Jan ŠléglandClaude Opus 5.5 447f3a5080 docs: note the unconditional log2 series and its f16 path ...

v_log_f16 is lowered through the same f32 log2, so the entry names both
and states the f16 effect measured over every f16 input (one NVIDIA
result moves onto the hardware's bits, none on RADV). The series is
chosen with OpSelect, so its cost applies to every v_log.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Fnrk5Dazjb2tLooVSpdRQq

2026-10-11 01:33:44 +02:00

..

ARCHITECTURE.md

Merge pull request #2917 from XairRrR/feat/agc-prepare-unpredicted-variants

2026-10-11 01:31:32 +02:00

BUILD.md

fix(agc): report a refused queue worker priority as a driver failure and reject unknown APS5_THREAD_PRIORITY values

2026-10-10 12:23:12 -06:00

CONVENTIONS.md

refactor(docs): separated user and dev in docs + compatibility file

2026-09-28 22:58:17 +03:00

HW_ORACLE.md

feat(shader): follow FP16_OVFL in every f16 result

2026-10-10 04:32:38 -03:00

PROGRESS.md

fix(progress): link include-shared libraries and label empty ones

2026-10-10 08:49:58 +02:00

TechnicalDebt.md

docs: note the unconditional log2 series and its f16 path

2026-10-11 01:33:44 +02:00