ept.git/ep_run/agg_bench.log, branch master Unnamed repository; edit this file 'description' to name the repository. aggregate speed bench: speed tier hf+sd 1.47x; accuracy tier hf+sd+t80+avg (0.91x, cos 0.94, avg free); compile demoted 2026-07-06T14:17:36+00:00 Yuren Hao yurenh2@illinois.edu 2026-07-06T14:17:36+00:00 9a8b2796ca12e4d4c24717485a635a301aa6d07f Full-ep_step wall times on quiet A6000 (warm s2000, B24), res parity across all 8 configs. compile only 1.12x at this shape (historical 1.46x was a different workload split); FULL cmp_sdpa saves 4% over eager at t80 — not worth the guard complexity. tforce_sdpa added (flash baked into compiled graph, flag-free so grad paths never see SDPA). bp_lm --tie probe in flight. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn 
Full-ep_step wall times on quiet A6000 (warm s2000, B24), res parity across
all 8 configs. compile only 1.12x at this shape (historical 1.46x was a
different workload split); FULL cmp_sdpa saves 4% over eager at t80 — not
worth the guard complexity. tforce_sdpa added (flash baked into compiled
graph, flag-free so grad paths never see SDPA). bp_lm --tie probe in flight.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn