jgrusewski fd31742627 fix(cuda): Plan A v2 — dd049d9a4 + 3 kernel-only bug fixes (no Phase 2.0) ...

Plan A v1 (commit 2d68ef5d8 = revert Phase 2.1 on top of 104fe81ca)
failed at cluster (alpha-rl-4sjzw): entropy collapsed to 0.69 by step 3000.
The Phase 2.0 V envelope clamp (c52282fb4) — kept in v1 — was likely
the culprit: it bounds V_pred to a tight envelope, biasing advantages
and breaking PPO gradient flow.

Plan A v2: start from dd049d9a4 (proven working at cluster wr=0.57
+$6.3M @ step 15000 today via alpha-rl-lpbp8) and apply ONLY the
kernel-only bug fixes that don't affect training dynamics:

- variable_selection.cu: VSN stride 40→56 fix (104fe81ca) — prevents
  step-4 NaN from reading wrong-stride window_tensor
- bucket_transition_kernels.cu: h_mag_per_bucket multi-warp 32→128
  (7e38e46e6) — correct warp-shuffle reduction for HIDDEN_DIM=128
- compute_advantage_return.cu: branch-gate done flag (a6acc25ec) —
  done's terminal-state semantics applied via gate, resolves step-4 NaN

NOT applied (intentionally):
- c52282fb4 Phase 2.0 V envelope clamp — biases V regression target
- b4aadff75 V envelope ±200→±10 — extends Phase 2.0
- 10d4614fb atomicAdd removal — kernel non-determinism is real but
  intrusive (Rust changes), dd049d9a4 works without this fix
- db4d9a16f Phase 2.1 dueling — broken decomposition per spec analysis
- 72672c9e7+ Phase 2.2/2.3/H/P1+P2+P3 — built on broken Phase 2.1

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-05-29 21:34:29 +02:00

9.7 KiB

Raw Blame History

View Raw