Merge branch 'v0.9.2-step3v-int8' into 'v0.9.2-step3v'

add step3v_w8a16

See merge request dcutoolkit/deeplearing/vllm!257

vllm/model_executor/layers/fused_moe/fused_moe_step3vw8a16.py

+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+"""Fused MoE kernel."""
+import functools
+import json
+import os
+import math
+from typing import Any, Callable, Dict, Optional, List, Optional, Tuple
+
+import torch
+import vllm.envs as envs
+import vllm.model_executor.layers.fused_moe.modular_kernel as mk
+from vllm import _custom_ops as ops
+from vllm.logger import init_logger
+# yapf: disable
+from vllm.model_executor.layers.fused_moe.config import (
+    FusedMoEQuantConfig, get_config_quant_dtype)
+from vllm.model_executor.layers.fused_moe.cutlass_moe import (
+    _valid_cutlass_block_scaled_grouped_gemm,
+    run_cutlass_block_scaled_fused_experts)
+# yapf: enable
+from vllm.model_executor.layers.fused_moe.deep_gemm_moe import (
+    _valid_deep_gemm, deep_gemm_moe_fp8)
+from vllm.model_executor.layers.fused_moe.moe_align_block_size import (
+    moe_align_block_size)
+
+try:
+    from lmslim.layers.gemm.int8_utils import (
+    per_token_group_quant_int8, per_token_quant_int8)
+    from lmslim.layers.fused_moe.fuse_moe_int8 import (fused_experts_impl_int8, get_w8a8moe_json)
+    from lmslim.layers.fused_moe.fuse_moe_w4a8 import fused_experts_impl_w4a8
+except Exception:
+    print("INFO: Please install lmslim if you want to infer the quantitative model of moe.\n") 
+
+from vllm.model_executor.layers.fused_moe.prepare_finalize import (
+    MoEPrepareAndFinalizeNoEP)
+from vllm.model_executor.layers.fused_moe.utils import (
+    _resize_cache, moe_kernel_quantize_input)
+from vllm.platforms import current_platform
+from vllm.triton_utils import tl, triton
+from vllm.utils import direct_register_custom_op
+
+# from .rocm_aiter_fused_moe import is_rocm_aiter_moe_enabled
+
+
+logger = init_logger(__name__)
+
+if envs.VLLM_USE_GLOBAL_CACHE13:
+    moe_cache_singleton = None
+    
+
+@torch.compile
+def moe_sum_reduce_torch_compile(x, out, routed_scaling_factor):
+    torch.sum(x, dim=1, out=out)
+    out.mul_(routed_scaling_factor)
+
+
+@triton.jit
+def _moe_sum_reduce_kernel(
+        input_ptr,
+        input_stride_0,
+        input_stride_1,
+        input_stride_2,
+        output_ptr,
+        output_stride_0,
+        output_stride_1,
+        token_num: int,
+        topk_num: int,
+        hidden_dim: int,
+        routed_scaling_factor: tl.constexpr,
+        BLOCK_M: tl.constexpr,
+        BLOCK_DIM: tl.constexpr,
+        NUM_STAGE: tl.constexpr,
+):
+    input_stride_0 = tl.cast(input_stride_0, dtype=tl.int64)
+    input_stride_1 = tl.cast(input_stride_1, dtype=tl.int64)
+    output_stride_0 = tl.cast(output_stride_0, dtype=tl.int64)
+
+    token_block_id = tl.program_id(0)
+    dim_block_id = tl.program_id(1)
+
+    token_start = token_block_id * BLOCK_M
+    token_end = min((token_block_id + 1) * BLOCK_M, token_num)
+
+    dim_start = dim_block_id * BLOCK_DIM
+    dim_end = min((dim_block_id + 1) * BLOCK_DIM, hidden_dim)
+
+    offs_dim = dim_start + tl.arange(0, BLOCK_DIM)
+
+    for token_index in range(token_start, token_end):
+        accumulator = tl.zeros((BLOCK_DIM,), dtype=tl.float32)
+        input_t_ptr = input_ptr + token_index * input_stride_0 + offs_dim
+        for i in tl.range(0, topk_num, num_stages=NUM_STAGE):
+            tmp = tl.load(
+                input_t_ptr + i * input_stride_1, mask=offs_dim < dim_end, other=0.0
+            )
+            accumulator += tmp
+        accumulator = accumulator * routed_scaling_factor
+        store_t_ptr = output_ptr + token_index * output_stride_0 + offs_dim
+        tl.store(
+            store_t_ptr,
+            accumulator.to(input_ptr.dtype.element_ty),
+            mask=offs_dim < dim_end,
+        )
+
+
+def moe_sum_reduce_triton(
+        input: torch.Tensor, output: torch.Tensor, routed_scaling_factor: float
+):
+    assert input.is_contiguous()
+    assert output.is_contiguous()
+
+    token_num, topk_num, hidden_dim = input.shape
+    assert output.shape[0] == token_num and output.shape[1] == hidden_dim
+
+    if token_num <= 32:
+        BLOCK_M = 1
+        BLOCK_DIM = 512
+        NUM_STAGE = 2
+        num_warps = 4
+
+    elif token_num <= 128:
+        BLOCK_M = 1
+        BLOCK_DIM = 1024
+        NUM_STAGE = 0
+        num_warps = 2
+
+    elif token_num <= 4096:
+        BLOCK_M = 1
+        BLOCK_DIM = 2048
+        NUM_STAGE = 0
+        num_warps = 2
+    else:
+        BLOCK_M = 1
+        BLOCK_DIM = 2048
+        NUM_STAGE = 2
+        num_warps = 8
+
+    grid = (
+        triton.cdiv(token_num, BLOCK_M),
+        triton.cdiv(hidden_dim, BLOCK_DIM),
+    )
+
+    _moe_sum_reduce_kernel[grid](
+        input,
+        *input.stride(),
+        output,
+        *output.stride(),
+        token_num=token_num,
+        topk_num=topk_num,
+        hidden_dim=hidden_dim,
+        routed_scaling_factor=routed_scaling_factor,
+        BLOCK_M=BLOCK_M,
+        BLOCK_DIM=BLOCK_DIM,
+        NUM_STAGE=NUM_STAGE,
+        num_warps=num_warps,
+    )
+    return
+
+
+def moe_reduce_dispatch(
+    intermediate_cache3: torch.Tensor,
+    out_hidden_states: torch.Tensor,
+    begin_chunk_idx: int,
+    end_chunk_idx: int,
+):
+    inter_cache_view = intermediate_cache3.view(*intermediate_cache3.shape)
+    n = intermediate_cache3.shape[0]
+
+    # 根据 n 大小选择不同的 reduce 实现
+    if 1 <= n <= 4:
+        moe_sum_reduce_torch_compile(inter_cache_view, out_hidden_states[begin_chunk_idx:end_chunk_idx], 1.0)
+    elif 4 < n <= 1024:
+        moe_sum_reduce_triton(inter_cache_view, out_hidden_states[begin_chunk_idx:end_chunk_idx], 1.0)
+    elif 1024 < n <= 32768:
+        ops.moe_sum_opt1(inter_cache_view, out_hidden_states[begin_chunk_idx:end_chunk_idx])
+    else:
+        ops.moe_sum(inter_cache_view, out_hidden_states[begin_chunk_idx:end_chunk_idx])
+
+ 
+def get_moe_cache(top_k_num,N,K,device,dtype):
+    global moe_cache_singleton
+    if moe_cache_singleton is None:
+        moe_cache_singleton = torch.empty(envs.VLLM_FUSED_MOE_CHUNK_SIZE * top_k_num *max(N, K), device=device, dtype=dtype)
+        logger.info(f"Initializing moe_cache_singleton shape: {moe_cache_singleton.shape}, memory: {moe_cache_singleton.element_size() * moe_cache_singleton.numel() / 1024**2:.2f} MB")
+    return moe_cache_singleton
+
+@triton.jit
+def write_zeros_to_output(c_ptr, stride_cm, stride_cn, pid_n, N, offs_token,
+                          token_mask, BLOCK_SIZE_M, BLOCK_SIZE_N,
+                          compute_type):
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=compute_type)
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[
+        None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+@triton.jit
+def fused_moe_kernel_awq(
+        # Pointers to matrices
+        a_ptr, # [4, 7168] 
+        b_ptr, # [256, 512, 3584]
+        c_ptr, # (8, 8, 512)
+        b_scale_ptr, # (256, 512, 56)
+        b_zp_ptr, # (256, 256, 56)
+        topk_weights_ptr, 
+        sorted_token_ids_ptr, # [0, 1, 2, 3, 4]
+        expert_ids_ptr,
+        num_tokens_post_padded_ptr,
+        # Matrix dimensions
+        N: tl.constexpr,
+        K: tl.constexpr,
+        EM, # pading后的总索引长度
+        num_valid_tokens, # 有效索引的上限
+        # The stride variables represent how much to increase the ptr by when
+        # moving by 1 element in a particular dimension. E.g. `stride_am` is
+        # how much to increase `a_ptr` by to get the element one row down
+        # (A has M rows).
+        stride_am,
+        stride_ak,
+        stride_be,
+        stride_bk, #1
+        stride_bn,
+        stride_cm,
+        stride_cn,
+        stride_bse,
+        stride_bsk,#1
+        stride_bsn,
+        stride_bze,
+        stride_bzk,
+        stride_bzn,
+        block_k_diviable: tl.constexpr,
+        group_size: tl.constexpr, # 128
+        # Meta-parameters
+        BLOCK_SIZE_M: tl.constexpr,
+        BLOCK_SIZE_N: tl.constexpr,
+        BLOCK_SIZE_K: tl.constexpr,
+        GROUP_SIZE_M: tl.constexpr,
+        MUL_ROUTED_WEIGHT: tl.constexpr,
+        top_k: tl.constexpr,
+        compute_type: tl.constexpr,
+        has_zp: tl.constexpr,
+        use_int4_w4a16: tl.constexpr,
+        use_int8_w8a16: tl.constexpr):
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id) # [block_m]
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N +
+               tl.arange(0, BLOCK_SIZE_N)) % N # [block_n]
+    offs_k = tl.arange(0, BLOCK_SIZE_K) # 0, 1, 2, ...... , 127 # # [block_k]
+    offs_k2 = tl.arange(0, BLOCK_SIZE_K // 2) # 0, 1, 2, ...... , 127 # # [block_k]
+    a_ptrs = a_ptr + (offs_token[:, None] // top_k * stride_am +
+                      offs_k[None, :] * stride_ak) # [block_m, block_k] 
+
+    off_experts = tl.load(expert_ids_ptr + pid_m)
+
+    if use_int4_w4a16:
+        # [0, 1, 2, ...... , 126, 127] --> [0, 0, 1, 1 ...... , 63, 63] 
+        # [128, 129, 130, ...... , 254, 255] --> [64, 64, 65, 65 ...... , 127, 127] 
+        
+        # b_ptrs = b_ptr + off_experts * stride_be + \
+        #     (offs_k[:, None] // 2) * stride_bk + offs_bn[None, :] * stride_bn
+        b_ptrs = b_ptr + off_experts * stride_be + \
+            offs_bn[:, None] * stride_bn + (offs_k2[None, :]) * stride_bk 
+        # tl.device_print("stride_bn",stride_bsn)>1
+        # tl.device_print("stride_bk",stride_bk)=1
+        b_shifter = (offs_k[:, None] % 2) * 4  # 0, 4
+    elif use_int8_w8a16:
+        b_ptrs = b_ptr + off_experts * stride_be + \
+            offs_k[:, None] * stride_bk + offs_bn[None, :] * stride_bn
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4 #  0, 4
+
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(a_ptrs,
+                    mask=token_mask[:, None] &
+                    (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+                    other=0.0)
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = tl.interleave(b, b)
+            b= b.trans()
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = b_scale_ptr + off_experts * stride_bse + \
+            offs_bn[None, :] * stride_bsk + \
+            ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsn
+        qzeros_scles = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+
+        scales_int16 = tl.cast(qzeros_scles,tl.uint16)
+        b_scale = tl.cast(scales_int16,tl.float16,bitcast=True)
+        # tl.device_print("b_scale dequant",b_scale)
+
+        mid = qzeros_scles >> 16
+        # b_zp = tl.cast(mid,tl.float16,bitcast=False)
+        b_zp = tl.cast(mid,tl.float16)
+        # b_zp = tl.cast(zeros_int16,tl.float16,bitcast=False)
+
+        # tl.device_print("bzp",b_zp)
+
+        # We accumulate along the K dimension.
