Skip to content

Actions: liangliangchang/llama.cpp

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
103 workflow runs
103 workflow runs

Filter by Workflow

⌥ + click/return to exclude

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
hip: disable direct host access on gfx1151 .github/workflows/build-cann.yml #45: Commit 865374b pushed by liangliangchang
Failure lichang.gfx11-master-sync
lichang.gfx11-master-sync
Failure
hip: restore RDNA3.5 Q4_K tile load schedule Python Type-Check #4: Commit b53bea0 pushed by liangliangchang
1d 0h 0m 2s lichang.gfx11-master-sync
lichang.gfx11-master-sync
1d 0h 0m 2s
hip: restore RDNA3.5 Q4_K tile load schedule .github/workflows/build-cann.yml #44: Commit b53bea0 pushed by liangliangchang
Failure lichang.gfx11-master-sync
lichang.gfx11-master-sync
Failure
hip: avoid emulated byte permutes for Q1_0 .github/workflows/build-cann.yml #43: Commit 43439ac pushed by liangliangchang
Failure lichang.gfx11-master-sync
lichang.gfx11-master-sync
Failure
test: include Q1_0 in gfx1151 MMQ matrix .github/workflows/build-cann.yml #42: Commit 07ba1da pushed by liangliangchang
Failure lichang.gfx11-master-sync
lichang.gfx11-master-sync
Failure
test: cover gfx1151 production shapes across MMQ types Update Operations Documentation #3: Commit 5c0a3c1 pushed by liangliangchang
1d 0h 0m 1s lichang.gfx11-master-sync
lichang.gfx11-master-sync
1d 0h 0m 1s
test: cover gfx1151 production shapes across MMQ types Python Type-Check #3: Commit 5c0a3c1 pushed by liangliangchang
1d 0h 0m 2s lichang.gfx11-master-sync
lichang.gfx11-master-sync
1d 0h 0m 2s
test: cover gfx1151 production shapes across MMQ types Python check requirements.txt #2: Commit 5c0a3c1 pushed by liangliangchang
1d 0h 0m 2s lichang.gfx11-master-sync
lichang.gfx11-master-sync
1d 0h 0m 2s
test: cover gfx1151 production shapes across MMQ types Check Pre-Tokenizer Hashes #3: Commit 5c0a3c1 pushed by liangliangchang
1d 0h 0m 2s lichang.gfx11-master-sync
lichang.gfx11-master-sync
1d 0h 0m 2s
test: cover gfx1151 production shapes across MMQ types .github/workflows/build-cann.yml #41: Commit 5c0a3c1 pushed by liangliangchang
Failure lichang.gfx11-master-sync
lichang.gfx11-master-sync
Failure
perf(hip): pipeline RDNA3.5 Q4_K WMMA .github/workflows/build-cann.yml #40: Commit 2470481 pushed by liangliangchang
Failure lichang.q4k-opt
lichang.q4k-opt
Failure
fix(hip): tune RDNA3.5 MMQ tile policy for Q4_K and small-N shapes .github/workflows/build-cann.yml #39: Commit a610b9c pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
fix(hip): restore RDNA3.5 activation B-pipeline without MTP regression. .github/workflows/build-cann.yml #38: Commit cf31ddd pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
fix(hip): drop RDNA3.5 activation B-pipeline that regressed MTP accept. .github/workflows/build-cann.yml #37: Commit 68bbe93 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
refactor(hip): drop q4k_legacy flags from MMQ host tile policy .github/workflows/build-cann.yml #36: Commit 81ab3f4 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
test: dedupe RDNA3.5 MMQ prefill cases in test-backend-ops .github/workflows/build-cann.yml #35: Commit a2d70ae pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
fix(hip): RDNA3.5 MMQ vec_dot accuracy and Q4_K MTP-safe path .github/workflows/build-cann.yml #34: Commit effc3ca pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
fix(hip): restore Q4_0 MMQ vec_dot at mmq_x<=32 on RDNA3.5 .github/workflows/build-cann.yml #33: Commit c1594c8 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
test: add RDNA3.5 MMQ prefill kernel cases for PR #32. .github/workflows/build-cann.yml #32: Commit c12b327 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
Merge branch 'gfx11' of rocm/llama.cpp into lichang.mmq_opt .github/workflows/build-cann.yml #31: Commit cb01a99 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
Merge latest rocm/gfx11 into lichang.mmq_opt Python Type-Check #2: Commit 01637a8 pushed by liangliangchang
1d 0h 0m 1s lichang.mmq_opt
lichang.mmq_opt
1d 0h 0m 1s
Merge latest rocm/gfx11 into lichang.mmq_opt Check Pre-Tokenizer Hashes #2: Commit 01637a8 pushed by liangliangchang
1d 0h 0m 2s lichang.mmq_opt
lichang.mmq_opt
1d 0h 0m 2s
Merge latest rocm/gfx11 into lichang.mmq_opt Update Operations Documentation #2: Commit 01637a8 pushed by liangliangchang
1d 0h 0m 1s lichang.mmq_opt
lichang.mmq_opt
1d 0h 0m 1s
Merge latest rocm/gfx11 into lichang.mmq_opt .github/workflows/build-cann.yml #30: Commit 01637a8 pushed by liangliangchang
Failure lichang.mmq_opt
lichang.mmq_opt
Failure
Fix Q6_K B-tile prefetch using wrong K offset in vec_dot Check Pre-Tokenizer Hashes #1: Commit 74503ef pushed by jimw567
1d 0h 0m 1s lichang.mmq_opt
lichang.mmq_opt
1d 0h 0m 1s