mirror of https://github.com/ROCm/composable_kernel.git synced 2026-07-01 04:07:56 +00:00

Files

Yi DING e397140640 Use filename but not path to filter compilation (#2556 )

[ROCm/composable_kernel commit: 4338346b10]

2025-07-24 17:38:14 +08:00

01_gemm

Add MoE & FP8 Blockscale WP Kernels for GFX950 (#2297 )

2025-06-12 09:25:59 +08:00

02_gemm_bilinear

…

03_gemm_bias_relu

…

04_gemm_add_add_fastgelu

Optimized GEMMs for MX FP4/8 (#2294 )

2025-06-05 13:54:15 -06:00

09_convnd_fwd

Add a daily CI build on gfx908. (#1987 )

2025-03-17 18:08:53 -07:00

10_convnd_fwd_multiple_d_multiple_reduce

Fix example_convnd_fwd_max_xdl_int8 failures on MI300 (#1666 )

2024-11-14 08:40:50 -08:00

12_reduce

…

13_pool2d_fwd

…

14_gemm_quantization

…

15_grouped_gemm

Fix example_grouped_gemm_multiple_d_xdl_fp16 on gfx950 (#2203 )

2025-05-19 14:25:50 -07:00

16_gemm_multi_d_multi_reduces

Fix example_convnd_fwd_max_xdl_int8 failures on MI300 (#1666 )

2024-11-14 08:40:50 -08:00

17_convnd_bwd_data

…

18_batched_gemm_reduce

remove support for gfx940 and gfx941 targets (#1944 )

2025-03-05 11:07:33 -08:00

19_binary_elementwise

…

20_grouped_conv_bwd_weight

Apply universal gemm to bwd_weight_cshuffle operator (#1873 )

2025-02-18 10:10:22 +01:00

21_gemm_layernorm

OCP FP8 support for gfx12. (#1710 )

2024-12-03 08:42:55 -08:00

22_cgemm

…

23_softmax

…

24_batched_gemm

Optimized GEMMs for MX FP4/8 (#2294 )

2025-06-05 13:54:15 -06:00

25_gemm_bias_e_permute

…

26_contraction

…

27_layernorm2d_fwd

…

28_grouped_gemm_bias_e_permute

…

29_batched_gemm_bias_e_permute

…

30_grouped_conv_fwd_multiple_d

Remove CK_USE_AMD_MFMA_GFX950 (#1935 )

2025-03-04 10:32:25 -08:00

31_batched_gemm_gemm

Merge from internal (#1857 )

2025-02-07 15:05:05 -07:00

32_batched_gemm_scale_softmax_gemm

OCP FP8 support for gfx12. (#1710 )

2024-12-03 08:42:55 -08:00

33_multiple_reduce

…

34_batchnorm

Fix build errors on windows (#2456 )

2025-07-16 07:58:23 -07:00

35_splitK_gemm

Optimized GEMMs for MX FP4/8 (#2294 )

2025-06-05 13:54:15 -06:00

36_sparse_embedding

…

37_batched_gemm_add_add_relu_gemm_add

OCP FP8 support for gfx12. (#1710 )

2024-12-03 08:42:55 -08:00

38_grouped_conv_bwd_data_multiple_d

[CK][CONV] Support NCHW in class DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1 (#2459 )

2025-07-17 08:19:57 +08:00

39_permute

…

40_conv2d_fwd_quantization

…

41_grouped_conv_conv_fwd

Merge from internal (#1857 )

2025-02-07 15:05:05 -07:00

42_groupnorm_fwd

…

43_splitk_gemm_bias_e_permute

…

44_elementwise_permute

Prevent instantiation of undefined FP8 operators. (#1639 )

2024-11-05 13:58:29 -08:00

45_elementwise_normalization

…

46_gemm_add_multiply

…

47_gemm_bias_softmax_gemm_permute

OCP FP8 support for gfx12. (#1710 )

2024-12-03 08:42:55 -08:00

48_pool3d_fwd

…

49_maxpool2d_bwd

…

50_put_element

…

51_avgpool3d_bwd

…

52_im2col_col2im

…

53_layernorm2d_bwd

…

54_groupnorm_bwd

…

59_grouped_gemm_multi_ABD

OCP FP8 support for gfx12. (#1710 )

2024-12-03 08:42:55 -08:00

60_gemm_multi_ABD

…

61_contraction_multi_ABD

…

62_convnd_activ

remove support for gfx940 and gfx941 targets (#1944 )

2025-03-05 11:07:33 -08:00

63_layernorm4d_fwd

…

64_fpAintB_gemm

…

65_gemm_multiply_multiply

Fix blockscale fp8 gemm examples (#2476 )

2025-07-10 07:12:13 -07:00

66_complex_contraction_bilinear

chore: unset executable permission (#2303 )

2025-06-10 09:13:59 -07:00

67_gemm_microscaling

MX GEMM - FP6 Support in GEMM MX v3 Pipeline (#2481 )

2025-07-11 13:07:05 -06:00

ck_tile

[CK_TILE] Support for elementwise kernel (#2246 )

2025-07-24 11:21:45 +02:00

CMakeLists.txt

Use filename but not path to filter compilation (#2556 )

2025-07-24 17:38:14 +08:00

README.md

Add basic documentation structure (#1715 )

2024-12-04 00:46:47 +01:00

README.md

Back to the main page

Composable Kernel examples