Commit Graph

  • 06d2e609cd Revert "Add attn sink (#2892)" (#3250) asleepzzz 2025-11-20 23:55:15 +08:00
  • d115b3be4a Revert "Add attn sink (#2892)" (#3250) asleepzzz 2025-11-20 23:55:15 +08:00
  • 5adaa201ed Revert "Add attn sink (#2892)" (#3250) asleepzzz 2025-11-20 23:55:15 +08:00
  • bd5135a83a fix bug for batch mode ltqin 2025-11-20 15:02:41 +00:00
  • c77462c090 Initial plan copilot/sub-pr-3236 copilot-swe-agent[bot] 2025-11-20 14:01:04 +00:00
  • d1c35e8426 Merge commit '9fa4e8d5ab0b80855b5aeafb2e7907302c1c004d' into develop assistant-librarian[bot] 2025-11-20 12:20:10 +00:00
  • f552cd7841 ref data copying Tianxing Wu 2025-11-20 11:34:39 +00:00
  • f4ba63deb7 Add attn sink (#2892) Linjun-AMD 2025-11-20 19:24:05 +08:00
  • cb7f05a8d3 Add attn sink (#2892) Linjun-AMD 2025-11-20 19:24:05 +08:00
  • 9fa4e8d5ab Add attn sink (#2892) Linjun-AMD 2025-11-20 19:24:05 +08:00
  • 4650aad069 Fix pre-hook commit error Kumar 2025-11-20 15:10:21 +05:30
  • 9f40865289 fix clang-format fail bf16x3_fix_1120 yingmaolu 2025-11-20 15:41:51 +08:00
  • 2dfc9991b8 fix:bf16x3:disable some instances temporarily yingmaolu 2025-11-20 15:30:57 +08:00
  • c0ee47d434 Fix clang format Kumar 2025-11-20 12:40:42 +05:30
  • c38a4ecc26 Add safe iteration Kumar 2025-11-20 12:27:10 +05:30
  • bc279fdf5f Merge branch 'develop' into rocking/fmha-fp8-pertensor rocking 2025-11-20 14:17:47 +08:00
  • 7e384ec9e2 Support hdim=256 rocking 2025-11-20 00:14:40 -06:00
  • dc2c613a9c Refactor code + cleaning work Kumar 2025-11-20 11:28:11 +05:30
  • fb2d106a2f Merge commit '84540edff312c18ba50c01e995774da37faa0a29' into develop assistant-librarian[bot] 2025-11-20 05:13:08 +00:00
  • 38605f5091 fix typo (#3244) Illia Silin 2025-11-19 20:23:09 -08:00
  • c36a154c66 fix typo (#3244) Illia Silin 2025-11-19 20:23:09 -08:00
  • 84540edff3 fix typo (#3244) Illia Silin 2025-11-19 20:23:09 -08:00
  • 809c1ead72 Merge commit '47e2ed838e3547bba1b48d3f559f20f46fd07b87' into develop assistant-librarian[bot] 2025-11-20 02:43:03 +00:00
  • 0d9f230577 [CK_TILE] Add Flatmm MX FP8 (#3208) Yi DING 2025-11-20 10:35:15 +08:00
  • e27e760d5a [CK_TILE] Add Flatmm MX FP8 (#3208) Yi DING 2025-11-20 10:35:15 +08:00
  • 47e2ed838e [CK_TILE] Add Flatmm MX FP8 (#3208) Yi DING 2025-11-20 10:35:15 +08:00
  • 158fec303c chore(copyright): update copyright header for test directory AviralGoelAMD 2025-11-19 15:38:14 +00:00
  • d9f0bdd5e3 chore(copyright): update copyright header for test directory AviralGoelAMD 2025-11-19 15:38:14 +00:00
  • 4e49e0228b chore(copyright): update copyright header for test directory AviralGoelAMD 2025-11-19 15:38:14 +00:00
  • ac0fb4fec5 [ck_tile] enable test grouped_gemm_quant and gemm_streamk on gfx12 (#3196) linqunAMD 2025-11-20 08:40:27 +08:00
  • 0739113989 [ck_tile] enable test grouped_gemm_quant and gemm_streamk on gfx12 (#3196) linqunAMD 2025-11-20 08:40:27 +08:00
