.h,sha256=sj8M951UVeIy2SeS4rK-OBJaAoCLwAUEk2e6to2hMAs,824
145
+ triton/runtime/tcc/include/sec_api/string_s.h,sha256=7vMvtQW5ijYQkj6N2z3nJMVbRDidJc73z1DuPNFPXWg,1787
146
+ triton/runtime/tcc/include/sec_api/tchar_s.h,sha256=pTGMzrJBlidpFpwyo85b-5oHWlLtusMarTOw17iXtUQ,8328
147
+ triton/runtime/tcc/include/sec_api/time_s.h,sha256=munLejFkrQCT44h7DKCbtnSY2lG7ROm-UAtg5yo4XcA,2331
148
+ triton/runtime/tcc/include/sec_api/wchar_s.h,sha256=WkcKw1iy2VEgIYL57B-UUzHCOo15YprU7bCLfXPPruQ,7492
149
+ triton/runtime/tcc/include/sec_api/sys/timeb_s.h,sha256=0VNBfsZOt7FQR0m8pkd-_VG0si3mcFGPT9wnAQgBRcA,653
150
+ triton/runtime/tcc/include/sys/fcntl.h,sha256=K09mD_2JlK-gOHQHBR48p-zI_kS-sq3S1DHNUs6K2cQ,359
151
+ triton/runtime/tcc/include/sys/file.h,sha256=tud5xTFAwRe8Nr0zXGS_yxOuTCxIa5R4OzIUmm6y0yA,367
152
+ triton/runtime/tcc/include/sys/locking.h,sha256=KXhhRemvNKH5bnNohVsZ6IefyA01oXLZupfTx_wvYxE,648
153
+ triton/runtime/tcc/include/sys/stat.h,sha256=1_lxCw-C9HyjkcP9s-P4dorkaH3yvv3x47L8uTLvwGU,6875
154
+ triton/runtime/tcc/include/sys/time.h,sha256=UYdB8oZUVDTfZ2Vy5Tv4VTsElqcTiULcayD_JStCk-Q,1717
155
+ triton/runtime/tcc/include/sys/timeb.h,sha256=70P59RZgq4KCcH9xacw9l3h45iN0PSPsVlZj_itOl4I,2445
156
+ triton/runtime/tcc/include/sys/types.h,sha256=zjFiRYBx3786jqpJEGZsDqwfsk0mFFB47OyrrKY1ozI,2249
157
+ triton/runtime/tcc/include/sys/unistd.h,sha256=H1lahcru73OFoL2pSvUYlrIU7iYFZISvUDU-k5PeGSk,351
158
+ triton/runtime/tcc/include/sys/utime.h,sha256=UkMS4-ijJffVr8Id24_LzrhdRRF14H7xvq23-C-jaLM,3429
159
+ triton/runtime/tcc/include/tcc/tcc_libm.h,sha256=jeuSdgYhXEKsx0f_EPlnvg4N2HNXEDyBt1BiVkHG3ZU,18759
160
+ triton/runtime/tcc/include/winapi/basetsd.h,sha256=XJy6oWq_V0AO0xtJqrfuAVeI2-fTtY89U8hts4B91vA,5614
161
+ triton/runtime/tcc/include/winapi/basetyps.h,sha256=NIQu4zicsTpyorh-yTCq2__OiQbrMUgBgM_1Qcf0QTQ,2413
162
+ triton/runtime/tcc/include/winapi/guiddef.h,sha256=fbGx_kZRP1eKPHd8POMA2EA9Mfv7bQDqz_kyhtLtEpM,4165
163
+ triton/runtime/tcc/include/winapi/poppack.h,sha256=gclR4fuHqo9uiHGgcyd_HNHMubZvbvqSr_NbzQCmByY,282
164
+ triton/runtime/tcc/include/winapi/pshpack1.h,sha256=jrZ90jPVo4fW3BgUy262xt6aEjQ4-u_Ke0QmkcryMEk,285
165
+ triton/runtime/tcc/include/winapi/pshpack2.h,sha256=1FnL1UaSn9RJgNMsFoCo8XbXF86d8WL1xcRD39zMnkI,285
166
+ triton/runtime/tcc/include/winapi/pshpack4.h,sha256=zTuhJYpd2ccUh50-SZsCHIXumCfAa6wvwsHmd7WQlTE,285
167
+ triton/runtime/tcc/include/winapi/pshpack8.h,sha256=Mv57X_I4fJFq0TTvW1sKxnRH2g4NzPQFwxViqscY1tg,285
168
+ triton/runtime/tcc/include/winapi/qos.h,sha256=NrnAEHFOIcdJ2dGIBxh81Jfu4yhuP-ng8seYJYDeVt0,2055
169
+ triton/runtime/tcc/include/winapi/shellapi.h,sha256=VQtBWV4D4FBI5NeNsDIIlogBMdke3yZ7LCMLIoxP7Qw,1649
170
+ triton/runtime/tcc/include/winapi/winbase.h,sha256=4WqDqKDdrW8R1SS9rqpTDt95ntSod2k4ntZym3EuuX0,159922
171
+ triton/runtime/tcc/include/winapi/wincon.h,sha256=1EagModl5S-Rg9CIzc-YYw-0yMOME9Y2fQmtXiWctiY,14781
172