+        b = ((b - b_zp) * b_scale).to(tl.float16)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token,
+                             mask=token_mask,
+                             other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[
+        None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
+@triton.jit
+def fused_moe_kernel_gptq_awq(
+        # Pointers to matrices
+        a_ptr,
+        b_ptr,
+        c_ptr,
+        b_scale_ptr,
+        b_zp_ptr,
+        topk_weights_ptr,
+        sorted_token_ids_ptr,
+        expert_ids_ptr,
+        num_tokens_post_padded_ptr,
+        # Matrix dimensions
+        N: tl.constexpr,
+        K: tl.constexpr,
+        EM,
+        num_valid_tokens,
+        # The stride variables represent how much to increase the ptr by when
+        # moving by 1 element in a particular dimension. E.g. `stride_am` is
+        # how much to increase `a_ptr` by to get the element one row down
+        # (A has M rows).
+        stride_am,
+        stride_ak,
+        stride_be,
+        stride_bk,
+        stride_bn,
+        stride_cm,
+        stride_cn,
+        stride_bse,
+        stride_bsk,
+        stride_bsn,
+        stride_bze,
+        stride_bzk,
+        stride_bzn,
+        block_k_diviable: tl.constexpr,
+        group_size: tl.constexpr,
+        # Meta-parameters
+        BLOCK_SIZE_M: tl.constexpr,
+        BLOCK_SIZE_N: tl.constexpr,
+        BLOCK_SIZE_K: tl.constexpr,
+        GROUP_SIZE_M: tl.constexpr,
+        MUL_ROUTED_WEIGHT: tl.constexpr,
+        top_k: tl.constexpr,
+        compute_type: tl.constexpr,
+        has_zp: tl.constexpr,
+        use_int4_w4a16: tl.constexpr,
+        use_int8_w8a16: tl.constexpr):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    group_id = pid // num_pid_in_group
+    first_pid_m = group_id * GROUP_SIZE_M
+    group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(
+        tl.int64)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    offs_bn = (pid_n * BLOCK_SIZE_N +
+               tl.arange(0, BLOCK_SIZE_N).to(tl.int64)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (offs_token[:, None] // top_k * stride_am +
+                      offs_k[None, :] * stride_ak)    
+
+    off_experts = tl.load(expert_ids_ptr + pid_m)
+
+    if use_int4_w4a16:
+        b_ptrs = b_ptr + off_experts * stride_be + \
+            (offs_k[:, None] // 2) * stride_bk + offs_bn[None, :] * stride_bn
+        b_shifter = (offs_k[:, None] % 2) * 4
+    elif use_int8_w8a16:
+        b_ptrs = b_ptr + off_experts * stride_be + \
+            offs_k[:, None] * stride_bk + offs_bn[None, :] * stride_bn
+
+    if not has_zp and use_int4_w4a16:
+        b_zp_num = 8
+    if not has_zp and use_int8_w8a16:
+        b_zp_num = 128
+    elif has_zp and use_int4_w4a16:
+        b_zp_shifter = (offs_bn[None, :] % 2) * 4
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+
+        if not block_k_diviable:
+            k_mask = offs_k[:, None] < K - k * BLOCK_SIZE_K
+            k_other = 0.0
+        else:
+            k_mask = None
+            k_other = None
+
+        a = tl.load(a_ptrs,
+                    mask=token_mask[:, None] &
+                    (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+                    other=0.0)
+        b = tl.load(b_ptrs)
+        if use_int4_w4a16:
+            b = (b >> b_shifter) & 0xF
+
+        b_scale_ptrs = b_scale_ptr + off_experts * stride_bse + \
+            offs_bn[None, :] * stride_bsn + \
+            ((offs_k[:, None] + BLOCK_SIZE_K * k) // group_size) * stride_bsk
+        b_scale = tl.load(b_scale_ptrs, mask=k_mask, other=k_other)
+        b_scale = b_scale.to(tl.float32)
+
+        if has_zp and use_int4_w4a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = b_zp_ptr + off_experts * stride_bze + \
+                (offs_bn[None, :] // 2) * stride_bzn + \
+                offs_k_true * stride_bzk
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = ((b_zp >> b_zp_shifter) & 0xF)
+            b_zp = b_zp.to(tl.float32)
+        elif has_zp and use_int8_w8a16:
+            offs_k_true = (offs_k[:, None] + BLOCK_SIZE_K * k) // group_size
+            b_zp_ptrs = b_zp_ptr + off_experts * stride_bze + \
+                offs_bn[None, :] * stride_bzn + \
+                offs_k_true * stride_bzk
+            b_zp = tl.load(b_zp_ptrs, mask=k_mask, other=k_other)
+            b_zp = b_zp.to(tl.float32)
+
+        # We accumulate along the K dimension.
+        if has_zp:
+            b = ((b.to(tl.float32) - b_zp) * b_scale).to(compute_type)
+        else:
+            b = ((b.to(tl.float32) - b_zp_num) * b_scale).to(compute_type)
+        accumulator = tl.dot(a, b, acc=accumulator)
+
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        if use_int4_w4a16:
+            b_ptrs += (BLOCK_SIZE_K // 2) * stride_bk
+        else:
+            b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token,
+                             mask=token_mask,
+                             other=0)
+        accumulator = accumulator * moe_weight[:, None]
+
+    accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[
+        None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
+@triton.jit
+def fused_moe_kernel(
+    # Pointers to matrices
+    a_ptr,
+    b_ptr,
+    c_ptr,
+    a_scale_ptr,
+    b_scale_ptr,
+    topk_weights_ptr,
+    sorted_token_ids_ptr,
+    expert_ids_ptr,
+    num_tokens_post_padded_ptr,
+    # Matrix dimensions
+    N,
+    K,
+    EM,
+    num_valid_tokens,
+    # The stride variables represent how much to increase the ptr by when
+    # moving by 1 element in a particular dimension. E.g. `stride_am` is
+    # how much to increase `a_ptr` by to get the element one row down
+    # (A has M rows).
+    stride_am,
+    stride_ak,
+    stride_be,
+    stride_bk,
+    stride_bn,
+    stride_cm,
+    stride_cn,
+    stride_asm,
+    stride_ask,
+    stride_bse,
+    stride_bsk,
+    stride_bsn,
+    # Block size for block-wise quantization
+    group_n: tl.constexpr,
+    group_k: tl.constexpr,
+    # Meta-parameters
+    BLOCK_SIZE_M: tl.constexpr,
+    BLOCK_SIZE_N: tl.constexpr,
+    BLOCK_SIZE_K: tl.constexpr,
+    GROUP_SIZE_M: tl.constexpr,
+    MUL_ROUTED_WEIGHT: tl.constexpr,
+    top_k: tl.constexpr,
+    compute_type: tl.constexpr,
+    use_fp8_w8a8: tl.constexpr,
+    use_int8_w8a8: tl.constexpr,
+    use_int8_w8a16: tl.constexpr,
+    per_channel_quant: tl.constexpr,
+):
+    """
+    Implements the fused computation for a Mixture of Experts (MOE) using
+    token and expert matrices.
+
+    Key Parameters:
+    - A: The input tensor representing tokens with shape (*, K), where '*' can
+        be any shape representing batches and K is the feature dimension of
+        each token.
+    - B: The stacked MOE weight tensor with shape (E, N, K), where E is
+        the number of experts, K is the input feature dimension, and N is
+        the output feature dimension.
+    - C: The output cache tensor with shape (M, topk, N), where M is the
+        total number of tokens post padding, topk is the number of times
+        each token is repeated, and N is the output feature dimension.
+    - sorted_token_ids: A tensor containing the sorted indices of tokens,
+        repeated topk times and arranged by the expert index they are
+        assigned to.
+    - expert_ids: A tensor containing the indices of the expert for each
+        block. It determines which expert matrix from B should be used for
+        each block in A.
+    This kernel performs the multiplication of a token by its corresponding
+    expert matrix as determined by `expert_ids`. The sorting of
+    `sorted_token_ids` by expert index and padding ensures divisibility by
+    BLOCK_SIZE_M, which is necessary to maintain consistency in block matrix
+    multiplication across different blocks processed by the same expert.
+    """
+    # -----------------------------------------------------------
+    # Map program ids `pid` to the block of C it should compute.
+    # This is done in a grouped ordering to promote L2 data reuse.
+    pid = tl.program_id(axis=0)
+    # num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+    # num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+    # num_pid_in_group = GROUP_SIZE_M * num_pid_n
+    # group_id = pid // num_pid_in_group
+    # first_pid_m = group_id * GROUP_SIZE_M
+    # group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+    # pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+    # pid_n = (pid % num_pid_in_group) // group_size_m
+    if GROUP_SIZE_M ==1:
+        num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+        pid_m = pid // num_pid_n
+        pid_n = pid % num_pid_n
+    else:
+        num_pid_m = tl.cdiv(EM, BLOCK_SIZE_M)
+        num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
+        num_pid_in_group = GROUP_SIZE_M * num_pid_n
+        group_id = pid // num_pid_in_group
+        first_pid_m = group_id * GROUP_SIZE_M
+        group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
+        pid_m = first_pid_m + ((pid % num_pid_in_group) % group_size_m)
+        pid_n = (pid % num_pid_in_group) // group_size_m
+
+    # ----------------------------------------------------------
+    # Create pointers for the first blocks of A and B.
+    # We will advance this pointer as we move in the K direction
+    # and accumulate
+    # `a_ptrs` is a block of [BLOCK_SIZE_M, BLOCK_SIZE_K] pointers
+    # `b_ptrs` is a block of [BLOCK_SIZE_K, BLOCK_SIZE_N] pointers
+    num_tokens_post_padded = tl.load(num_tokens_post_padded_ptr)
+    if pid_m * BLOCK_SIZE_M >= num_tokens_post_padded:
+        return
+
+    offs_token_id = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
+    offs_token = tl.load(sorted_token_ids_ptr + offs_token_id)
+    token_mask = offs_token < num_valid_tokens
+
+    off_experts = tl.load(expert_ids_ptr + pid_m)
+    if off_experts == -1:
+        # -----------------------------------------------------------
+        # Write back zeros to the output when the expert is not
+        # in the current expert parallel rank.
