DotDebianandClaude Opus 5.5 f617b873cc test(shader): mask the ds_wrxchg2 lanes with a per-lane compare ...

The masked case loaded EXEC from a scalar literal. A scalar mask is read
by subgroup lane, so on a host whose subgroups are narrower than the wave
(llvmpipe: 8 lanes) lane 16 took bit 0 and stayed active. v_cmpx_gt_u32
on the thread id gives the same low 16 lanes on any subgroup size.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

2026-10-06 22:46:27 +02:00

..

Decoder

fix(jpegenc): honor chroma sampling and encode true grayscale JPEGs

2026-10-05 21:49:36 +02:00

libs

test(shader): mask the ds_wrxchg2 lanes with a per-lane compare

2026-10-06 22:46:27 +02:00

relinker

Merge pull request #767 from 011010/feat/relinker-windows-tls-alu-reads

2026-10-06 16:35:21 +13:00

shader/recompiler

feat(shader): implement ds_wrxchg2_rtn_b32 and ds_wrxchg2st64_rtn_b32

2026-10-06 22:46:27 +02:00