Skip to content

Actions: ROCm/fleet-chiplet-megakernel

Actions

Clang format

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
code-format.yml
54 workflow runs
54 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
gfx950: leave one LDS block free so a scheduler fits beside a worker Clang format #54: Commit ddeb577 pushed by sangeeta0201
59s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
59s
demo.py: make the measured MI355X decode recipe the default Clang format #53: Commit 0befef7 pushed by sangeeta0201
53s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
53s
Allow disabling top-k renormalization in moe_topk_softmax_mi300 Clang format #52: Pull request #13 synchronize by titoatwork
Action required titoatwork:fix-moe-topk-no-renorm
titoatwork:fix-moe-topk-no-renorm
Action required
Barrier between mark reads and compaction writes in moe_topk_softmax_mi300 Clang format #51: Pull request #12 synchronize by titoatwork
Action required titoatwork:fix-moe-topk-compaction-barrier
titoatwork:fix-moe-topk-compaction-barrier
Action required
Allow disabling top-k renormalization in moe_topk_softmax_mi300 Clang format #50: Pull request #13 opened by titoatwork
Action required titoatwork:fix-moe-topk-no-renorm
titoatwork:fix-moe-topk-no-renorm
Action required
Barrier between mark reads and compaction writes in moe_topk_softmax_mi300 Clang format #49: Pull request #12 opened by titoatwork
Action required titoatwork:fix-moe-topk-compaction-barrier
titoatwork:fix-moe-topk-compaction-barrier
Action required
Router-partial: allocate the scratch for single-layer graphs; publish… Clang format #48: Commit e624b0c pushed by sangeeta0201
42s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
42s
chore(deps): bump the github-actions group across 1 directory with 8 updates Clang format #47: Pull request #11 opened by dependabot Bot
35s dependabot/github_actions/github-actions-8fc4621966
dependabot/github_actions/github-actions-8fc4621966
35s
chore(deps): bump the github-actions group across 1 directory with 8 … Clang format #46: Commit a11fcde pushed by dependabot Bot
41s dependabot/github_actions/github-actions-8fc4621966
dependabot/github_actions/github-actions-8fc4621966
41s
chore: add security scanning workflows (#10) Clang format #45: Commit 0275d46 pushed by sangeeta0201
1m 18s amd_mi350
amd_mi350
1m 18s
chore: add security scanning workflows Clang format #44: Pull request #10 opened by haribabug
55s add-security-scanning
add-security-scanning
55s
chore: add security scanning workflows Clang format #43: Commit f5d90aa pushed by haribabug
1m 46s add-security-scanning
add-security-scanning
1m 46s
Fix the stale-SGPR fault in the QKV L2 warm path Clang format #42: Commit 1288b43 pushed by sangeeta0201
56s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
56s
Fix the stale-SGPR fault in the QKV L2 warm path Clang format #41: Commit 76d8311 pushed by sangeeta0201
46s fleet-det-stack
fleet-det-stack
46s
MPK_LTK_EARLY_TAG: router tag at logit time, per-XCD normed-row ready… Clang format #40: Commit cf52673 pushed by sangeeta0201
54s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
54s
Make -O3 the default hipcc opt level for the MI300/MI350 path Clang format #39: Commit e24d15d pushed by sangeeta0201
1m 25s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
1m 25s
TopK: skip the routing-index clear and reduce softmax rows with DPP Clang format #38: Commit 8631fa7 pushed by sangeeta0201
47s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
47s
Destroy previous worker/scheduler streams on re-init Clang format #37: Pull request #9 opened by badnikhil
Action required badnikhil:fix-stream-leak-on-reinit
badnikhil:fix-stream-leak-on-reinit
Action required
Let persistent chunks span KV pages Clang format #36: Commit 0fcfdfc pushed by sangeeta0201
1m 7s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
1m 7s
Gate chunked decode on the repo content criterion Clang format #35: Commit b87f692 pushed by sangeeta0201
1m 7s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
1m 7s
Amortize vLLM decode with persistent token chunks Clang format #34: Commit a939dd4 pushed by sangeeta0201
2m 43s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
2m 43s
Clarify native Fleet performance path Clang format #33: Commit d437912 pushed by sangeeta0201
1m 23s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
1m 23s
Optimize Fleet MK vLLM greedy decode Clang format #32: Commit 3aae77d pushed by sangeeta0201
1m 24s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
1m 24s
Fix Fleet MK vLLM decode hang and document the 0.27.1 path Clang format #31: Commit 85d6ef5 pushed by sangeeta0201
1m 11s fleet-mk-vllm-0.27
fleet-mk-vllm-0.27
1m 11s
Decode 1.70 -> 1.64 ms: pin each routed expert to one XCD pair Clang format #30: Commit 9d478ea pushed by sangeeta0201
52s amd_mi355_gpt_oss120b
amd_mi355_gpt_oss120b
52s