+        write_zeros_to_output(c_ptr, stride_cm, stride_cn, pid_n, N,
+                            offs_token, token_mask, BLOCK_SIZE_M,
+                            BLOCK_SIZE_N, compute_type)
+        return
+
+    offs_bn = (pid_n * BLOCK_SIZE_N +
+               tl.arange(0, BLOCK_SIZE_N)) % N
+    offs_k = tl.arange(0, BLOCK_SIZE_K)
+    a_ptrs = a_ptr + (offs_token[:, None] // top_k * stride_am +
+                      offs_k[None, :] * stride_ak)
+
+    b_ptrs = b_ptr + off_experts * stride_be + (offs_k[:, None] * stride_bk +
+                                                offs_bn[None, :] * stride_bn)
+    if use_int8_w8a16:
+        b_scale_ptrs = b_scale_ptr + off_experts * stride_bse + offs_bn[
+            None, :] * stride_bsn
+        b_scale = tl.load(b_scale_ptrs)
+
+    if use_fp8_w8a8 or use_int8_w8a8:
+        # block-wise
+        if group_k > 0 and group_n > 0:
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            offs_bsn = offs_bn // group_n
+            b_scale_ptrs = (b_scale_ptr + off_experts * stride_bse +
+                            offs_bsn * stride_bsn)
+        # channel-wise
+        elif per_channel_quant:
+            b_scale_ptrs = b_scale_ptr + off_experts * stride_bse + offs_bn[
+                None, :] * stride_bsn
+            b_scale = tl.load(b_scale_ptrs)
+            # Load per-token scale for activations
+            a_scale_ptrs = a_scale_ptr + (offs_token // top_k) * stride_asm
+            a_scale = tl.load(a_scale_ptrs, mask=token_mask, other=0.0)[:,
+                                                                        None]
+        # tensor-wise
+        else:
+            a_scale = tl.load(a_scale_ptr)
+            b_scale = tl.load(b_scale_ptr + off_experts)
+
+    # -----------------------------------------------------------
+    # Iterate to compute a block of the C matrix.
+    # We accumulate into a `[BLOCK_SIZE_M, BLOCK_SIZE_N]` block
+    # of fp32 values for higher accuracy.
+    # `accumulator` will be converted back to fp16 after the loop.
+    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
+    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
+        # Load the next block of A and B, generate a mask by checking the
+        # K dimension.
+        a = tl.load(a_ptrs,
+                    mask=token_mask[:, None] &
+                    (offs_k[None, :] < K - k * BLOCK_SIZE_K),
+                    other=0.0)
+        b = tl.load(b_ptrs,
+                    mask=offs_k[:, None] < K - k * BLOCK_SIZE_K,
+                    other=0.0)
+        # We accumulate along the K dimension.
+        if use_int8_w8a16:
+            accumulator = tl.dot(a, b.to(compute_type), acc=accumulator)
+        elif use_fp8_w8a8 or use_int8_w8a8:
+            if group_k > 0 and group_n > 0:
+                k_start = k * BLOCK_SIZE_K
+                offs_ks = k_start // group_k
+                a_scale = tl.load(a_scale_ptrs + offs_ks * stride_ask,
+                                  mask=token_mask,
+                                  other=0.0)
+                b_scale = tl.load(b_scale_ptrs + offs_ks * stride_bsk)
+
+                accumulator += tl.dot(a, b) * a_scale[:,
+                                                      None] * b_scale[None, :]
+            else:
+                if use_fp8_w8a8:
+                    # acc used to enable fp8_fast_accum
+                    accumulator = tl.dot(a, b, acc=accumulator)
+                else:
+                    accumulator += tl.dot(a, b)
+        else:
+            accumulator += tl.dot(a, b)
+        # Advance the ptrs to the next K block.
+        a_ptrs += BLOCK_SIZE_K * stride_ak
+        b_ptrs += BLOCK_SIZE_K * stride_bk
+
+    if MUL_ROUTED_WEIGHT:
+        moe_weight = tl.load(topk_weights_ptr + offs_token,
+                             mask=token_mask,
+                             other=0)
+        accumulator = accumulator * moe_weight[:, None]
+    if use_int8_w8a16:
+        accumulator = (accumulator * b_scale).to(compute_type)
+    elif use_fp8_w8a8 or use_int8_w8a8:
+        if group_k > 0 and group_n > 0:
+            accumulator = accumulator.to(compute_type)
+        else:
+            accumulator = (accumulator * a_scale * b_scale).to(compute_type)
+    else:
+        accumulator = accumulator.to(compute_type)
+    # -----------------------------------------------------------
+    # Write back the block of the output
+    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
+    c_ptrs = c_ptr + stride_cm * offs_token[:, None] + stride_cn * offs_cn[
+        None, :]
+    c_mask = token_mask[:, None] & (offs_cn[None, :] < N)
+    tl.store(c_ptrs, accumulator, mask=c_mask)
+
+
+def invoke_fused_moe_kernel(A: torch.Tensor,
+                            B: torch.Tensor,
+                            C: torch.Tensor,
+                            A_scale: Optional[torch.Tensor],
+                            B_scale: Optional[torch.Tensor],
+                            B_zp: Optional[torch.Tensor],
+                            topk_weights: Optional[torch.Tensor],
+                            sorted_token_ids: torch.Tensor,
+                            expert_ids: torch.Tensor,
+                            num_tokens_post_padded: torch.Tensor,
+                            mul_routed_weight: bool,
+                            top_k: int,
+                            config: dict[str, Any],
+                            compute_type: tl.dtype,
+                            use_fp8_w8a8: bool,
+                            use_int8_w8a8: bool,
+                            use_int8_w8a16: bool,
+                            use_int4_w4a16: bool,
+                            use_int4_w4a8: bool,
+                            per_channel_quant: bool,
+                            block_shape: Optional[list[int]] = None,
+                            use_nn_moe: Optional[bool]=False) -> None:
+    assert topk_weights is not None or not mul_routed_weight
+    assert topk_weights is None or topk_weights.stride(1) == 1
+    assert sorted_token_ids.stride(0) == 1
+
+    if use_fp8_w8a8 or use_int8_w8a8:
+        assert B_scale is not None
+        assert (block_shape is None
+                or triton.cdiv(B.size(-2), block_shape[0]) == B_scale.size(-2))
+        assert (block_shape is None
+                or triton.cdiv(B.size(-1), block_shape[1]) == B_scale.size(-1))
+
+    elif use_int8_w8a16 or use_int4_w4a16:
+        assert B_scale is not None
+        assert block_shape is None or block_shape[0] == 0
+    else:
+        assert A_scale is None
+        assert B_scale is None
+
+    M = A.size(0)
+    num_tokens = M * top_k
+
+    EM = sorted_token_ids.size(0)
+    if A.size(0) < config["BLOCK_SIZE_M"]:
+        # optimize for small batch_size.
+        # We assume that top_ids of each token is unique, so
+        # so num_valid_experts <= batch_size <= BLOCK_SIZE_M,
+        # and we can skip some invalid blocks.
+        EM = min(sorted_token_ids.size(0),
+                 A.size(0) * top_k * config['BLOCK_SIZE_M'])
+    grid = lambda META: (triton.cdiv(EM, META['BLOCK_SIZE_M']) * triton.cdiv(
+        B.size(1) if not use_nn_moe else B.size(2), META['BLOCK_SIZE_N']), )
+
+    if (use_int8_w8a16 or use_int4_w4a16) and \
+            block_shape is not None and block_shape[1] > 0:
+        assert B_scale is not None and B_scale.ndim == 3
+        assert B_zp is None or B_zp.ndim == 3
+        if os.environ.get('moe_wna16_use_cuda') == '1':
+            use_moe_wna16_cuda = should_moe_wna16_use_cuda(
+            num_valid_tokens=num_tokens,
+            group_size=block_shape[1],
+            num_experts=B.size(0),
+            bit=4 if use_int4_w4a16 else 8)
+
+            config = config.copy()
+            config.update(
+                get_moe_wna16_block_config(config=config,
+                                           use_moe_wna16_cuda=use_moe_wna16_cuda,
+                                           num_valid_tokens=num_tokens,
+                                           size_k=A.size(1),
+                                           size_n=B.size(1),
+                                           num_experts=B.size(1),
+                                           group_size=block_shape[1],
+                                           real_top_k=top_k,
+                                           block_size_m=config["BLOCK_SIZE_M"]))
+
+            if use_moe_wna16_cuda:
+                bit = 4 if use_int4_w4a16 else 8
+                ops.moe_wna16_gemm(A, C, B, B_scale, B_zp,
+                                topk_weights if mul_routed_weight else None,
+                                sorted_token_ids, expert_ids,
+                                num_tokens_post_padded, top_k,
+                                config["BLOCK_SIZE_M"], config["BLOCK_SIZE_N"],
+                                config["BLOCK_SIZE_K"], bit)
+                return
+
+        if os.environ.get('AWQ_MOE_SZ') == '1':
+            fused_moe_kernel_awq[grid](
+                A,
+                B,
+                C,
+                B_scale,
+                B_zp,
+                topk_weights,
+                sorted_token_ids,
+                expert_ids,
+                num_tokens_post_padded,
+                B.size(1),
+                A.size(1),
+                EM,
+                topk_ids.numel(),
+                A.stride(0),
+                A.stride(1),
+                B.stride(0),
+                B.stride(2),
+                B.stride(1),
+                C.stride(1),
+                C.stride(2),
+                B_scale.stride(0),
+                B_scale.stride(2),
+                B_scale.stride(1),
+                B_zp.stride(0) if B_zp is not None else 0,
+                B_zp.stride(2) if B_zp is not None else 0,
+                B_zp.stride(1) if B_zp is not None else 0,
+                block_k_diviable=A.size(1) % config["BLOCK_SIZE_K"] == 0,
+                group_size=block_shape[1],
+                MUL_ROUTED_WEIGHT=mul_routed_weight,
+                top_k=top_k,
+                compute_type=compute_type,
+                has_zp=B_zp is not None,
+                use_int4_w4a16=use_int4_w4a16,
+                use_int8_w8a16=use_int8_w8a16,
+                **config,
+            )
+        else:
+            fused_moe_kernel_gptq_awq[grid](
+                A,
+                B,
+                C,
+                B_scale,
+                B_zp,
+                topk_weights,
+                sorted_token_ids,
+                expert_ids,
+                num_tokens_post_padded,
+                B.size(1),
+                A.size(1),
+                EM,
+                num_tokens,
+                A.stride(0),
+                A.stride(1),
+                B.stride(0),
+                B.stride(2),
+                B.stride(1),
+                C.stride(1),
+                C.stride(2),
+                B_scale.stride(0),
+                B_scale.stride(2),
+                B_scale.stride(1),
+                B_zp.stride(0) if B_zp is not None else 0,
+                B_zp.stride(2) if B_zp is not None else 0,
+                B_zp.stride(1) if B_zp is not None else 0,
+                block_k_diviable=A.size(1) % config["BLOCK_SIZE_K"] == 0,
+                group_size=block_shape[1],
+                MUL_ROUTED_WEIGHT=mul_routed_weight,
+                top_k=top_k,
+                compute_type=compute_type,
+                has_zp=B_zp is not None,
+                use_int4_w4a16=use_int4_w4a16,
+                use_int8_w8a16=use_int8_w8a16,
+                **config,
+            )
+    else:
+        # config = config.copy()
+        # BLOCK_SIZE_K = config.pop("BLOCK_SIZE_K")
+        # if block_shape is not None:
+        #     BLOCK_SIZE_K = min(BLOCK_SIZE_K, min(block_shape[0],
+        #                                          block_shape[1]))
+        fused_moe_kernel[grid](
+            A,
+            B,
+            C,
+            A_scale,
+            B_scale,
+            topk_weights,
+            sorted_token_ids,
+            expert_ids,
+            num_tokens_post_padded,
+            B.size(1) if not use_nn_moe else B.size(2),
+            A.size(1),
+            EM,
+            num_tokens,
+            A.stride(0),
+            A.stride(1),
+            B.stride(0),
+            B.stride(2) if not use_nn_moe else B.stride(1),
+            B.stride(1) if not use_nn_moe else B.stride(2),
+            C.stride(1),
+            C.stride(2),
+            A_scale.stride(0)
+            if A_scale is not None and A_scale.ndim == 2 else 0,
+            A_scale.stride(1)
+            if A_scale is not None and A_scale.ndim == 2 else 0,
+            B_scale.stride(0)
+            if B_scale is not None and B_scale.ndim >= 2 else 0,
+            B_scale.stride(2)
+            if B_scale is not None and B_scale.ndim == 3 else 0,
+            B_scale.stride(1)
+            if B_scale is not None and B_scale.ndim >= 2 else 0,
+            0 if block_shape is None else block_shape[0],
+            0 if block_shape is None else block_shape[1],
+            MUL_ROUTED_WEIGHT=mul_routed_weight,
+            top_k=top_k,
+            compute_type=compute_type,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a8=use_int8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            per_channel_quant=per_channel_quant,
+            **config,
+        )
+
+
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+def get_config_file_name(E: int,
+                         N: int,
+                         dtype: Optional[str],
+                         block_shape: Optional[List[int]] = None, use_nn_moe: Optional[bool] = False) -> str:
+    device_name = current_platform.get_device_name().replace(" ", "_")
+    dtype_selector = "" if not dtype else f",dtype={dtype}"
+    block_shape_selector = ("" if not block_shape or not all(block_shape) else
+                            f",block_shape={block_shape}").replace(" ", "")
+    if not use_nn_moe:
+        return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}.json"  # noqa: E501
+    else:
+        return f"E={E},N={N},device_name={device_name}{dtype_selector}{block_shape_selector}_nn.json"
+
+# Adapted from: https://github.com/sgl-project/sglang/pull/2628
+@functools.lru_cache
+def get_moe_configs(
+    E: int,
+    N: int,
+    dtype: Optional[str],
+    block_n: Optional[int] = None,
+    block_k: Optional[int] = None,
+    use_nn_moe: Optional[bool] = False,
+) -> Optional[Dict[int, Any]]:
+    """
+    Return optimized configurations for the fused MoE kernel.
+
+    The return value will be a dictionary that maps an irregular grid of
+    batch sizes to configurations of the fused_moe kernel. To evaluate the
+    kernel on a given batch size bs, the closest batch size in the grid should
+    be picked and the associated configuration chosen to invoke the kernel.
+    """
+
+    # First look up if an optimized configuration is available in the configs
+    # directory
+    block_shape = [block_n, block_k] if block_n and block_k else None
+    json_file_name = get_config_file_name(E, N, dtype, block_shape, use_nn_moe=use_nn_moe)
+
+    config_file_path = os.path.join(
+        os.path.dirname(os.path.realpath(__file__)), "configs", json_file_name)
+    if torch.cuda.get_device_properties(torch.cuda.current_device()).multi_processor_count == 120:
+        config_file_path_120 = config_file_path.replace(".json","_120.json")
+        if os.path.exists(config_file_path_120):
+            with open(config_file_path_120) as f:
+                logger.info("Using configuration from %s for MoE layer.",
+                            config_file_path_120)
+                # If a configuration has been found, return it
+                return {int(key): val for key, val in json.load(f).items()}
+
+    if os.path.exists(config_file_path):
+        with open(config_file_path) as f:
+            logger.info("Using configuration from %s for MoE layer.",
+                        config_file_path)
+            # If a configuration has been found, return it
+            return {int(key): val for key, val in json.load(f).items()}
+
+    # If no optimized configuration is available, we will use the default
+    # configuration
+    logger.warning(
+        ("Using default MoE config. Performance might be sub-optimal! "
+         "Config file not found at %s"), config_file_path)
+    return None
+
+
+def get_moe_wna16_block_config(config: dict[str,
+                                            int], use_moe_wna16_cuda: bool,
+                               num_valid_tokens: int, size_k: int, size_n: int,
+                               num_experts: int, group_size: int,
+                               real_top_k: int, block_size_m: int):
+    if "BLOCK_SIZE_N" in config and "BLOCK_SIZE_K" in config:
+        # optimal block config is set
+        return {}
+    if not use_moe_wna16_cuda:
+        # triton moe wna16 kernel
+        if num_valid_tokens // real_top_k == 1:
+            # if bs=1, use a smaller BLOCK_SIZE_N
+            return {"BLOCK_SIZE_N": 32, "BLOCK_SIZE_K": 64}
+        else:
+            return {"BLOCK_SIZE_N": 64, "BLOCK_SIZE_K": 32}
+    else:
+        # cuda moe wna16 kernel
+        # set default block_size 128, and increase them when num_blocks
+        # is too large.
+        block_size_n = 128
+        block_size_k = 128
+        if block_size_k <= group_size:
+            block_size_k = group_size
+
+        num_n_blocks = size_k // block_size_k
+        num_k_blocks = size_n // block_size_k
+        num_m_blocks = (num_valid_tokens + block_size_m - 1) / block_size_m + \
+            num_experts
+        if num_valid_tokens // real_top_k <= block_size_m:
+            num_m_blocks = min(num_m_blocks, num_valid_tokens)
+        num_blocks = num_m_blocks * num_n_blocks * num_k_blocks
+
+        if size_k % 256 == 0 and num_blocks >= 256 and \
+                block_size_k < 256:
+            block_size_k = 256
+            num_blocks = num_blocks // (256 // block_size_k)
+
+        if num_m_blocks <= 16 and size_k % (block_size_k * 2) == 0 and \
+                size_k % (block_size_k * 2) == 0 and block_size_k <= 512 and \
+                num_blocks >= 512:
+            block_size_k = block_size_k * 2
+            num_blocks = num_blocks // 2
+
+        if num_blocks > 1024:
+            block_size_n = 256
+            num_n_blocks = num_n_blocks // 2
+            num_blocks = num_blocks // 2
+
+        if size_n <= 1024 and num_blocks >= 1024:
+            # The kernel performance got much better with BLOCK_SIZE_N=1024
+            # when num_blocks is large, event when N is small.
+            # Not sure why, maybe it force the CUDA SM process only one block
+            # at the same time.
+            block_size_n = 1024
+
+        return {"BLOCK_SIZE_N": block_size_n, "BLOCK_SIZE_K": block_size_k}
+
+
+def should_moe_wna16_use_cuda(num_valid_tokens: int, group_size: int,
+                              num_experts: int, bit: int):
+    return bit == 4 and group_size in [32, 64, 128] and \
+        num_valid_tokens / num_experts <= 6
+
+
+def get_default_config(
+    M: int,
+    E: int,
+    N: int,
+    K: int,
+    topk: int,
+    dtype: Optional[str],
+    is_marlin: bool,
+    block_shape: Optional[List[int]] = None,
+    use_nn_moe: Optional[bool]=False,
+) -> dict[str, int]:
+    if dtype == "fp8_w8a8" and block_shape is not None:
+        # Block-wise quant: BLOCK_SIZE_N must be divisible by block_shape[0]
+        # BLOCK_SIZE_K must be divisible by block_shape[1]
+        # num_stages=3 can cause triton.runtime.errors.OutOfResources
+        # on ROCm, set it to 2 instead.
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": block_shape[0],
+            "BLOCK_SIZE_K": block_shape[1],
+            "GROUP_SIZE_M": 32,
+            "num_warps": 4,
+            "num_stages": 3 if not current_platform.is_rocm() else 2,
+        }
+
+    # elif dtype in ["int4_w4a16", "int8_w8a16"] and block_shape is not None:
+    #     # moe wna16 kernels
+    #     # only set BLOCK_SIZE_M
+    #     # BLOCK_SIZE_N and BLOCK_SIZE_K would be set later
+    #     bit = 4 if dtype == "int4_w4a16" else 8
+    #     use_moe_wna16_cuda = should_moe_wna16_use_cuda(M * topk,
+    #                                                    block_shape[1], E, bit)
+    #     if use_moe_wna16_cuda:
+    #         config = {"BLOCK_SIZE_M": min(16, M)}
+    #     elif M <= 20:
+    #         config = {"BLOCK_SIZE_M": 16, "GROUP_SIZE_M": 1}
+    #     elif M <= 40:
+    #         config = {"BLOCK_SIZE_M": 32, "GROUP_SIZE_M": 1}
+    #     else:
+    #         config = {"BLOCK_SIZE_M": 64, "GROUP_SIZE_M": 1}
+    elif is_marlin:
+        for block_size_m in [8, 16, 32, 48, 64]:
+            if M * topk / E / block_size_m < 0.9:
+                break
+        return {"BLOCK_SIZE_M": block_size_m}
+    elif M <= E:
+        config = {
+            "BLOCK_SIZE_M": 16,
+            "BLOCK_SIZE_N": 32,
+            "BLOCK_SIZE_K": 64,
+            "GROUP_SIZE_M": 1,
+        }
+    else:
+        config = {
+            "BLOCK_SIZE_M": 64,
+            "BLOCK_SIZE_N": 64,
+            "BLOCK_SIZE_K": 32,
+            "GROUP_SIZE_M": 8,
+        }
+
+    if use_nn_moe:
+        config["num_ldmatrixes"] = 1
+    return config
+
+
+def try_get_optimal_moe_config(
+    w1_shape: tuple[int, ...],
+    w2_shape: tuple[int, ...],
+    top_k: int,
+    dtype: Optional[str],
+    M: int,
+    is_marlin: bool = False,
+    block_shape: Optional[List[int]] = None,
+    use_nn_moe: Optional[bool] = False,
+) -> dict[str, int]:
+    from vllm.model_executor.layers.fused_moe import get_config
+    override_config = get_config()
+    if override_config:
+        config = override_config
+    else:
+        # First try to load optimal config from the file
+        if not use_nn_moe:
+            E, _, N = w2_shape
+        else:
+            E, N, _  = w2_shape
+        # if dtype == "int4_w4a16":
+        #     N = N * 2
+        block_n = block_shape[0] if block_shape else 0
+        block_k = block_shape[1] if block_shape else 0
+        configs = get_moe_configs(E, N, dtype, block_n, block_k, use_nn_moe=use_nn_moe)
+
+        if configs:
+            # If an optimal configuration map has been found, look up the
+            # optimal config
+            config = configs[min(configs.keys(), key=lambda x: abs(x - M))]
+        else:
+            # Else use the default config
+            config = get_default_config(M, E, N, w1_shape[2] if not use_nn_moe else w1_shape[1], top_k, dtype,
+                                        is_marlin, block_shape, use_nn_moe=use_nn_moe)
+    return config
+
+
+def vllm_topk_softmax(topk_weights: torch.Tensor, topk_indices: torch.Tensor,
+                      token_expert_indices: torch.Tensor,
+                      gating_output: torch.Tensor,
+                      renormalize: bool) -> tuple[torch.Tensor, ...]:
+    ops.topk_softmax(
+        topk_weights,
+        topk_indices,
+        token_expert_indices,
+        gating_output,
+    )
+    if renormalize:
+        topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
+
+    return topk_weights, topk_indices
+
+
+def dispatch_topk_func() -> Callable[..., tuple[torch.Tensor, ...]]:
+    # if is_rocm_aiter_moe_enabled():
+    #     from .rocm_aiter_fused_moe import rocm_aiter_topk_softmax
+    #     return rocm_aiter_topk_softmax
+    return vllm_topk_softmax
+
+
+def fused_topk(
+    hidden_states: torch.Tensor,
+    gating_output: torch.Tensor,
+    topk: int,
+    renormalize: bool,
+    indices_type: Optional[torch.dtype] = None,
+) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    assert hidden_states.size(0) == gating_output.size(0), (
+        "Number of tokens mismatch")
+
+    M, _ = hidden_states.size()
+
+    topk_weights = torch.empty(M,
+                               topk,
+                               dtype=torch.float32,
+                               device=hidden_states.device)
+    topk_ids = torch.empty(
+        M,
+        topk,
+        dtype=torch.int32 if indices_type is None else indices_type,
+        device=hidden_states.device)
+    token_expert_indices = torch.empty(M,
+                                       topk,
+                                       dtype=torch.int32,
+                                       device=hidden_states.device)
+
+    gating_output_float = gating_output.float()  # TODO(woosuk): Optimize this.
+
+    topk_func = dispatch_topk_func()
+    topk_weights, topk_ids = topk_func(topk_weights, topk_ids,
+                                       token_expert_indices,
+                                       gating_output_float, renormalize)
+
+    return topk_weights, topk_ids, token_expert_indices
+
+
+def is_power_of_two(n):
+    return n > 0 and math.log2(n).is_integer()
+
+
+# This is used by the Deepseek-V2 and Deepseek-V3 model
+@torch.compile(dynamic=True, backend=current_platform.simple_compile_backend)
+def grouped_topk(
+    hidden_states: torch.Tensor,
+    gating_output: torch.Tensor,
+    topk: int,
+    renormalize: bool,
+    num_expert_group: int = 0,
+    topk_group: int = 0,
+    scoring_func: str = "softmax",
+    e_score_correction_bias: Optional[torch.Tensor] = None
+) -> tuple[torch.Tensor, torch.Tensor]:
+
+    assert hidden_states.size(0) == gating_output.size(0), (
+        "Number of tokens mismatch")
+
+    if scoring_func == "softmax":
+        scores = torch.softmax(gating_output, dim=-1)
+    elif scoring_func == "sigmoid":
+        scores = gating_output.sigmoid()
+    else:
+        raise ValueError(f"Unsupported scoring function: {scoring_func}")
+
+    num_token = scores.size(0)
+    if e_score_correction_bias is not None:
+        # Store original scores before applying correction bias. We use biased
+        # scores for expert selection but original scores for routing weights
+        original_scores = scores
+        scores = scores + e_score_correction_bias.unsqueeze(0)
+        group_scores = (scores.view(num_token, num_expert_group,
+                                    -1).topk(2, dim=-1)[0].sum(dim=-1))
+    else:
+        group_scores = scores.view(num_token, num_expert_group,
+                                   -1).max(dim=-1).values  # [n, n_group]
+    group_idx = torch.topk(group_scores, k=topk_group, dim=-1,
+                           sorted=False)[1]  # [n, top_k_group]
+    group_mask = torch.zeros_like(group_scores)  # [n, n_group]
+    group_mask.scatter_(1, group_idx, 1)  # [n, n_group]
+    score_mask = group_mask.unsqueeze(-1).expand(
+        num_token, num_expert_group,
+        scores.size(-1) // num_expert_group).reshape(num_token, -1)  # [n, e]
+    tmp_scores = scores.masked_fill(~score_mask.bool(),
+                                    float("-inf"))  # [n, e]
+
+    if e_score_correction_bias is not None:
+        topk_ids = torch.topk(tmp_scores, k=topk, dim=-1, sorted=False)[1]
+        # Use original unbiased scores for the routing weights
+        topk_weights = original_scores.gather(1, topk_ids)
+    else:
+        topk_weights, topk_ids = torch.topk(tmp_scores,
+                                            k=topk,
+                                            dim=-1,
+                                            sorted=False)
+
+    if renormalize:
+        topk_weights = topk_weights / topk_weights.sum(dim=-1, keepdim=True)
+
+    return topk_weights.to(torch.float32), topk_ids.to(torch.int32)
+
+
+def get_config_dtype_str(
+        dtype: torch.dtype,
+        use_int4_w4a16: Optional[bool] = False,
+        use_int8_w8a16: Optional[bool] = False,
+        use_fp8_w8a8: Optional[bool] = False,
+        use_int8_w8a8: Optional[bool] = False,
+        use_int4_w4a8: Optional[bool] = False) -> Optional[str]:
+    if use_fp8_w8a8:
+        return "fp8_w8a8"
+    elif use_int8_w8a8:
+        return "int8_w8a8"
+    elif use_int8_w8a16:
+        return "int8_w8a16"
+    elif use_int4_w4a16:
+        return "int4_w4a16"
+    elif use_int4_w4a8:
+        return "int4_w4a8"
+    elif dtype == torch.float:
+        # avoiding cases where kernel fails when float32 MoE
+        # use fp16/bfloat16 configs
+        return "float32"
+    return None
+
+
+def inplace_fused_experts_step3v(hidden_states: torch.Tensor,
+                          w1: torch.Tensor,
+                          w2: torch.Tensor,
+                          topk_weights: torch.Tensor,
+                          topk_ids: torch.Tensor,
+                          activation: Optional[str] = None,
+                          apply_router_weight_on_input: bool = False,
+                          use_fp8_w8a8: bool = False,
+                          use_int8_w8a8: bool = False,
+                          use_int8_w8a16: bool = False,
+                          use_int4_w4a16: bool = False,
+                          use_int4_w4a8: bool =False,
+                          per_channel_quant: bool = False,
+                          global_num_experts: int = -1,
+                          expert_map: Optional[torch.Tensor] = None,
+                          w1_scale: Optional[torch.Tensor] = None,
+                          w2_scale: Optional[torch.Tensor] = None,
+                          w1_zp: Optional[torch.Tensor] = None,
+                          w2_zp: Optional[torch.Tensor] = None,
+                          a1_scale: Optional[torch.Tensor] = None,
+                          a2_scale: Optional[torch.Tensor] = None,
+                          block_shape: Optional[List[int]] = None,
+                          use_nn_moe: Optional[bool] = False,
+                          shared_output: Optional[torch.Tensor] = None,
+                          routed_scaling_factor: Optional[float] = 1.0,
+                          use_step3v_w8a16: Optional[bool] = False) -> None:
+    fused_experts_impl(hidden_states, w1, w2, topk_weights, topk_ids, True,
+                       activation, apply_router_weight_on_input, use_fp8_w8a8,
+                       use_int8_w8a8, use_int8_w8a16, use_int4_w4a16,use_int4_w4a8,
+                       per_channel_quant, global_num_experts, expert_map,
+                       w1_scale, w2_scale, w1_zp, w2_zp, a1_scale, a2_scale,
+                       block_shape, use_nn_moe, shared_output, routed_scaling_factor,use_step3v_w8a16)
+
+
+def inplace_fused_experts_step3v_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        activation: Optional[str] = None,
+        apply_router_weight_on_input: bool = False,
+        use_fp8_w8a8: bool = False,
+        use_int8_w8a8: bool = False,
+        use_int8_w8a16: bool = False,
+        use_int4_w4a16: bool = False,
+        use_int4_w4a8: bool =False,
+        per_channel_quant: bool = False,
+        global_num_experts: int = -1,
+        expert_map: Optional[torch.Tensor] = None,
+        w1_scale: Optional[torch.Tensor] = None,
+        w2_scale: Optional[torch.Tensor] = None,
+        w1_zp: Optional[torch.Tensor] = None,
+        w2_zp: Optional[torch.Tensor] = None,
+        a1_scale: Optional[torch.Tensor] = None,
+        a2_scale: Optional[torch.Tensor] = None,
+        block_shape: Optional[List[int]] = None,
+        use_nn_moe: Optional[bool] = False,
+        shared_output: Optional[torch.Tensor] = None,
+        routed_scaling_factor: Optional[float] = 1.0) -> None:
+    pass
+
+
+direct_register_custom_op(
+    op_name="inplace_fused_experts_step3v",
+    op_func=inplace_fused_experts_step3v,
+    mutates_args=["hidden_states"],
+    fake_impl=inplace_fused_experts_step3v_fake,
+    tags=(torch.Tag.needs_fixed_stride_order, ),
+)
+
+
+def outplace_fused_experts_step3v(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        activation: Optional[str] = None,
+        apply_router_weight_on_input: bool = False,
+        use_fp8_w8a8: bool = False,
+        use_int8_w8a8: bool = False,
+        use_int8_w8a16: bool = False,
+        use_int4_w4a16: bool = False,
+        use_int4_w4a8: bool =False,
+        per_channel_quant: bool = False,
+        global_num_experts: int = -1,
+        expert_map: Optional[torch.Tensor] = None,
+        w1_scale: Optional[torch.Tensor] = None,
+        w2_scale: Optional[torch.Tensor] = None,
+        w1_zp: Optional[torch.Tensor] = None,
+        w2_zp: Optional[torch.Tensor] = None,
+        a1_scale: Optional[torch.Tensor] = None,
+        a2_scale: Optional[torch.Tensor] = None,
+        block_shape: Optional[List[int]] = None,
+        use_nn_moe: Optional[bool] = False,
+        shared_output: Optional[torch.Tensor] = None,
+        routed_scaling_factor: Optional[float] = 1.0) -> torch.Tensor:
+    return fused_experts_impl(hidden_states, w1, w2, topk_weights, topk_ids,
+                              False, activation, apply_router_weight_on_input,
+                              use_fp8_w8a8, use_int8_w8a8, use_int8_w8a16,
+                              use_int4_w4a16,use_int4_w4a8, per_channel_quant,
+                              global_num_experts, expert_map, w1_scale,
+                              w2_scale, w1_zp, w2_zp, a1_scale, a2_scale,
+                              block_shape, use_nn_moe, shared_output, routed_scaling_factor)
+
+
+def outplace_fused_experts_step3v_fake(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        activation: Optional[str] = None,
+        use_fp8_w8a8: bool = False,
+        use_int8_w8a8: bool = False,
+        use_int8_w8a16: bool = False,
+        use_int4_w4a16: bool = False,
+        use_int4_w4a8: bool =False,
+        per_channel_quant: bool = False,
+        global_num_experts: int = -1,
+        expert_map: Optional[torch.Tensor] = None,
+        w1_scale: Optional[torch.Tensor] = None,
+        w2_scale: Optional[torch.Tensor] = None,
+        w1_zp: Optional[torch.Tensor] = None,
+        w2_zp: Optional[torch.Tensor] = None,
+        a1_scale: Optional[torch.Tensor] = None,
+        a2_scale: Optional[torch.Tensor] = None,
+        block_shape: Optional[List[int]] = None,
+        use_nn_moe: Optional[bool] = False,
+        shared_output: Optional[torch.Tensor] = None,
+        routed_scaling_factor: Optional[float] = 1.0) -> torch.Tensor:
+    return torch.empty_like(hidden_states)
+
+
+direct_register_custom_op(
+    op_name="outplace_fused_experts_step3v",
+    op_func=outplace_fused_experts_step3v,
+    mutates_args=[],
+    fake_impl=outplace_fused_experts_step3v_fake,
+    tags=(torch.Tag.needs_fixed_stride_order, ),
+)
+
+
+def torch_vllm_inplace_fused_experts(**kwargs) -> torch.Tensor:
+    torch.ops.vllm.inplace_fused_experts_step3v(**kwargs)
+    hidden_states = kwargs['hidden_states']
+    return hidden_states
+
+
+def torch_vllm_outplace_fused_experts(**kwargs) -> torch.Tensor:
+    return torch.ops.vllm.outplace_fused_experts(**kwargs)
+
+
+def dispatch_fused_experts_func(inplace: bool) -> Callable[..., torch.Tensor]:
+    if inplace:
+        return torch_vllm_inplace_fused_experts
+    return torch_vllm_outplace_fused_experts
+
+
+# TODO (bnell): replace this with modular op.  Can get rid of inplace/outplace
+# torch ops.
