20231088/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

forward fix PR 14245, restore build on ROCm 6.2 (#14709)

Browse Source 
Signed-off-by: Jeff Daily <jeff.daily@amd.com>
This commit is contained in:
Jeff Daily 2025-03-13 20:40:15 -07:00 committed by GitHub
parent 7888e1d0a3
commit 2a602b055a
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 12 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
csrc/quantization/fp8/amd/quant_utils.cuh
View File

@ -19,12 +19,24 @@ __device__ __forceinline__ fp8_type cvt_c10(float const r) {
return {};
}
// __hip_fp8_e4m3 only exists starting in ROCm 6.3. The macro
// HIP_FP8_TYPE_OCP comes from the hip_fp8.h header and also makes
// its first appearance in ROCm 6.3. Since VLLM_DISPATCH_FP8_TYPES
// on ROCm instantiates both OCP and FNUZ kernels, we need to replace
// the new HW cvt with something reasonable that doesn't rely on the
// ROCm 6.3 feature. This allows compiling on ROCm 6.2 or newer.
template <>
__device__ __forceinline__ c10::Float8_e4m3fn cvt_c10(float const r) {
#if HIP_FP8_TYPE_OCP
return c10::Float8_e4m3fn(
__hip_cvt_float_to_fp8(r, __hip_fp8_e4m3::__default_saturation,
__hip_fp8_e4m3::__default_interpret),
c10::Float8_e4m3fn::from_bits());
#else
// Cast implemented by pytorch. Uses bit manipulation instead of HW cvt.
// HW cvt above is faster when it is available (ROCm 6.3 or newer).
return static_cast<c10::Float8_e4m3fn>(r);
#endif
}
template <>