Commit Graph

  • 1893079fdc WIP: Builder for expected test results. Ville Pietilä 2025-11-03 17:06:01 +00:00
  • b00303a831 Replace CK_TILE_PIPELINE macros with a common enum Emily Martins 2025-10-27 17:26:04 +00:00
  • 4ec9a97bfe Replace CK_TILE_PIPELINE macros with a common enum Emily Martins 2025-10-27 17:26:04 +00:00
  • 2ec57a8e70 Replace CK_TILE_PIPELINE macros with a common enum Emily Martins 2025-10-27 17:26:04 +00:00
  • bc26a7282b Merge commit 'afe1ff618df6fb28532331560f9b40a0b396a1da' into develop assistant-librarian[bot] 2025-11-03 16:13:52 +00:00
  • 15f010138b Fix namespace. Ville Pietilä 2025-11-03 16:06:14 +00:00
  • 699f7daae3 Ud fix moe sorting gfx908 (#2720) Michael Mcminn 2025-11-03 10:31:31 -05:00
  • d33b51181b Ud fix moe sorting gfx908 (#2720) Michael Mcminn 2025-11-03 10:31:31 -05:00
  • afe1ff618d Ud fix moe sorting gfx908 (#2720) Michael Mcminn 2025-11-03 10:31:31 -05:00
  • 6f61dd56c5 Optimize grouped conv bwd wei split_k off calc Bartlomiej Kocot 2025-11-03 15:10:55 +00:00
  • 0d68298d43 Separate test configs from other testing assets. Ville Pietilä 2025-11-03 14:39:54 +00:00
  • 9fc3d64e16 Rename test config file. Ville Pietilä 2025-11-03 14:33:56 +00:00
  • 6ded2ab92a Remove code duplication in fwd conv builder tests. Ville Pietilä 2025-11-03 14:06:45 +00:00
  • bc22b83b19 Add kUseTrLoad = false in non-trload pipeline Qianfeng Zhang 2025-11-03 12:40:16 +00:00
  • f72822d3a0 change squant to quant type ltqin 2025-11-03 12:31:45 +00:00
  • 33df038b64 Merge commit 'd405641f06162f2a6b1bf15f890caa7105beebe4' into develop assistant-librarian[bot] 2025-11-03 10:13:50 +00:00
  • 9b78a23bb7 Add workspace definition file. Ville Pietilä 2025-11-03 09:57:38 +00:00
  • fd15355261 Add validation rules for builder parameters. Ville Pietilä 2025-11-03 09:57:27 +00:00
  • aeaa457e75 Move instances assets to a dedicated directory. Ville Pietilä 2025-11-03 09:56:59 +00:00
  • e40ab20b9e Clarifying the using of CK_TILE_HOST and CK_TILE_HOST_DEVICE trying to save compiling time Qianfeng Zhang 2025-11-03 08:39:43 +00:00
  • 7c8d79af33 Ck tile engine gemm unit tests exapand test coverage (#3025) msaffari-amd 2025-11-03 10:29:16 +01:00
  • 6187719c40 Ck tile engine gemm unit tests exapand test coverage (#3025) msaffari-amd 2025-11-03 10:29:16 +01:00
  • d405641f06 Ck tile engine gemm unit tests exapand test coverage (#3025) msaffari-amd 2025-11-03 10:29:16 +01:00
  • ead8f4df80 Merge commit '3ae3992c18045446f1b733b306265efbd14c5d57' into develop assistant-librarian[bot] 2025-11-03 07:13:15 +00:00
  • aeeed60666 [CK_BUILDER] Add conv factories for DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle and DeviceGroupedConvFwdMultipleD_Wmma_CShuffle (#3138) Ville Pietilä 2025-11-03 09:03:25 +02:00