+def fused_experts(
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        topk_weights: torch.Tensor,
+        topk_ids: torch.Tensor,
+        inplace: bool = False,
+        activation: str = "silu",
+        apply_router_weight_on_input: bool = False,
+        use_fp8_w8a8: bool = False,
+        use_int8_w8a8: bool = False,
+        use_int8_w8a16: bool = False,
+        use_int4_w4a16: bool = False,
+        use_int4_w4a8: bool =False,
+        per_channel_quant: bool = False,
+        global_num_experts: int = -1,
+        expert_map: Optional[torch.Tensor] = None,
+        w1_scale: Optional[torch.Tensor] = None,
+        w2_scale: Optional[torch.Tensor] = None,
+        w1_zp: Optional[torch.Tensor] = None,
+        w2_zp: Optional[torch.Tensor] = None,
+        a1_scale: Optional[torch.Tensor] = None,
+        a2_scale: Optional[torch.Tensor] = None,
+        block_shape: Optional[List[int]] = None,
+        allow_deep_gemm: bool = False,
+        allow_cutlass_block_scaled_grouped_gemm: bool = False,
+        use_nn_moe: Optional[bool] = False,
+        shared_output: Optional[torch.Tensor] = None,
+        routed_scaling_factor: Optional[float] = 1.0,
+        use_step3v_w8a16: bool = False,) -> torch.Tensor:
+    # For now, disable DeepGemm for small N (<= 512) until better
+    # permute/unpermute ops are available.
+    N = w1.size(1)
+    if (allow_deep_gemm and use_fp8_w8a8 and N > 512
+            and _valid_deep_gemm(hidden_states, w1, w2)):
+        assert apply_router_weight_on_input is False
+        return deep_gemm_moe_fp8(
+            hidden_states=hidden_states,
+            w1=w1,
+            w2=w2,
+            topk_weights=topk_weights,
+            topk_ids=topk_ids,
+            inplace=inplace,
+            activation=activation,
+            global_num_experts=global_num_experts,
+            expert_map=expert_map,
+            w1_scale=w1_scale,
+            w2_scale=w2_scale,
+            a1_scale=a1_scale,
+            a2_scale=a2_scale,
+            apply_router_weight_on_input=apply_router_weight_on_input,
+        )
+    elif (allow_cutlass_block_scaled_grouped_gemm and use_fp8_w8a8
+          and _valid_cutlass_block_scaled_grouped_gemm(hidden_states, w1, w2)):
+        assert apply_router_weight_on_input is False
+        return run_cutlass_block_scaled_fused_experts(
+            a=hidden_states,
+            w1=w1,
+            w2=w2,
+            w1_scale=w1_scale,
+            w2_scale=w2_scale,
+            topk_weights=topk_weights,
+            topk_ids=topk_ids)
+    else:
+        return dispatch_fused_experts_func(inplace)(
+            hidden_states=hidden_states,
+            w1=w1,
+            w2=w2,
+            topk_weights=topk_weights,
+            topk_ids=topk_ids,
+            activation=activation,
+            apply_router_weight_on_input=apply_router_weight_on_input,
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a8=use_int8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int4_w4a8=use_int4_w4a8,
+            per_channel_quant=per_channel_quant,
+            global_num_experts=global_num_experts,
+            expert_map=expert_map,
+            w1_scale=w1_scale,
+            w2_scale=w2_scale,
+            w1_zp=w1_zp,
+            w2_zp=w2_zp,
+            a1_scale=a1_scale,
+            a2_scale=a2_scale,
+            block_shape=block_shape,
+            use_nn_moe=use_nn_moe,
+            shared_output=shared_output,
+            routed_scaling_factor=routed_scaling_factor,
+            use_step3v_w8a16 =use_step3v_w8a16)
+
+
+def fused_experts_impl(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    inplace: bool = False,
+    activation: str = "silu",
+    apply_router_weight_on_input: bool = False,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    use_int4_w4a8: bool =False,
+    per_channel_quant: bool = False,
+    global_num_experts: int = -1,
+    expert_map: Optional[torch.Tensor] = None,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+    use_nn_moe: Optional[bool] = False,
+    shared_output: Optional[torch.Tensor] = None,
+    routed_scaling_factor: Optional[float] = 1.0,
+    use_step3v_w8a16: Optional[bool] = False
+) -> torch.Tensor:
+    num_tokens = hidden_states.size(0)
+    if use_step3v_w8a16: #step3v w8a16
+        use_nn_moe = True
+    if use_nn_moe:
+        E, _, N = w1.size()
+    else:
+        E, N, _ = w1.size()
+    K = w2.size(1)
+
+    if global_num_experts == -1:
+        global_num_experts = E
+    top_k_num = topk_ids.size(1)
+    # We execute the fused_moe kernel in chunks to circumvent this issue:
+    # https://github.com/vllm-project/vllm/issues/5938
+    CHUNK_SIZE = envs.VLLM_FUSED_MOE_CHUNK_SIZE
+    M = min(num_tokens, CHUNK_SIZE)
+
+    if envs.VLLM_USE_GLOBAL_CACHE13:
+        cache13 = get_moe_cache(top_k_num, N,K if not use_nn_moe else w2.shape[2], device=hidden_states.device, dtype=hidden_states.dtype)
+    else:
+        cache13 = torch.empty(M * top_k_num * max(N, K if not use_nn_moe else w2.shape[2]), device=hidden_states.device, dtype=hidden_states.dtype)
+    if use_int8_w8a8 is True:
+        return fused_experts_impl_int8(hidden_states=hidden_states,
+                                       w1=w1,
+                                       w2=w2,
+                                       topk_weights=topk_weights,
+                                       topk_ids=topk_ids,
+                                       cache13 = cache13,
+                                       inplace=inplace,
+                                       activation=activation,
+                                       apply_router_weight_on_input=apply_router_weight_on_input,
+                                       use_fp8_w8a8=False,
+                                       use_int8_w8a8=True,
+                                       use_int8_w8a16=False,
+                                       use_int4_w4a16=False,
+                                       per_channel_quant=per_channel_quant,
+                                       global_num_experts=global_num_experts,
+                                       expert_map=expert_map,
+                                       w1_scale=w1_scale,
+                                       w2_scale=w2_scale,
+                                       w1_zp=w1_zp,
+                                       w2_zp=w2_zp,
+                                       a1_scale=a1_scale,
+                                       a2_scale=a2_scale,
+                                       block_shape=block_shape,
+                                       use_nn_moe=False,
+                                       shared_output=shared_output,
+                                       routed_scaling_factor=routed_scaling_factor
+                                       )
+    elif use_int4_w4a8 is True:
+        return fused_experts_impl_w4a8(hidden_states=hidden_states,
+                                       w1=w1,
+                                       w2=w2,
+                                       topk_weights=topk_weights,
+                                        topk_ids=topk_ids,
+                                        inplace=inplace,
+                                        cache13 = cache13,
+                                        activation=activation,
+                                        apply_router_weight_on_input=apply_router_weight_on_input,
+                                        use_fp8_w8a8= False,
+                                        use_int8_w8a8= False,
+                                        use_int8_w8a16= False,
+                                        use_int4_w4a16 = False,
+                                        use_int4_w4a8 = True,
+                                        per_channel_quant=per_channel_quant,
+                                        global_num_experts=global_num_experts,
+                                        expert_map=expert_map,
+                                        w1_scale=w1_scale,
+                                        w2_scale=w2_scale,
+                                        w1_zp=w1_zp,
+                                        w2_zp=w2_zp,
+                                        a1_scale=a1_scale,
+                                        a2_scale=a2_scale,
+                                        block_shape=block_shape,
+                                        use_nn_moe= False,
+                                        shared_output=shared_output,
+                                        routed_scaling_factor=routed_scaling_factor
+                                       )
+    
+    #
+    if use_int4_w4a16:
+        assert hidden_states.size(1) // 2 == w1.size(2), (
+            "Hidden size mismatch")
+    elif use_nn_moe:
+        assert hidden_states.size(1) == w1.size(1), "Hidden size mismatch"
+    else:
+        assert hidden_states.size(1) == w1.size(2), (
+            f"Hidden size mismatch {hidden_states.size(1)} != {w1.size(2)}")
+
+    assert topk_weights.size() == topk_ids.size(), "topk shape mismatch"
+    assert hidden_states.is_contiguous(), "Hidden_states must be contiguous"
+    assert w1.stride(-1) == 1, "Stride of last dimension must be 1"
+    assert w2.stride(-1) == 1, "Stride of last dimension must be 1"
+    assert hidden_states.dtype in [
+        torch.float32, torch.float16, torch.bfloat16
+    ]
+
+    config_dtype = get_config_dtype_str(use_fp8_w8a8=use_fp8_w8a8,
+                                        use_int8_w8a8=use_int8_w8a8,
+                                        use_int8_w8a16=use_int8_w8a16,
+                                        use_int4_w4a16=use_int4_w4a16,
+                                        use_int4_w4a8=use_int4_w4a8,
+                                        dtype=hidden_states.dtype)
+
+    qtype = get_config_quant_dtype(use_fp8_w8a8=use_fp8_w8a8,
+                                   use_int8_w8a8=use_int8_w8a8,
+                                   use_int8_w8a16=use_int8_w8a16,
+                                   use_int4_w4a16=use_int4_w4a16,
+                                   use_int4_w4a8=use_int4_w4a8)
+
+    get_config_func = functools.partial(
+        try_get_optimal_moe_config,
+        w1.size(),
+        w2.size(),
+        top_k_num,
+        config_dtype,
+        block_shape=block_shape,
+        use_nn_moe=use_nn_moe,
+    )
+
+    config = get_config_func(M)
+
+    # We can reuse the memory between these because by the time we need
+    # cache3, we're done with cache1
+    intermediate_cache1 = cache13[:M * top_k_num * N].view(M, top_k_num, N)
+    intermediate_cache3 = cache13[:M * top_k_num * (K if not use_nn_moe else w2.shape[2])].view(M, top_k_num, K if not use_nn_moe else w2.shape[2])
+
+    # This needs separate memory since it's used concurrently with cache1
+    intermediate_cache2 = torch.empty((M * top_k_num, N // 2),
+                                      device=hidden_states.device,
+                                      dtype=hidden_states.dtype)
+
+    if hidden_states.dtype == torch.bfloat16:
+        compute_type = tl.bfloat16
+    elif hidden_states.dtype == torch.float16:
+        compute_type = tl.float16
+    elif hidden_states.dtype == torch.float32:
+        compute_type = tl.float32
+    else:
+        raise ValueError(f"Unsupported compute_type: {hidden_states.dtype}")
+
+    if inplace:
+        out_hidden_states = hidden_states
+    else:
+        out_hidden_states = torch.empty_like(hidden_states)
+
+    for chunk in range((num_tokens // CHUNK_SIZE) + 1):
+        begin_chunk_idx, end_chunk_idx = (chunk * CHUNK_SIZE,
+                                          min((chunk + 1) * CHUNK_SIZE,
+                                              num_tokens))
+        curr_hidden_states = hidden_states[begin_chunk_idx:end_chunk_idx]
+        tokens_in_chunk, _ = curr_hidden_states.size()
+
+        if tokens_in_chunk == 0:
+            break
+
+        if tokens_in_chunk < CHUNK_SIZE and chunk > 0:
+            # Adjust the intermediate cache size and config for the last
+            # chunk. Note that in most cases we only have one chunk
+            # so the cache size and config are already set correctly and
+            # do not need to be adjusted.