+ triton/runtime/tcc/include/winapi/windef.h,sha256=E36aQ6E25K4Zs6TIRAI8ahYRsjaFAANk9r4xQ9saTHU,5674
173
+ triton/runtime/tcc/include/winapi/windows.h,sha256=eyvZ3ahFwMO9jiar7-CWYM4jOGvCo3jBhevcncUIGTw,2168
174
+ triton/runtime/tcc/include/winapi/winerror.h,sha256=WQE0AAsbXE-3r7zFSkRaQiKNdBZKnoskQ00amT92hS4,150512
175
+ triton/runtime/tcc/include/winapi/wingdi.h,sha256=31k3rBgFsnq7oDJ30sNMrujLQ4ftuJStzXPmFyqfvZQ,121301
176
+ triton/runtime/tcc/include/winapi/winnls.h,sha256=ScB2ApvYYkVfPw1v5HEr4Zy-YMjCXdgdIKTM2rJ0X1A,29401
177
+ triton/runtime/tcc/include/winapi/winnt.h,sha256=YSge0Uf8W7VYi0lCss4jwBogP5vJfkYY0GaOZsZj3c0,192825
178
+ triton/runtime/tcc/include/winapi/winreg.h,sha256=bbuWncIekNkETavNGQJowbsz5EWGLOKkpTbppxNPpOs,13963
179
+ triton/runtime/tcc/include/winapi/winsock2.h,sha256=2rONhOHGvxD-yBv4Cz2TCyHhP9ftfhonV8J2tis4dLo,57132
180
+ triton/runtime/tcc/include/winapi/winuser.h,sha256=Q1a_zfUgnE7FjeSG4hc85LF-DOdaQisib93dGFl8mQU,179678
181
+ triton/runtime/tcc/include/winapi/winver.h,sha256=fYRmeOwqjHD4YwjPa-WF12CSTGIN_PtLBI9g2IV3tp0,5381
182
+ triton/runtime/tcc/include/winapi/ws2ipdef.h,sha256=lS1jv4914xIFURm5ZO4L7caJuROVn_WGsMfZGmhFcDI,332
183
+ triton/runtime/tcc/include/winapi/ws2tcpip.h,sha256=0V__ggtdG8onGTUhnrfG3dLKUSEwf8dqf132-480ngs,12443
184
+ triton/runtime/tcc/lib/cuda.def,sha256=c8XdUSeajwFv0jSgW1lDoiCcEauHygWrl8Ws4wRrB_g,15535
185
+ triton/runtime/tcc/lib/gdi32.def,sha256=NfGNAGtoQaOuBuheG0mv4AqlinvYm9aPNbSycCUP50A,5052
186
+ triton/runtime/tcc/lib/kernel32.def,sha256=ank2QlCtswD44SShHIX-0vqd5XQI6r0eu7HO3PszuM4,12882
187
+ triton/runtime/tcc/lib/libtcc1.a,sha256=QMlzkJsmkK91dki2ntBomthyr7EDkf4akOvkdArJV0g,56326
188
+ triton/runtime/tcc/lib/msvcrt.def,sha256=oB3JKUV0IE47w6R7bKFEnIlfef-SFeWC4jg6g-YUg7g,15880
189
+ triton/runtime/tcc/lib/python3.def,sha256=o8MijZkdWOnHNPU4fmJrT1qw3fWSlaegC969H_X5_VE,16173
190
+ triton/runtime/tcc/lib/python310.def,sha256=3ofvYJiv78p-4DljnpZVvy9D5SvSfF-H26MldEEVJlw,33539
191
+ triton/runtime/tcc/lib/python311.def,sha256=8rxqpShmaBhu2HzvmkjsJy7kxXjNDBtD3nHTa88QKqY,33820
192
+ triton/runtime/tcc/lib/python312.def,sha256=yIwbeXDcGnPfAjTo89oEojrlQtrW6aVPvGm7GMX1RNs,35695
193
+ triton/runtime/tcc/lib/python313.def,sha256=HcOu0MS-PFcj2_OskAKsXfv9w27GebWijs-rc-RXmQI,34658
194
+ triton/runtime/tcc/lib/python313t.def,sha256=LssGwvVZnHdyZV2HHGcAqhla1blHjulz8Q7Wft8ERrs,34774
195
+ triton/runtime/tcc/lib/python314.def,sha256=eyBv4CLxdCSXXPKKbHeAJxeZfRljZ-GXBtLsP5Jce4g,37714
196
+ triton/runtime/tcc/lib/python314t.def,sha256=YaXruw0YyKSgsjLzQ5qFcxTQBGB98C-WgfpdOx9gH1M,37936
197
+ triton/runtime/tcc/lib/python39.def,sha256=iv2tgW38gyRGVdJ126Za7jrtW-ZAH1gFLD62JHj1e4E,34237
198
+ triton/runtime/tcc/lib/python3t.def,sha256=vem4PTOKJ-w6B_UCZdGFCscddRA-XWRONNOTMdWkXYQ,18101
199
+ triton/runtime/tcc/lib/user32.def,sha256=EcYohyyDgmz9fLBoOR-vszLeJ2YkBUoNGvSnuXrkum0,10439
200
+ triton/tools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
