[CPU Backend] CPU top-k and top-p sampling kernels using Triton (#43633)

Signed-off-by: Li, Tianmu <tianmu.li@intel.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
This commit is contained in:
Tianmu Li
2026-05-29 15:02:39 +08:00
committed by GitHub
co-authored by mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
parent 04516eabc8
commit 94d3f4d205
5 changed files with 23 additions and 6 deletions
+2 -1
View File
@@ -5,6 +5,7 @@ import torch
from torch import Generator
from vllm.platforms import current_platform
from vllm.triton_utils import HAS_TRITON
from vllm.v1.sample.ops.topk_topp_sampler import apply_top_k_top_p_pytorch
DEVICE_TYPE = current_platform.device_type
@@ -151,7 +152,7 @@ def test_flashinfer_sampler():
# =============================================================================
@pytest.mark.skipif("cpu" in DEVICE_TYPE, reason="CUDA/XPU not available")
@pytest.mark.skipif(not HAS_TRITON, reason="Triton not available on this platform")
class TestTritonTopkTopp:
"""Tests for the Triton top-k/top-p kernel."""