igneum-network/igneum
1
0
Fork 0
Code Issues Pull requests Releases Activity
igneum/proto-cuda/packs-readwidth
History
igneum-labs 357ccb9507 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-05 20:18:35 +00:00
..
mixA-0 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixA-1 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixA-2 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixA-3 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixA-4 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixA-5 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-0 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-1 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-2 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-3 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-4 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
mixB-5 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
scr0k32 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr2k32 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr2k128 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr4k32 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr4k128 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr8k32 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
scr8k128 read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
w4 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
w16 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
w64 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00
w64x4 read-width experiment (gate 1): load classes W=4/16/64, per-load width mix, scratch RMW variant behind a generator flag; 20 packs; CPU verifier and acceptance mirror; emulator shims 2026-10-05 19:46:54 +00:00