+            intermediate_cache1 = intermediate_cache1[:tokens_in_chunk]
+            intermediate_cache2 = intermediate_cache2[:tokens_in_chunk *
+                                                      topk_ids.size(1)]
+            intermediate_cache3 = intermediate_cache3[:tokens_in_chunk]
+            if not use_int8_w8a8:
+                config = get_config_func(tokens_in_chunk)
+
+        curr_topk_ids = topk_ids[begin_chunk_idx:end_chunk_idx]
+        curr_topk_weights = topk_weights[begin_chunk_idx:end_chunk_idx]
+
+        qcurr_hidden_states, a1q_scale = moe_kernel_quantize_input(
+            A=curr_hidden_states,
+            A_scale=a1_scale,
+            quant_dtype=qtype,
+            per_act_token_quant=per_channel_quant,
+            block_shape=block_shape)
+
+        if use_int4_w4a16:
+            sorted_token_ids, expert_ids, num_tokens_post_padded = (
+                moe_align_block_size(curr_topk_ids, config['BLOCK_SIZE_M'],
+                                    global_num_experts, expert_map, curr_hidden_states.shape[0]))
+        else:
+            sorted_token_ids, expert_ids, num_tokens_post_padded = (
+                moe_align_block_size(curr_topk_ids, config['BLOCK_SIZE_M'],
+                                    global_num_experts, expert_map))
+
+        if use_step3v_w8a16 : #step3v w8a16
+            w1 = (w1.reshape([w1_scale.shape[0],w1_scale.shape[1], -1, w1_scale.shape[-1]]) * w1_scale.unsqueeze(2)).reshape(w1.shape) #step3v w8反量化
+            invoke_fused_moe_kernel(qcurr_hidden_states,
+                                            w1,
+                                            intermediate_cache1,
+                                            None,
+                                            None,
+                                            w1_zp,
+                                            curr_topk_weights,
+                                            sorted_token_ids,
+                                            expert_ids,
+                                            num_tokens_post_padded,
+                                            apply_router_weight_on_input,
+                                            top_k_num,
+                                            config,
+                                            compute_type=compute_type,
+                                            use_fp8_w8a8=use_fp8_w8a8,
+                                            use_int8_w8a8=use_int8_w8a8,
+                                            use_int8_w8a16=use_int8_w8a16,
+                                            use_int4_w4a16=use_int4_w4a16,
+                                            use_int4_w4a8=use_int4_w4a8,
+                                            per_channel_quant=per_channel_quant,
+                                            block_shape=block_shape,
+                                            use_nn_moe=use_nn_moe)
+        else:        
+            invoke_fused_moe_kernel(qcurr_hidden_states,
+                                    w1,
+                                    intermediate_cache1,
+                                    a1q_scale,
+                                    w1_scale,
+                                    w1_zp,
+                                    curr_topk_weights,
+                                    sorted_token_ids,
+                                    expert_ids,
+                                    num_tokens_post_padded,
+                                    apply_router_weight_on_input,
+                                    top_k_num,
+                                    config,
+                                    compute_type=compute_type,
+                                    use_fp8_w8a8=use_fp8_w8a8,
+                                    use_int8_w8a8=use_int8_w8a8,
+                                    use_int8_w8a16=use_int8_w8a16,
+                                    use_int4_w4a16=use_int4_w4a16,
+                                    use_int4_w4a8=use_int4_w4a8,
+                                    per_channel_quant=per_channel_quant,
+                                    block_shape=block_shape,
+                                    use_nn_moe=use_nn_moe)
+
+
+        if activation == "silu":
+            torch.ops._C.silu_and_mul(intermediate_cache2,
+                                      intermediate_cache1.view(-1, N))
+        elif activation == "gelu":
+            torch.ops._C.gelu_and_mul(intermediate_cache2,
+                                      intermediate_cache1.view(-1, N))
+        else:
+            raise ValueError(f"Unsupported FusedMoe activation: {activation}")
+
+        qintermediate_cache2, a2q_scale = moe_kernel_quantize_input(
+            A=intermediate_cache2,
+            A_scale=a2_scale,
+            quant_dtype=qtype,
+            per_act_token_quant=per_channel_quant,
+            block_shape=block_shape)
+        
+        if use_step3v_w8a16 : #step3v w8a16
+            w2 = (w2.reshape([w2_scale.shape[0],w2_scale.shape[1], -1, w2_scale.shape[-1]]) * w2_scale.unsqueeze(2)).reshape(w2.shape) #step3v w8反量化
+            invoke_fused_moe_kernel(qintermediate_cache2,
+                                    w2,
+                                    intermediate_cache3,
+                                    None,
+                                    None,
+                                    w2_zp,
+                                    curr_topk_weights,
+                                    sorted_token_ids,
+                                    expert_ids,
+                                    num_tokens_post_padded,
+                                    not apply_router_weight_on_input,
+                                    1,
+                                    config,
+                                    compute_type=compute_type,
+                                    use_fp8_w8a8=use_fp8_w8a8,
+                                    use_int8_w8a8=use_int8_w8a8,
+                                    use_int8_w8a16=use_int8_w8a16,
+                                    use_int4_w4a16=use_int4_w4a16,
+                                    use_int4_w4a8=use_int4_w4a8,
+                                    per_channel_quant=per_channel_quant,
+                                    block_shape=block_shape,
+                                    use_nn_moe=use_nn_moe)
+        else:
+            invoke_fused_moe_kernel(qintermediate_cache2,
+                                    w2,
+                                    intermediate_cache3,
+                                    a2q_scale,
+                                    w2_scale,
+                                    w2_zp,
+                                    curr_topk_weights,
+                                    sorted_token_ids,
+                                    expert_ids,
+                                    num_tokens_post_padded,
+                                    not apply_router_weight_on_input,
+                                    1,
+                                    config,
+                                    compute_type=compute_type,
+                                    use_fp8_w8a8=use_fp8_w8a8,
+                                    use_int8_w8a8=use_int8_w8a8,
+                                    use_int8_w8a16=use_int8_w8a16,
+                                    use_int4_w4a16=use_int4_w4a16,
+                                    use_int4_w4a8=use_int4_w4a8,
+                                    per_channel_quant=per_channel_quant,
+                                    block_shape=block_shape,
+                                    use_nn_moe=use_nn_moe)
+
+
+        if envs.VLLM_USE_LIGHTOP: 
+            from lightop import op as op
+            op.moe_sum(input=intermediate_cache3.view(*intermediate_cache3.size()),
+                    output=out_hidden_states[begin_chunk_idx:end_chunk_idx], bias=shared_output[begin_chunk_idx:end_chunk_idx], 
+                    expert_mask=None, num_local_tokens=None, factor=routed_scaling_factor)
+        # else:
+        #     ops.moe_sum(intermediate_cache3.view(*intermediate_cache3.size()),
+        #                 out_hidden_states[begin_chunk_idx:end_chunk_idx]) 
+        #     if shared_output is not None:
+        #         if hidden_states.dtype != torch.float16 or dpsk_fp16_quick:
+        #             out_hidden_states[begin_chunk_idx:end_chunk_idx] = out_hidden_states[begin_chunk_idx:end_chunk_idx] * routed_scaling_factor + shared_output[begin_chunk_idx:end_chunk_idx]
+        #         else:
+        #             # Fix FP16 overflow
+        #             # See DeepseekV2DecoderLayer for more details.
+        #             out_hidden_states[begin_chunk_idx:end_chunk_idx] + shared_output[begin_chunk_idx:end_chunk_idx] * (1. / routed_scaling_factor)
+        #     else:
+        #         if hidden_states.dtype != torch.float16 or dpsk_fp16_quick:
+        #             ops.moe_sum(intermediate_cache3.view(*intermediate_cache3.size()),
+        #                     out_hidden_states[begin_chunk_idx:end_chunk_idx]) * routed_scaling_factor
+        else:
+            if envs.VLLM_USE_LIGHTOP_MOE_SUM:
+                from lightop import op as op
+                op.moe_sum(input=intermediate_cache3.view(*intermediate_cache3.size()),
+                    output=out_hidden_states[begin_chunk_idx:end_chunk_idx], bias=None, 
+                    expert_mask=None, num_local_tokens=None, factor=1.0)
+            elif envs.VLLM_USE_OPT_MOE_SUM:
+                moe_reduce_dispatch(intermediate_cache3.view(*intermediate_cache3.size()), out_hidden_states[begin_chunk_idx:end_chunk_idx], begin_chunk_idx, end_chunk_idx)
+            else:
+                ops.moe_sum(intermediate_cache3.view(*intermediate_cache3.size()),
+                                out_hidden_states[begin_chunk_idx:end_chunk_idx])
+
+
+    return out_hidden_states
+
+
+def fused_moe(
+    hidden_states: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    gating_output: torch.Tensor,
+    topk: int,
+    renormalize: bool,
+    inplace: bool = False,
+    activation: str = "silu",
+    use_grouped_topk: bool = False,
+    num_expert_group: Optional[int] = None,
+    topk_group: Optional[int] = None,
+    custom_routing_function: Optional[Callable] = None,
+    use_fp8_w8a8: bool = False,
+    use_int8_w8a8: bool = False,
+    use_int8_w8a16: bool = False,
+    use_int4_w4a16: bool = False,
+    use_int4_w4a8: bool =False,
+    per_channel_quant: bool = False,
+    global_num_experts: int = -1,
+    expert_map: Optional[torch.Tensor] = None,
+    w1_scale: Optional[torch.Tensor] = None,
+    w2_scale: Optional[torch.Tensor] = None,
+    w1_zp: Optional[torch.Tensor] = None,
+    w2_zp: Optional[torch.Tensor] = None,
+    a1_scale: Optional[torch.Tensor] = None,
+    a2_scale: Optional[torch.Tensor] = None,
+    block_shape: Optional[List[int]] = None,
+    use_nn_moe: Optional[bool] = False,
+    shared_output: Optional[torch.Tensor] = None,
+    routed_scaling_factor: Optional[float] = None,
+    use_step3v_w8a16: bool = False,    
+) -> torch.Tensor:
+    """
+    This function computes a Mixture of Experts (MoE) layer using two sets of
+    weights, w1 and w2, and top-k gating mechanism.