  • d2e32b4305 [ck_tile] enable test grouped_gemm_quant and gemm_streamk on gfx12 (#3196) linqunAMD 2025-11-20 08:40:27 +08:00
  • 22e5874b05 Merge branch 'develop' into rocking/fmha-fp8-pertensor rocking 2025-11-20 05:51:49 +08:00
  • ca48bf3b98 Merge commit 'cd8af997e6d1fde6bc4397bd6ab4fca46510e776' into develop assistant-librarian[bot] 2025-11-19 21:11:39 +00:00
  • 4a5e7d098d [CK] s_prefetch unit test fixes. Michal Kulikowski 2025-11-10 11:19:37 +01:00
  • dd53cdad01 [CK] s_prefetch unit test fixes. Michal Kulikowski 2025-11-10 11:19:37 +01:00
  • cd8af997e6 [CK] s_prefetch unit test fixes. Michal Kulikowski 2025-11-10 11:19:37 +01:00
  • 8fc5eca798 [CK] Added s_prefetch unit test. -added s_buffer_load_b32/64 assembly -added amd_s_buffer_load_impl Michal Kulikowski 2025-11-05 14:09:04 +01:00
  • 6c23879329 [CK] Added s_prefetch unit test. -added s_buffer_load_b32/64 assembly -added amd_s_buffer_load_impl Michal Kulikowski 2025-11-05 14:09:04 +01:00
  • f3ef7acca0 [CK] Added s_prefetch unit test. -added s_buffer_load_b32/64 assembly -added amd_s_buffer_load_impl Michal Kulikowski 2025-11-05 14:09:04 +01:00
  • b9ee41c660 [CK_Builder ]fixed accidental drop of get_elementwise_operation during merge and added usage of get_elementwise_operation() to other builder instances (#3238) kabrahamAMD 2025-11-19 21:31:05 +01:00
  • 2e71ebe0b1 [CK_Builder ]fixed accidental drop of get_elementwise_operation during merge and added usage of get_elementwise_operation() to other builder instances (#3238) kabrahamAMD 2025-11-19 21:31:05 +01:00
  • 964f8e1f60 [CK_Builder ]fixed accidental drop of get_elementwise_operation during merge and added usage of get_elementwise_operation() to other builder instances (#3238) kabrahamAMD 2025-11-19 21:31:05 +01:00
  • 7ed276c492 Merge commit 'e6e2e04edbd5766afb388fc4ba64d57a9b52452e' into develop assistant-librarian[bot] 2025-11-19 18:15:38 +00:00
  • 8bcd57d8d4 [Inductor] Copy logic for ck-tile gemm instance configuration in Inductor max-autotune integration and test it (#2910) Max Podkorytov 2025-11-19 09:38:02 -08:00
  • 7098fd7442 [Inductor] Copy logic for ck-tile gemm instance configuration in Inductor max-autotune integration and test it (#2910) Max Podkorytov 2025-11-19 09:38:02 -08:00
  • e6e2e04edb [Inductor] Copy logic for ck-tile gemm instance configuration in Inductor max-autotune integration and test it (#2910) Max Podkorytov 2025-11-19 09:38:02 -08:00
  • 29d162bec7 Merge commit '7fe7aa76f52ad7bdb0bb08c1f2d1de468cc8c070' into develop assistant-librarian[bot] 2025-11-19 17:12:38 +00:00
  • 2dabd450c1 [CK_BUILDER] fixes (#3222) Robin Voetter 2025-11-19 18:05:25 +01:00
  • e0bdf4511d [CK_BUILDER] fixes (#3222) Robin Voetter 2025-11-19 18:05:25 +01:00
  • 7fe7aa76f5 [CK_BUILDER] fixes (#3222) Robin Voetter 2025-11-19 18:05:25 +01:00
  • abf4a7ea2f Merge commit '9837ba5af2d9a9fad3b5e7eddd871101c7402487' into develop assistant-librarian[bot] 2025-11-19 16:14:27 +00:00
