Search results

[Perf Regression] 20 config(s) regressed @ 80b5b02b

performance

regression

Status: Open.

#1463 In ROCm/ATOM;

· github-actions opened on Jul 4, 2026

[Issue]: Concurrent TP8's from QPX partitions fail

Status: Open.

#1462 In ROCm/ATOM;

· mustafa-khalid opened on Jul 3, 2026

[Perf Regression] 30 config(s) regressed @ f9017b58

performance

regression

Status: Open.

#1446 In ROCm/ATOM;

· github-actions opened on Jul 3, 2026

[Perf Regression] 5 config(s) regressed @ 92d3d6a0

performance

regression

Status: Open.

#1442 In ROCm/ATOM;

· github-actions opened on Jul 2, 2026

[Perf Regression] 9 config(s) regressed @ 4b7b05dd

performance

regression

Status: Open.

#1440 In ROCm/ATOM;

· github-actions opened on Jul 2, 2026

[Perf Regression] 9 config(s) regressed @ 4b7b05dd

performance

regression

Status: Open.

#1439 In ROCm/ATOM;

· github-actions opened on Jul 2, 2026

[Perf Regression] 41 config(s) regressed @ 04b120d3

performance

regression

Status: Open.

#1438 In ROCm/ATOM;

· github-actions opened on Jul 2, 2026

[Bug] GLM-5.2-FP8 (glm_moe_dsa) emits garbage output past ~1.5k input tokens (tp8, gfx950)

Status: Open.

#1435 In ROCm/ATOM;

· yhl-amd opened on Jul 2, 2026

[Bug] DeepSeek-V4: prefix caching corrupts output on cross-request hit (SWA window read from stale per-request ring)

Status: Open.

#1417 In ROCm/ATOM;

· yhl-amd opened on Jun 30, 2026

MiniMax-M3-MXFP4: image requests rejected — processor has no chat template (multimodal eval blocked)

Status: Open.

#1407 In ROCm/ATOM;

· Princejain1101 opened on Jun 30, 2026

[Issue]: DeepSeek-V4-Flash Memory access fault under 64/128 concurrency with 3000+ requests

Status: Open.

#1400 In ROCm/ATOM;

· ab1stone opened on Jun 29, 2026

[Perf Regression] 1 config(s) regressed @ f696e2bc

performance

regression

Status: Open.

#1398 In ROCm/ATOM;

· github-actions opened on Jun 29, 2026