[MoE Refactor] Mxfp4 oracle rebased (#37128)

Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Yongye Zhu
2026-03-21 03:37:04 +00:00
committed by GitHub
co-authored by Claude Opus 4.6
parent c7f98b4d0a
commit 87bd91892f
18 changed files with 1707 additions and 1381 deletions
@@ -6,6 +6,7 @@ import pytest
import torch
import torch.nn.functional as F
from vllm.platforms import current_platform
from vllm.utils.import_utils import has_triton_kernels
if not has_triton_kernels():
@@ -14,6 +15,7 @@ if not has_triton_kernels():
allow_module_level=True,
)
import triton_kernels.matmul_ogs_details.opt_flags as opt_flags
import triton_kernels.swiglu
from triton_kernels.matmul_ogs import FlexCtx, PrecisionConfig
from triton_kernels.numerics import InFlexData
@@ -303,6 +305,12 @@ def test_equiv(num_token, a_dtype, w_dtype, tp, workspace_init):
pc2,
) = init_compute_data(M, K, N, E, a_dtype, w_dtype, num_warps=8)
if current_platform.is_device_capability_family(100):
constraints = {
"is_persistent": True,
}
opt_flags.update_opt_flags_constraints(constraints)
if a_dtype == "bf16" and w_dtype == "mx4":
quant_config = mxfp4_w4a16_moe_quant_config(
w1_scale=pc1,
+1 -1
View File
@@ -82,7 +82,7 @@ def test_mxfp4_loading_and_execution_moe(vllm_runner, model_case: ModelCase):
model_case.model_id,
tensor_parallel_size=model_case.tp,
load_format="dummy",
cudagraph_capture_sizes=[16],
compilation_config={"cudagraph_capture_sizes": [16]},
) as llm:
# Disabled as check_model is broken: https://github.com/vllm-project/vllm/pull/18465#issuecomment-3329880562
# def check_model(model):