  • 34c1fc5ae7 chore(copyright): update copyright header for tutorial directory (#3230) Aviral Goel 2025-11-19 10:20:53 -05:00
  • eceaba0da4 chore(copyright): update copyright header for tutorial directory (#3230) Aviral Goel 2025-11-19 10:20:53 -05:00
  • 9837ba5af2 chore(copyright): update copyright header for tutorial directory (#3230) Aviral Goel 2025-11-19 10:20:53 -05:00
  • 3c389eb3f1 Refactor Jenkinsfile (#3229) Illia Silin 2025-11-19 07:20:25 -08:00
  • c2247fa9a5 Refactor Jenkinsfile (#3229) Illia Silin 2025-11-19 07:20:25 -08:00
  • 3e8e6f7e4f Refactor Jenkinsfile (#3229) Illia Silin 2025-11-19 07:20:25 -08:00
  • 2e77585ae1 Merge commit '1eb26460aa621028c5d5a8a20cf593ed8a3a3cc5' into develop assistant-librarian[bot] 2025-11-19 15:13:07 +00:00
  • 5b4ee55185 correct clang-format mohsen saffari 2025-11-19 15:32:39 +01:00
  • 94b3569da0 [ck_tile] Pooling example - Improved tile sizes (#3233) Yashvardhan Agarwal 2025-11-19 16:30:18 +02:00
  • 5f7f81660d [ck_tile] Pooling example - Improved tile sizes (#3233) Yashvardhan Agarwal 2025-11-19 16:30:18 +02:00
  • 1eb26460aa [ck_tile] Pooling example - Improved tile sizes (#3233) Yashvardhan Agarwal 2025-11-19 16:30:18 +02:00
  • c618e4b4e8 [CK_TILE] Vector stores c col layout part4 - fix formatting vec_stores_c_col_v4 Aleksander Dudek 2025-11-19 08:14:20 -06:00
  • d18e076d8a [CK_TILE] Vector stores c col layout part4 Aleksander Dudek 2025-11-19 08:10:58 -06:00
  • 5bfda1f662 Add validity checks for MoE FlatMM scatter and enable bf16 hardware atomic Mohsen Saffari 2025-11-19 14:02:24 +00:00
  • 493857497c update test_moe_comprehensive script flatmm-moe-gemm-improve-testing-coverage Mohsen Saffari 2025-11-19 13:05:48 +00:00
  • c52fbb9259 Re-enable Linqun's Xdl Wmma instances kiefer 2025-10-02 09:47:33 +00:00
  • 7521983fc0 Re-enable old wmma instances kiefer 2025-10-01 10:16:38 +00:00
  • 33c276a3fc work on reference pages for gemm pipeline policies philipm/ck-tile-docs Philip Maybank 2025-11-19 11:13:57 +00:00
  • 055f643f4f Merge commit 'ad57f6ef0bcaeef7988bfd3954aac06554f12afb' into develop assistant-librarian[bot] 2025-11-19 11:12:09 +00:00
  • 08e3e9ea2d Disable FP8 / BF8 testing on CDNA1/2, it doesn't work anymore and needs to be either fixed or removed. kiefer 2025-11-18 12:51:35 +00:00
  • b5e2f26808 [CK_BUILDER] Put global CK functions in an the CK namespace (#3232) John Shumway 2025-11-19 02:23:02 -08:00
  • 1036380fba [CK_BUILDER] Put global CK functions in an the CK namespace (#3232) John Shumway 2025-11-19 02:23:02 -08:00
  • ad57f6ef0b [CK_BUILDER] Put global CK functions in an the CK namespace (#3232) John Shumway 2025-11-19 02:23:02 -08:00
  • dd11be744d block scale code for batch mode ltqin 2025-11-19 09:59:21 +00:00
  • a183b4dc29 add blockscale parameters to kernel ltqin 2025-11-19 06:19:18 +00:00