201
+ triton/tools/build_extern.py,sha256=jCr-2hu3nLGBIJhCGUQ1jAyzLttughjkiPGEwRFjLR0,13673
202
+ triton/tools/compile.py,sha256=ECNKNHnf7RZLldrv3REeXBtsGUmcjU3xqNr9tnqxeJ8,8730
203
+ triton/tools/disasm.py,sha256=T9jiTkdK_0nI3R_4uea0zvfioYdcR-zIZwTfuucgw6g,5026
204
+ triton/tools/link.py,sha256=u7qtfZRLriZkAMEGNvj8YF-k1cthmLL7BwHYqBgT63E,11871
205
+ triton/tools/mxfp.py,sha256=YQdpBrGkOVNOtnLeRjMCeVFHWkSwUubGeWsItIjO8TU,11737
206
+ triton/tools/ragged_tma.py,sha256=2WiUvniV80LkRfA1v_uCf8uzHsLaeBy0r4ruxX8uKhc,3167
207
+ triton/tools/tensor_descriptor.py,sha256=CFrcp_oNIwJxTt8LGHycVYzygnNwqE1AQoaJkIr7cGY,1466
208
+ triton/tools/extra/cuda/compile.c,sha256=ug4F_NBvwo1YybPQuePgRddlYFuBEVv3M5l60XlJFoQ,2190
209
+ triton/tools/extra/cuda/compile.h,sha256=n9QKIFZTL4RSsiXtAxBP9XGSnxjyaevQQ9bBpwDsvAg,332
210
+ triton/tools/extra/hip/compile.cpp,sha256=aJUqzL3cLuVWMP0--eATyVOT84-eLjT8j6hmrP33dwQ,1858
211
+ triton/tools/extra/hip/compile.h,sha256=BIRh2lo4kKXWcsomk89DKx3sY3_XLMGoCyJ5HYvl3GU,348
212
+ triton_windows-3.5.0.post21.dist-info/licenses/LICENSE,sha256=kmQPuXIi_Qppj_KM4MN4LBcmI_jWxgm1V2NqgPKPuUY,1132
213
+ triton_windows-3.5.0.post21.dist-info/METADATA,sha256=3YoWhex1CJ-sUQq1-HDFXPUuywmeMNekRlkNGcEeAFQ,1761
214
+ triton_windows-3.5.0.post21.dist-info/WHEEL,sha256=JLOMsP7F5qtkAkINx5UnzbFguf8CqZeraV8o04b0I8I,101
215
+ triton_windows-3.5.0.post21.dist-info/entry_points.txt,sha256=cztF9ZYXxoMhibI_OttiKCl1EBP2LQaV8naJ-BcuES4,76
216
+ triton_windows-3.5.0.post21.dist-info/top_level.txt,sha256=WBiIZyv6n9Y7MIh-HPHSv2w1RDk7EFL__7ZgQRrmHYs,7
217
+ triton_windows-3.5.0.post21.dist-info/RECORD,,
@@ -0,0 +1,3 @@
1
+ [triton.backends]
2
+ amd = triton.backends.amd
3
+ nvidia = triton.backends.nvidia
@@ -0,0 +1,23 @@
1
+ /*
2
+ * Copyright 2018-2020 Philippe Tillet
3
+ * Copyright 2020-2022 OpenAI
4
+ *
5
+ * Permission is hereby granted, free of charge, to any person obtaining
6
+ * a copy of this software and associated documentation files
7
+ * (the "Software"), to deal in the Software without restriction,
8
+ * including without limitation the rights to use, copy, modify, merge,
9
+ * publish, distribute, sublicense, and/or sell copies of the Software,
10
+ * and to permit persons to whom the Software is furnished to do so,
11
+ * subject to the following conditions:
12
+ *
13
+ * The above copyright notice and this permission notice shall be
14
+ * included in all copies or substantial portions of the Software.
15
+ *
16
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
17
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
18
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
19
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
20
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
21
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
22
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
23
+ */
@@ -0,0 +1 @@
1
+ triton