  • 651dc5343b [CK_BUILDER] Add conv factories for DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle and DeviceGroupedConvFwdMultipleD_Wmma_CShuffle (#3138) Ville Pietilä 2025-11-03 09:03:25 +02:00
  • 3ae3992c18 [CK_BUILDER] Add conv factories for DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle and DeviceGroupedConvFwdMultipleD_Wmma_CShuffle (#3138) Ville Pietilä 2025-11-03 09:03:25 +02:00
  • b84217c5a7 Merge commit '16e85cf179fd8e98f56d664642d37a6775d7bc4d' into develop assistant-librarian[bot] 2025-11-03 01:41:17 +00:00
  • 9f069d6e35 [CK_TILE] B matrix 2D block scale gemm (#3074) Sami Remes 2025-11-03 00:49:20 +00:00
  • fd7bc5e9ab [CK_TILE] B matrix 2D block scale gemm (#3074) Sami Remes 2025-11-03 00:49:20 +00:00
  • 16e85cf179 [CK_TILE] B matrix 2D block scale gemm (#3074) Sami Remes 2025-11-03 00:49:20 +00:00
  • a25f7cdeb8 fix the test file ThomasNing 2025-11-02 18:23:02 +00:00
  • c494b23725 support the change for blockscale 2d ThomasNing 2025-11-02 04:22:39 +00:00
  • 316be5c6b2 Merge commit '73f637894da54ac2014d3f7be675f1bf75a689c1' into develop assistant-librarian[bot] 2025-11-02 04:15:35 +00:00
  • f4b880d058 refactor: remove gemm preshuffle pipeline v1 by removing all references from codebase (#3132) Aviral Goel 2025-11-02 00:06:28 -04:00
  • 8302f9ec4a refactor: remove gemm preshuffle pipeline v1 by removing all references from codebase (#3132) Aviral Goel 2025-11-02 00:06:28 -04:00
  • 73f637894d refactor: remove gemm preshuffle pipeline v1 by removing all references from codebase (#3132) Aviral Goel 2025-11-02 00:06:28 -04:00
  • e31829384d Change in updating max_uih_seqlen in the example Qianfeng Zhang 2025-11-02 03:54:51 +00:00
  • 39cb8c33d1 Use supplement_array_by_last_element(num_targets, ) in example Qianfeng Zhang 2025-11-02 03:30:13 +00:00
  • 80e08b6efe Use supplement_array_by_last_element() in example to simplify the codes Qianfeng Zhang 2025-11-01 16:20:38 +00:00
  • 10133e5d51 Update to README.md Qianfeng Zhang 2025-11-01 13:16:50 +00:00
  • 8408ec0a02 Add scripts for testing the using of separate sequence lengths for k/v Qianfeng Zhang 2025-11-01 13:12:36 +00:00
  • 17e404be3b Support separate sequence lengths for q and kv Qianfeng Zhang 2025-10-31 14:04:32 +00:00
  • 6b4b6fb8a3 fix CMake ThomasNing 2025-11-02 01:49:49 +00:00
  • c09de8fa6f Merge commit '45be7415864b839cc27b0455bc6eae177b4832cf' into develop assistant-librarian[bot] 2025-11-01 20:11:51 +00:00
  • 5be796d8a5 fix: fix bug in print tile window when printing bf8/fp8 tiles (#3120) Aviral Goel 2025-11-01 15:28:07 -04:00
  • 69f7ade10b fix: fix bug in print tile window when printing bf8/fp8 tiles (#3120) Aviral Goel 2025-11-01 15:28:07 -04:00
  • 45be741586 fix: fix bug in print tile window when printing bf8/fp8 tiles (#3120) Aviral Goel 2025-11-01 15:28:07 -04:00
  • 420486464a Merge commit 'ab1a8356b6f0cd2a92392663d81c8e6ee78e4123' into develop assistant-librarian[bot] 2025-11-01 14:11:22 +00:00
