Skip to content

Actions: ROCm/aiter

Actions

Triton Test

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
triton-test.yaml
2,500+ workflow runs
2,500+ workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
[Triton/Gluon] [gfx950] pa_decode_sparse: pick BLOCK_K by occupancy, not a constant (up to +43 %) Triton Test #10575: Pull request #5294 opened by stefanskiasan
In progress stefanskiasan:pa-decode-sparse-block-k-occupancy
stefanskiasan:pa-decode-sparse-block-k-occupancy
In progress
[Triton/Gluon] Add MXFP8 convert and fast-transpose kernels Triton Test #10574: Pull request #5203 synchronize by WuLei-AMD
Queued WuLei-AMD:dev_lumen_pr3.2_transpose_mxfp8
WuLei-AMD:dev_lumen_pr3.2_transpose_mxfp8
Queued
[Triton/Gluon] Add FP8 block-wise quantization kernels Triton Test #10573: Pull request #5177 synchronize by WuLei-AMD
Action required WuLei-AMD:dev_lumen_pr3_quant
WuLei-AMD:dev_lumen_pr3_quant
Action required
[Triton/Gluon] Add vocab-parallel cross-entropy kernel Triton Test #10572: Pull request #5165 synchronize by WuLei-AMD
Action required WuLei-AMD:dev_lumen_pr2_cross_entropy
WuLei-AMD:dev_lumen_pr2_cross_entropy
Action required
[Config] configs: GLM-5.2 a8w8-bpreshuffle rows for the shapes a live… Triton Test #10571: Commit 636098e pushed by valarLip
47m 32s main
main
47m 32s
[HIP] [CI] [JIT] [ROCm][MoE] Add fused MoE routing preamble for MXFP4 decode Triton Test #10570: Pull request #5150 synchronize by sshlyapn
47m 47s sshlyapn:sshliapn/features/fused_router
sshlyapn:sshliapn/features/fused_router
47m 47s
[HIP] [CI] [JIT] [ROCm][MoE] Add fused MoE routing preamble for MXFP4 decode Triton Test #10569: Pull request #5150 synchronize by sshlyapn
1m 25s sshlyapn:sshliapn/features/fused_router
sshlyapn:sshliapn/features/fused_router
1m 25s
perf(mqa-logits): build the FP4 prefill schedule in one kernel, not 2… Triton Test #10568: Commit 456b927 pushed by valarLip
1h 27m 46s main
main
1h 27m 46s
Guard against negative expert ids in MoE sorting P0 (#4839) Triton Test #10567: Commit b297d2d pushed by yifehuan
47m 15s main
main
47m 15s
[FlyDSL] [CI] FlyDSL FMHA forward-prefill A16W16 kernel for gfx1250 (… Triton Test #10566: Commit 7aaebf6 pushed by coderfeli
42m 51s main
main
42m 51s
[FlyDSL] [CI] FlyDSL FMHA forward-prefill A16W16 kernel for gfx1250 (fmha_fwd_prefill_m32x8) Triton Test #10565: Pull request #5168 synchronize by coderfeli
47m 58s jruan/gfx1250_fdsl_mha_d128_bf16
jruan/gfx1250_fdsl_mha_d128_bf16
47m 58s
[Triton/Gluon] Add 3D neighborhood flash attention (na3d_flash) Triton Test #10564: Pull request #5230 synchronize by jjuvonen-amd
44m 7s jjuvonen-amd:feat/3d-neighborhood-flash-attn
jjuvonen-amd:feat/3d-neighborhood-flash-attn
44m 7s
[FlyDSL] [CI] FlyDSL FMHA forward-prefill A16W16 kernel for gfx1250 (fmha_fwd_prefill_m32x8) Triton Test #10563: Pull request #5168 synchronize by coderfeli
2m 2s jruan/gfx1250_fdsl_mha_d128_bf16
jruan/gfx1250_fdsl_mha_d128_bf16
2m 2s
[Triton/Gluon] Move fp4 GEMM gluon to _gluon_kernels Triton Test #10562: Pull request #4997 synchronize by Boss2002n
49m 12s vinayak/gemm_afp4wfp4_to_gluon_kernels
vinayak/gemm_afp4wfp4_to_gluon_kernels
49m 12s
[Triton/Gluon] [GFX950] Add Unified Attention Gluon Kernel Triton Test #10561: Pull request #4614 synchronize by cagrikymk
2s cagri/unif_2d_gfx950_clean
cagri/unif_2d_gfx950_clean
2s
[Triton/Gluon] bf16 persistent gemm Triton Test #10560: Pull request #4860 synchronize by amirumoAMD
2m 23s amemoore/bf16-persistent-gemm
amemoore/bf16-persistent-gemm
2m 23s
[Triton/Gluon] bf16 persistent gemm Triton Test #10559: Pull request #4860 synchronize by amirumoAMD
13m 34s amemoore/bf16-persistent-gemm
amemoore/bf16-persistent-gemm
13m 34s
[Triton/Gluon] [GFX950] Add Unified Attention Gluon Kernel Triton Test #10558: Pull request #4614 synchronize by cagrikymk
1s cagri/unif_2d_gfx950_clean
cagri/unif_2d_gfx950_clean
1s
[Triton/Gluon] bf16 persistent gemm Triton Test #10557: Pull request #4860 synchronize by amirumoAMD
31m 36s amemoore/bf16-persistent-gemm
amemoore/bf16-persistent-gemm
31m 36s
[Triton/Gluon] [GFX950] Add Unified Attention Gluon Kernel Triton Test #10556: Pull request #4614 synchronize by cagrikymk
2s cagri/unif_2d_gfx950_clean
cagri/unif_2d_gfx950_clean
2s
[Triton/Gluon] [GFX950] Add Unified Attention Gluon Kernel Triton Test #10555: Pull request #4614 synchronize by cagrikymk
2s cagri/unif_2d_gfx950_clean
cagri/unif_2d_gfx950_clean
2s
[Triton/Gluon] [Bugfix] Guard RDNA unified attention against LDS overflow Triton Test #10554: Pull request #4868 synchronize by amd-xavierwang
Action required amd-xavierwang:rdna-unified-attn-lds-guard
amd-xavierwang:rdna-unified-attn-lds-guard
Action required
[Triton/Gluon] [GFX1250] fused_add_rmsnorm_pad() gluon equivalent function Triton Test #10553: Pull request #4146 synchronize by amd-jrosas
13m 38s jrosas/fused_add_rmsnorm_pad
jrosas/fused_add_rmsnorm_pad
13m 38s
[Triton/Gluon] [Bugfix] Guard RDNA unified attention against LDS overflow Triton Test #10552: Pull request #4868 synchronize by amd-xavierwang
Action required amd-xavierwang:rdna-unified-attn-lds-guard
amd-xavierwang:rdna-unified-attn-lds-guard
Action required
[Triton/Gluon] split K correct OOB afp4wfp4 Triton Test #10551: Pull request #5261 opened by afriedri
53m 22s afriedri:fp4_preshuffle_fix
afriedri:fp4_preshuffle_fix
53m 22s