  • 05f83b643f Merge commit 'd7b31978692a6747f5fc232e2ac424566e40b0b8' into develop assistant-librarian[bot] 2025-11-19 06:16:01 +00:00
  • 44936cfdec [CK_TILE] FMHA Reduce register spilling in fwd with dropout (workaround for CI failures with clang-22) (#3221) Anton Gorenko 2025-11-19 10:40:12 +05:00
  • a9ddd16f4f [CK_TILE] FMHA Reduce register spilling in fwd with dropout (workaround for CI failures with clang-22) (#3221) Anton Gorenko 2025-11-19 10:40:12 +05:00
  • d7b3197869 [CK_TILE] FMHA Reduce register spilling in fwd with dropout (workaround for CI failures with clang-22) (#3221) Anton Gorenko 2025-11-19 10:40:12 +05:00
  • 79f2db722e support a16_wint4 moe yadaish 2025-11-19 04:13:41 +00:00
  • e2bfdba309 Partial Progress : Final Structuring ThruptiRajLakshmanaGowda 2025-11-19 00:26:21 +00:00
  • 2275548400 debugging permuteN khuagarw 2025-11-18 21:59:30 +00:00
  • 751e5d85a6 Merge commit 'e91ee8578cc9e493f12ee01055a35a405571effc' into develop assistant-librarian[bot] 2025-11-18 19:12:13 +00:00
  • f00415fca6 Initial plan copilot/build-gemm-example copilot-swe-agent[bot] 2025-11-18 18:42:07 +00:00
  • b6c966df35 chore(copyright): update copyright header for docs & include directory (#3226) Aviral Goel 2025-11-18 13:23:14 -05:00
  • 3fd2a773f1 chore(copyright): update copyright header for docs & include directory (#3226) Aviral Goel 2025-11-18 13:23:14 -05:00
  • e91ee8578c chore(copyright): update copyright header for docs & include directory (#3226) Aviral Goel 2025-11-18 13:23:14 -05:00
  • 902250eab3 chore(copyright): update copyright header for include directory (#3224) Aviral Goel 2025-11-18 13:17:18 -05:00
  • 0cfa802f89 chore(copyright): update copyright header for include directory (#3224) Aviral Goel 2025-11-18 13:17:18 -05:00
  • f5ac3ee359 chore(copyright): update copyright header for include directory (#3224) Aviral Goel 2025-11-18 13:17:18 -05:00
  • 3774b900d1 [CK-Tile] Remove usage of tile partitioner's full gemm shape (#3204) Max Podkorytov 2025-11-18 09:56:40 -08:00
  • b1faa0c1c5 [CK-Tile] Remove usage of tile partitioner's full gemm shape (#3204) Max Podkorytov 2025-11-18 09:56:40 -08:00
  • a3a4eb12bd [CK-Tile] Remove usage of tile partitioner's full gemm shape (#3204) Max Podkorytov 2025-11-18 09:56:40 -08:00
  • 42d013d007 Update moe_flatmm_kernel to manage OOB Mohsen Saffari 2025-11-18 16:39:09 +00:00
  • 86a4127e31 Merge commit 'ac70206b2c8b43447e46ad382057fe56dc639803' into develop assistant-librarian[bot] 2025-11-18 15:13:30 +00:00
  • a07cd6bc71 feat: add support for bf16 for grouped_gemm & grouped_gemm_preshuffle… (#3225) Aviral Goel 2025-11-18 09:32:27 -05:00
  • adf8515169 feat: add support for bf16 for grouped_gemm & grouped_gemm_preshuffle… (#3225) Aviral Goel 2025-11-18 09:32:27 -05:00
  • ac70206b2c feat: add support for bf16 for grouped_gemm & grouped_gemm_preshuffle… (#3225) Aviral Goel 2025-11-18 09:32:27 -05:00