[ROCm] Remove stale SDPA and skinny GEMM workarounds (#50907)
Signed-off-by: Andreas Karatzas <Andreas.Karatzas@amd.com> Signed-off-by: Andreas Karatzas <akaratza@amd.com> Co-authored-by: OpenAI Codex <codex@openai.com>
This commit is contained in:
co-authored by
OpenAI Codex
parent
4f2f31b82e
commit
1ab2801dde
@@ -134,12 +134,16 @@ def rocm_unquantized_gemm_impl(
|
||||
GrpsShrB = min(N_p2 // 16, 4)
|
||||
# Given the above, how many CUs would we need?
|
||||
CuNeeded = rndup_cus * GrpsShrB
|
||||
# candidate for atomic reduce count splitk?
|
||||
# Deterministic reduction stores one float workspace value per K shard.
|
||||
fits_wvsplitkrc = (
|
||||
N_p2 * m * ((k + 512 - 1) // 512)
|
||||
) <= 128 * 1024 * 12 # deterministic
|
||||
fits_wvsplitkrc &= CuNeeded <= cu_count
|
||||
|
||||
skinny_operands_compatible = weight.is_contiguous() and (
|
||||
bias is None or bias.is_contiguous()
|
||||
)
|
||||
|
||||
use_skinny_reduce_counting = (
|
||||
envs.VLLM_ROCM_USE_SKINNY_GEMM
|
||||
and on_gfx950()
|
||||
@@ -151,12 +155,13 @@ def rocm_unquantized_gemm_impl(
|
||||
and k > 512
|
||||
and m % 16 == 0
|
||||
and fits_wvsplitkrc
|
||||
and weight.is_contiguous()
|
||||
and skinny_operands_compatible
|
||||
)
|
||||
)
|
||||
|
||||
if use_skinny_reduce_counting:
|
||||
return ops.wvSplitKrc(x, weight, cu_count, bias)
|
||||
x_view = x.reshape(-1, x.size(-1)).contiguous()
|
||||
return ops.wvSplitKrc(x_view, weight, cu_count, bias)
|
||||
|
||||
# gfx1250's aiter gemm_a16w16 uses the gluon backend, which requires
|
||||
# K % 256 == 0 (it walks K with fixed-size descriptors and won't pad a
|
||||
@@ -174,10 +179,13 @@ def rocm_unquantized_gemm_impl(
|
||||
# TODO GFX1250: Include once skinny GEMM is supported on gfx1250
|
||||
and x.dtype in [torch.float16, torch.bfloat16]
|
||||
and k % 8 == 0
|
||||
and skinny_operands_compatible
|
||||
)
|
||||
|
||||
if use_skinny:
|
||||
x_view = x.reshape(-1, x.size(-1))
|
||||
# The skinny kernels assume contiguous K elements. A shape-preserving
|
||||
# reshape can retain a transposed activation's non-contiguous strides.
|
||||
x_view = x.reshape(-1, x.size(-1)).contiguous()
|
||||
if m > 8 and 0 < n <= 5:
|
||||
cu_count = num_compute_units()
|
||||
out = ops.wvSplitK(weight, x_view, cu_count, bias)
|
||||
|
||||
Reference in New Issue
Block a user