Issues

• --reset-state does not reset bringup_rounds, so a reset session can enter ENABLEMENT already at its attempt cap

  AMD-AGI/Hyperloom#1766 · Filter by author MohitAMD opened on Oct 6, 2026

• ENABLEMENT: _close_enablement_lane missing from Coordinator._DELEGATED raises AttributeError at the attempt cap

  AMD-AGI/Hyperloom#1765 · Filter by author MohitAMD opened on Oct 6, 2026

• Roofline in Hyperloom

  domain:inferencetype:feature
  AMD-AGI/Hyperloom#1646 · Filter by author tsrikris opened on Sep 23, 2026
  3

  

• Magpie's --disable-radix-cache default on the AgentX path

  domain:inferencetype:bug
  AMD-AGI/Hyperloom#1601 · Filter by author amd-ethany opened on Sep 22, 2026
  1

  

• Proposal: sparse (IRT-calibrated) accuracy evaluations for the accuracy gate

  AMD-AGI/Hyperloom#1539 · Filter by author KHassanQazi opened on Sep 16, 2026
  4

  

• TraceLens Bypass Path Cleanup

  awaiting-author-responsedomain:inferencetype:feature
  AMD-AGI/Hyperloom#1517 · Filter by author tsrikris opened on Sep 15, 2026
  8

  

• KB warm replay does not compare the ROCm/AITER build a config was tuned on

  domain:inferencetype:feature
  AMD-AGI/Hyperloom#1507 · Filter by author rpoornac opened on Sep 14, 2026
  1

  

• --force-resume with a raised --target-gain closes immediately: target_reached_at survives the resume

  domain:inferencetype:bug
  AMD-AGI/Hyperloom#1469 · Filter by author rodosingh opened on Sep 10, 2026
  

• Track experimental Radeon 8065S (Halo/gfx1151) custom-workload bring-up

  AMD-AGI/Hyperloom#1467 · Filter by author powderluv opened on Sep 10, 2026
  

• Verify GEAK CI recovery: paused dispatch and no selected new contribution

  AMD-AGI/Hyperloom#1440 · Filter by author iraj465 opened on Sep 8, 2026
  

• GEAK records acceptance after native current-best promotion is vetoed

  AMD-AGI/Hyperloom#1436 · Filter by author iraj465 opened on Sep 8, 2026
  

• Retained GEAK overlays are omitted from concurrency sweep and exported launch script

  AMD-AGI/Hyperloom#1435 · Filter by author iraj465 opened on Sep 8, 2026
  

• GEAK fallback can promote candidates rejected by native accuracy validation

  AMD-AGI/Hyperloom#1433 · Filter by author iraj465 opened on Sep 8, 2026
  

• MiniMax MXFP4 SGLang drops SwiGLU-OAI semantics, blocking valid GEAK validation

  AMD-AGI/Hyperloom#1432 · Filter by author iraj465 opened on Sep 8, 2026
  

• Preserve environment values across the GEAK handoff and return path

  AMD-AGI/Hyperloom#1429 · Filter by author iraj465 opened on Sep 7, 2026
  

• RDNA support (gfx11/gfx12): interest in a contributed runner + profiling backend?

  AMD-AGI/Hyperloom#1196 · Filter by author The-Monk opened on Aug 17, 2026
  8

• Add RDNA4 GPU support (RX 9070 XT, RX 9070, RX 9060 XT, RX 9000 XT)

  domain:inferencedomain:trainingtype:feature
  AMD-AGI/Hyperloom#1033 · Filter by author Maxritz opened on Jul 26, 2026
  2