Pull requests / #1540

#1540 dq_q2_0: keep the Q2_0 signed-zero fix on every HIP arch (#1474)

open · @nekomario28 · 0 commentaires · Sur GitHub

AMD / HIPNVIDIA / CUDAModels & quants

Description

## Summary

- `hip_q2_zero` fails on gfx1201 (R9700) with ROCm 7.10: the half path folds a negative scale times +0 to +0, but the `dq_q2_0` workaround was gated to `(__gfx1012__ && HIP_VERSION_MAJOR < 7) || __gfx1151__` — the arch list only covers the reports seen so far, and gfx1201 on HIP 7 falls outside it.
- Drop the arch/version conditions, keep `__HIPCC__`. The workaround writes the scale's sign bit for code 1, which is value-identical on targets whose compiler never folds — this matches the reporter's proposed fix.

Fixes #1474.

## Test plan

- [x] `amdclang++ -x hip -fsyntax-only --offload-arch=gfx1201` on `src/kernels/cuda/iq_kernels.cu` with the build's compat flags: clean
- [ ] `ctest -R hip_q2_zero` on the reporter's R9700 / ROCm 7.10 build: expect PASS (128 signed-zero failures → 0)
- [ ] CUDA builds unaffected (`#if defined(__HIPCC__)` only)

Sur le site

Liens install, modèles, releases.