Skip to content

Actions: JGalego/lqcd-neuron

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
66 workflow runs
66 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
perf(neuron): half-spinor trick for sharded Wilson Dslash CI #66: Commit 2668370 pushed by JGalego
1d 0h 0m 2s main
main
1d 0h 0m 2s
perf(neuron): fewer shards via optlevel-1 downgrade for large lattices CI #65: Commit 2ec3533 pushed by JGalego
23h 38m 59s main
main
23h 38m 59s
perf(neuron): batched shard dispatch — all shards in one NEFF call CI #64: Commit cfb335c pushed by JGalego
39m 16s main
main
39m 16s
perf(neuron): single-NEFF sharded dispatch to eliminate swap overhead CI #63: Commit 51ce036 pushed by JGalego
1h 54m 12s main
main
1h 54m 12s
perf(neuron): parallel sharded dispatch with per-core pinning + multi… CI #62: Commit 033fddb pushed by JGalego
5h 21m 37s main
main
5h 21m 37s
fix(neuron): tighten shard cap to 80k and auto-retry sharded compile CI #61: Commit 44189ba pushed by JGalego
18h 10m 1s main
main
18h 10m 1s
ux(make): precheck AWS credentials before AWS-using targets CI #60: Commit b22bcf0 pushed by JGalego
5h 55m 29s main
main
5h 55m 29s
feat(bench): surface compiler fallbacks in bench-tail notes CI #59: Commit 7c7a11c pushed by JGalego
7m 30s main
main
7m 30s
fix(neuron): fall back to host-loop batched dslash on SRAM/HLO overflow CI #58: Commit 365a7f5 pushed by JGalego
5m 32s main
main
5m 32s
fix(neuron): guard compile_dslash_batched against SRAM/HLO overflow CI #57: Commit ea6a1f1 pushed by JGalego
1d 0h 0m 3s main
main
1d 0h 0m 3s
feat(neuron): surface compile-path provenance in bench output CI #56: Commit 2f73f3d pushed by JGalego
1h 56m 29s main
main
1h 56m 29s
fix(neuron): tighten shard volume cap to 150k sites for HLO budget CI #55: Commit 1b0db42 pushed by JGalego
25m 0s main
main
25m 0s
feat(neuron): T-axis spatial sharding for V > 24^4 lattices CI #54: Commit 85ecb19 pushed by JGalego
18m 21s main
main
18m 21s
feat(bench): add bench-rm, surface contributing run-ids in bench-tail CI #53: Commit 20529c4 pushed by JGalego
37m 31s main
main
37m 31s
feat(bench-runs): add status column with stale/abandoned detection CI #52: Commit 57c39d8 pushed by JGalego
12m 18s main
main
12m 18s
feat(bench): expose --optlevel and OPTLEVEL knob for neuronx-cc CI #51: Commit 8889827 pushed by JGalego
27m 48s main
main
27m 48s
Add cross-run averaging to bench-tail and log neuronx-cc invocation CI #50: Commit cd803d2 pushed by JGalego
1h 6m 7s main
main
1h 6m 7s
fix(bench): make ephemeral wallclock kill switch configurable CI #49: Commit 17ff91c pushed by JGalego
1h 57m 16s main
main
1h 57m 16s
ux(bench-runs): format S3 listing as 'started (UTC) run id' table CI #48: Commit 1342e6a pushed by JGalego
17m 12s main
main
17m 12s
feat(bench-tail): add Speedup column and bold the per-row best throug… CI #47: Commit 0466fd5 pushed by JGalego
3m 43s main
main
3m 43s
fix(bench): stop inlining run_bench_job.sh into EC2 user-data (16 KiB… CI #46: Commit 011476f pushed by JGalego
1h 49m 32s main
main
1h 49m 32s
docs(bench): document partial-result S3 streaming and bench-tail helpers CI #45: Commit ecbba72 pushed by JGalego
16m 31s main
main
16m 31s
feat(bench): stream partial results to S3 during in-flight runs CI #44: Commit b21326d pushed by JGalego
1m 29s main
main
1m 29s
scripts: harden bench job launch against unpushed code and silent stalls CI #43: Commit 3ca34d9 pushed by JGalego
15m 39s main
main
15m 39s
feat(bench): batch-size sweep + strict fused fallback CI #42: Commit 3df3a75 pushed by JGalego
13m 5s main
main
13m 5s