  • b2aa37f3f5 Add 2GB limitation for grouped conv bwd weight (#3054) Bartłomiej Kocot 2025-11-01 14:16:45 +01:00
  • 515fb27488 Add 2GB limitation for grouped conv bwd weight (#3054) Bartłomiej Kocot 2025-11-01 14:16:45 +01:00
  • ab1a8356b6 Add 2GB limitation for grouped conv bwd weight (#3054) Bartłomiej Kocot 2025-11-01 14:16:45 +01:00
  • e065ebfb86 Merge commit '1fbb47ad304566a90a374cef4731f1a257e5e179' into develop assistant-librarian[bot] 2025-11-01 13:15:56 +00:00
  • 5f45985732 [CK TILE] Grouped conv fwd split image (#2970) JH-Leon-KIM-AMD 2025-11-01 14:18:16 +02:00
  • 35df1d1b79 [CK TILE] Grouped conv fwd split image (#2970) JH-Leon-KIM-AMD 2025-11-01 14:18:16 +02:00
  • 1fbb47ad30 [CK TILE] Grouped conv fwd split image (#2970) JH-Leon-KIM-AMD 2025-11-01 14:18:16 +02:00
  • 346ee26218 solve the merge conflict ThomasNing 2025-10-31 23:27:28 +00:00
  • 89be44d78b Add in Changlog and restructure the quant 2d example ThomasNing 2025-10-31 23:22:08 +00:00
  • 6f90564708 fix formatting Sami Remes 2025-10-31 20:16:52 +00:00
  • fe92102baf add some documentation and 2d block scale example Sami Remes 2025-10-31 20:13:43 +00:00
  • d560ad2092 Merge commit '8f1274d9b655c2584b3643acac07ef813f31238e' into develop assistant-librarian[bot] 2025-10-31 19:11:51 +00:00
  • 658fb530ab test(grouped_gemm): add unit tests for grouped_gemm bquant with preshuffleB true (#3119) Aviral Goel 2025-10-31 15:07:06 -04:00
  • d17d3f0766 test(grouped_gemm): add unit tests for grouped_gemm bquant with preshuffleB true (#3119) Aviral Goel 2025-10-31 15:07:06 -04:00
  • 8f1274d9b6 test(grouped_gemm): add unit tests for grouped_gemm bquant with preshuffleB true (#3119) Aviral Goel 2025-10-31 15:07:06 -04:00
  • 27dc4d9833 [CK TILE ENGINE] GEMM Multi D Restructure (#3121) Thrupti Raj Lakshmana Gowda 2025-10-31 14:02:46 -05:00
  • 10e844d93c [CK TILE ENGINE] GEMM Multi D Restructure (#3121) Thrupti Raj Lakshmana Gowda 2025-10-31 14:02:46 -05:00
  • a33d98f8e2 [CK TILE ENGINE] GEMM Multi D Restructure (#3121) Thrupti Raj Lakshmana Gowda 2025-10-31 14:02:46 -05:00
  • b7a073f769 [CK-tile] unhardcode the number of LDS banks from universal gemm policy (#3130) Max Podkorytov 2025-10-31 11:58:11 -07:00
  • 49500a1b3d [CK-tile] unhardcode the number of LDS banks from universal gemm policy (#3130) Max Podkorytov 2025-10-31 11:58:11 -07:00
  • 04efd282cf [CK-tile] unhardcode the number of LDS banks from universal gemm policy (#3130) Max Podkorytov 2025-10-31 11:58:11 -07:00
  • e6be7bcc2a WMMA gemm_add_relu_add_layernorm (#2989) Enrico Degregori 2025-10-31 19:19:26 +01:00
  • 71bd07a783 WMMA gemm_add_relu_add_layernorm (#2989) Enrico Degregori 2025-10-31 19:19:26 +01:00
  • 4ebc48a3cd WMMA gemm_add_relu_add_layernorm (#2989) Enrico Degregori 2025-10-31 19:19:26 +01:00
