Jan ŠléglandClaude Opus 5.5 60301adc6d test(agc-driver): add a split-lane case with per-lane operands to the LDS mixed-width atomic test ...

On RADV the race between ds_add_u64 under the workgroup lock and a
concurrent ds_add_u32 only shows when the lanes get their operands from
per-lane buffer loads: with v_mov_b32 constants or loads from one address
no increment is lost there. The new case has the split-lane kernel's
parity split and loads each lane's (0xffffffff, 0, 1, 0) row from the input
buffer at tid * 16 with buffer_load_dwordx4, so the input buffer now holds
256 rows. With SpirvMemoryInstructions.cpp reverted it fails on RADV and
on the RTX 3080; with the change it passes on both.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Fnrk5Dazjb2tLooVSpdRQq

2026-10-11 01:53:37 +03:00

..

Decoder

…

libs

test(agc-driver): add a split-lane case with per-lane operands to the LDS mixed-width atomic test

2026-10-11 01:53:37 +03:00

relinker

fix(relinker): treat address-taken code as reachable in unused-filter=1

2026-10-11 00:35:27 +02:00

shader/recompiler

fix(shader): LDS 32-bit atomics take the workgroup lock of the program's 64-bit LDS atomics

2026-10-11 01:53:36 +03:00