  • 96199abfbe Merge commit 'e9596228ff7f6ddb68fbd2f0f9e964cfb6af61cf' into develop assistant-librarian[bot] 2025-10-31 18:15:38 +00:00
  • 2136eddf8a Fix synchronization issue in fwd qr pipeline with dropout (#3135) Anton Gorenko 2025-10-31 21:44:52 +05:00
  • 4f47945979 Fix synchronization issue in fwd qr pipeline with dropout (#3135) Anton Gorenko 2025-10-31 21:44:52 +05:00
  • e9596228ff Fix synchronization issue in fwd qr pipeline with dropout (#3135) Anton Gorenko 2025-10-31 21:44:52 +05:00
  • 3d78c17295 Merge commit '5ed2046bee509cd907b9e609ae18a871864f1738' into develop assistant-librarian[bot] 2025-10-31 15:12:07 +00:00
  • a8a377ca53 Add the last two forward instance traits. (#3134) John Shumway 2025-10-31 07:52:42 -07:00
  • 46dd130e26 Add the last two forward instance traits. (#3134) John Shumway 2025-10-31 07:52:42 -07:00
  • 5ed2046bee Add the last two forward instance traits. (#3134) John Shumway 2025-10-31 07:52:42 -07:00
  • d2474f5396 Adding new alert failure patterns (#3122) andrew clark 2025-10-31 08:38:31 -06:00
  • 9950df2ae7 Adding new alert failure patterns (#3122) andrew clark 2025-10-31 08:38:31 -06:00
  • 1977e4b96a Adding new alert failure patterns (#3122) andrew clark 2025-10-31 08:38:31 -06:00
  • c6b0458d1d Add copyright notices to missing files (#3133) John Afaganis 2025-10-31 08:35:11 -06:00
  • a987c5dc2e Add copyright notices to missing files (#3133) John Afaganis 2025-10-31 08:35:11 -06:00
  • 3f996ee738 Add copyright notices to missing files (#3133) John Afaganis 2025-10-31 08:35:11 -06:00
  • b7429e620c Kabraham/fix block gemm v1 b scale (#3129) kabrahamAMD 2025-10-31 15:19:01 +01:00
  • abc0a0b77f Kabraham/fix block gemm v1 b scale (#3129) kabrahamAMD 2025-10-31 15:19:01 +01:00
  • a7c52e8afa Kabraham/fix block gemm v1 b scale (#3129) kabrahamAMD 2025-10-31 15:19:01 +01:00
  • 70ac1657a1 Merge commit 'c2d79314469f569c13c205ff5383f284c90d7445' into develop assistant-librarian[bot] 2025-10-31 13:20:09 +00:00
  • 2e1831f8fd [CK TILE] Clear output buffers for grouped conv bwd (#3127) Bartłomiej Kocot 2025-10-31 14:11:54 +01:00
  • fcabe28158 [CK TILE] Clear output buffers for grouped conv bwd (#3127) Bartłomiej Kocot 2025-10-31 14:11:54 +01:00
  • c2d7931446 [CK TILE] Clear output buffers for grouped conv bwd (#3127) Bartłomiej Kocot 2025-10-31 14:11:54 +01:00
  • 9c776aeb7c update tests SWDEV-561448 Yashvardhan Agarwal 2025-10-23 13:54:45 +00:00
  • 5efccadbd9 merge clear workspace with P0_v1 Yashvardhan Agarwal 2025-10-21 10:26:08 +00:00
  • 6769664197 add printf info wjx/mxfp4_moe_bpreshuffle_v1 mtgu0705 2025-10-31 05:01:10 -05:00
  • 3d0f3abf65 update valarLip 2025-10-31 07:57:49 +00:00
  • 051b930450 add d=192 ck_tile/fmha_in_fp8_async ltqin 2025-10-31 05:23:24 +00:00
  • a0dd3fc932 Merge commit 'e135dd518d19a36466ce7c61bb9d3203ec18c8af' into develop assistant-librarian[bot] 2025-10-31 03:32:13 +00:00