+
+    Parameters:
+    - hidden_states (torch.Tensor): The input tensor to the MoE layer.
+    - w1 (torch.Tensor): The first set of expert weights.
+    - w2 (torch.Tensor): The second set of expert weights.
+    - gating_output (torch.Tensor): The output of the gating operation
+        (before softmax).
+    - topk (int): The number of top-k experts to select.
+    - renormalize (bool): If True, renormalize the top-k weights to sum to 1.
+    - inplace (bool): If True, perform the operation in-place.
+        Defaults to False.
+    - activation (str): The activation function to apply after the first
+        MoE layer.
+    - num_expert_group: Optional[int]: additional parameter for grouped_topk
+    - topk_group: Optional[int]: additional parameter for grouped_topk
+    - use_grouped_topk: If True, use grouped_topk instead of fused_topk
+        note: Deepseekv2 model uses grouped_topk
+    - use_fp8_w8a8 (bool): If True, use fp8 arithmetic to compute the inner
+        products for w1 and w2. Defaults to False.
+    - use_int8_w8a8 (bool): If True, use int8 arithmetic to compute the inner
+        products for w1 and w2. Defaults to False.
+    - use_int8_w8a16 (bool): If True, use matmul of int8 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - use_int4_w4a16 (bool): If True, use matmul of int4 weight and bf16/fp16
+        activation to compute the inner products for w1 and w2.
+        Defaults to False.
+    - global_num_experts (int): The total number of experts in the global
+        expert space.
+    - expert_map (Optional[torch.Tensor]):  A tensor mapping expert indices 
+        from the global expert space to the local expert space of the expert 
+        parallel shard.
+    - w1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        w1.
+    - w2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        w2.
+    - a1_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a1.
+    - a2_scale (Optional[torch.Tensor]): Optional scale to be used for
+        a2.
+    - block_shape: (Optional[List[int]]): Optional block size for block-wise
+        quantization.
+
+    Returns:
+    - torch.Tensor: The output tensor after applying the MoE layer.
+    """
+
+    if use_grouped_topk:
+        assert num_expert_group is not None and topk_group is not None
+        topk_weights, topk_ids = grouped_topk(hidden_states, gating_output,
+                                              topk, renormalize,
+                                              num_expert_group, topk_group)
+    elif custom_routing_function is None:
+        topk_weights, topk_ids, token_expert_indices = fused_topk(
+            hidden_states, gating_output, topk, renormalize)
+    else:
+        topk_weights, topk_ids = custom_routing_function(
+            hidden_states, gating_output, topk, renormalize)
+
+    return fused_experts(hidden_states,
+                         w1,
+                         w2,
+                         topk_weights,
+                         topk_ids,
+                         inplace=inplace,
+                         activation=activation,
+                         use_fp8_w8a8=use_fp8_w8a8,
+                         use_int8_w8a8=use_int8_w8a8,
+                         use_int8_w8a16=use_int8_w8a16,
+                         use_int4_w4a16=use_int4_w4a16,
+                         use_int4_w4a8=use_int4_w4a8,
+                         per_channel_quant=per_channel_quant,
+                         global_num_experts=global_num_experts,
+                         expert_map=expert_map,
+                         w1_scale=w1_scale,
+                         w2_scale=w2_scale,
+                         w1_zp=w1_zp,
+                         w2_zp=w2_zp,
+                         a1_scale=a1_scale,
+                         a2_scale=a2_scale,
+                         block_shape=block_shape,
+                         use_nn_moe=use_nn_moe,
+                         shared_output=shared_output,
+                         routed_scaling_factor=routed_scaling_factor,
+                         use_step3v_w8a16=use_step3v_w8a16)
+
+
+class TritonExperts(mk.FusedMoEPermuteExpertsUnpermute):
+
+    def __init__(
+        self,
+        use_fp8_w8a8: bool = False,
+        use_int8_w8a8: bool = False,
+        use_int8_w8a16: bool = False,
+        use_int4_w4a16: bool = False,
+        use_int4_w4a8: bool = False,
+        per_act_token_quant: bool = False,
+        block_shape: Optional[List[int]] = None,
+    ):
+        super().__init__(
+            FusedMoEQuantConfig.make(
+                use_fp8_w8a8=use_fp8_w8a8,
+                use_int8_w8a8=use_int8_w8a8,
+                use_int8_w8a16=use_int8_w8a16,
+                use_int4_w4a16=use_int4_w4a16,
+                use_int4_w4a8=use_int4_w4a8,
+                per_act_token_quant=per_act_token_quant,
+                block_shape=block_shape,
+            ))
+
+        self.use_fp8_w8a8 = use_fp8_w8a8
+        self.use_int4_w4a16 = use_int4_w4a16
+        self.use_int8_w8a8 = use_int8_w8a8
+        self.use_int8_w8a16 = use_int8_w8a16
+        self.use_int4_w4a8 = use_int4_w4a8
+
+    @property
+    def activation_formats(
+        self
+    ) -> tuple[mk.FusedMoEActivationFormat, mk.FusedMoEActivationFormat]:
+        return (mk.FusedMoEActivationFormat.Standard,
+                mk.FusedMoEActivationFormat.Standard)
+
+    def supports_chunking(self) -> bool:
+        return True
+
+    def supports_expert_map(self) -> bool:
+        return True
+
+    def workspace_shapes(
+        self,
+        a: torch.Tensor,
+        aq: torch.Tensor,
+        M: int,
+        N: int,
+        K: int,
+        topk: int,
+        global_num_experts: int,
+        local_num_experts: int,
+    ) -> tuple[tuple[int, ...], tuple[int, ...], tuple[int, ...], torch.dtype]:
+        workspace1 = (M, topk, max(N * 2, K))
+        workspace2 = (M, topk, N)
+        output = (M, topk, K)
+        return (workspace1, workspace2, output, a.dtype)
+
+    def apply(
+        self,
+        output: torch.Tensor,
+        hidden_states: torch.Tensor,
+        w1: torch.Tensor,
+        w2: torch.Tensor,
+        topk_ids: torch.Tensor,
+        activation: str,
+        global_num_experts: int,
+        expert_map: Optional[torch.Tensor],
+        w1_scale: Optional[torch.Tensor],
+        w2_scale: Optional[torch.Tensor],
+        w1_zp: Optional[torch.Tensor],
+        w2_zp: Optional[torch.Tensor],
+        a1q_scale: Optional[torch.Tensor],
+        a2_scale: Optional[torch.Tensor],
+        workspace13: torch.Tensor,
+        workspace2: torch.Tensor,
+        expert_num_tokens: Optional[torch.Tensor],
+    ):
+        # Check constraints.
+        if self.use_int4_w4a16:
+            assert hidden_states.size(-1) // 2 == w1.size(2), (
+                "Hidden size mismatch")
+        else:
+            assert hidden_states.size(-1) == w1.size(2), \
+                (f"Hidden size mismatch {hidden_states.size(-1)} "
+                 f"!= {w1.size(2)}")
+
+        assert hidden_states.is_contiguous(
+        ), "Hidden_states must be contiguous"
+        assert hidden_states.dim() == 2
+        assert w1.stride(-1) == 1, "Stride of last dimension must be 1"
+        assert w2.stride(-1) == 1, "Stride of last dimension must be 1"
+        assert hidden_states.dtype in [
+            torch.float32, torch.float16, torch.bfloat16, torch.float8_e4m3fn
+        ]
+
+        E, num_tokens, N, K, top_k_num = mk._moe_problem_size(
+            hidden_states, w1, w2, topk_ids)
+
+        if global_num_experts == -1:
+            global_num_experts = E
+
+        config_dtype = get_config_dtype_str(use_fp8_w8a8=self.use_fp8_w8a8,
+                                            use_int8_w8a16=self.use_int8_w8a16,
+                                            use_int4_w4a16=self.use_int4_w4a16,
+                                            use_int4_w4a8=self.use_int4_w4a8,
+                                            dtype=hidden_states.dtype)
+
+        config = try_get_optimal_moe_config(
+            w1.size(),
+            w2.size(),
+            top_k_num,
+            config_dtype,
+            num_tokens,
+            block_shape=self.block_shape,
+        )
+
+        if hidden_states.dtype == torch.bfloat16:
+            compute_type = tl.bfloat16
+        elif hidden_states.dtype == torch.float16:
+            compute_type = tl.float16
+        elif hidden_states.dtype == torch.float32:
+            compute_type = tl.float32
+        elif hidden_states.dtype == torch.float8_e4m3fn:
+            compute_type = tl.bfloat16
+        else:
+            raise ValueError(
+                f"Unsupported compute_type: {hidden_states.dtype}")
+
+        # We can reuse the memory between these because by the time we need
+        # cache3, we're done with cache1
+        intermediate_cache1 = _resize_cache(workspace13,
+                                            (num_tokens, top_k_num, N))
+        intermediate_cache2 = _resize_cache(workspace2,
+                                            (num_tokens * top_k_num, N // 2))
+
+        sorted_token_ids, expert_ids, num_tokens_post_padded = (
+            moe_align_block_size(topk_ids, config['BLOCK_SIZE_M'],
+                                 global_num_experts, expert_map))
+
+        invoke_fused_moe_kernel(hidden_states,
+                                w1,
+                                intermediate_cache1,
+                                a1q_scale,
+                                w1_scale,
+                                w1_zp,
+                                None,
+                                sorted_token_ids,
+                                expert_ids,
+                                num_tokens_post_padded,
+                                False,
+                                top_k_num,
+                                config,
+                                compute_type=compute_type,
+                                use_fp8_w8a8=self.use_fp8_w8a8,
+                                use_int8_w8a8=self.use_int8_w8a8,
+                                use_int8_w8a16=self.use_int8_w8a16,
+                                use_int4_w4a16=self.use_int4_w4a16,
+                                use_int4_w4a8=self.use_int4_w4a8,
+                                per_channel_quant=self.per_act_token_quant,
+                                block_shape=self.block_shape)
+
+        self.activation(activation, intermediate_cache2,
+                        intermediate_cache1.view(-1, N))
+
+        a2q_scale: Optional[torch.Tensor] = None
+
+        qintermediate_cache2, a2q_scale = moe_kernel_quantize_input(
+            intermediate_cache2, a2_scale, self.quant_dtype,
+            self.per_act_token_quant, self.block_shape)
+
+        invoke_fused_moe_kernel(qintermediate_cache2,
+                                w2,
+                                output,
+                                a2q_scale,
+                                w2_scale,
+                                w2_zp,
+                                None,
+                                sorted_token_ids,
+                                expert_ids,
+                                num_tokens_post_padded,
+                                False,
+                                1,
+                                config,
+                                compute_type=compute_type,
+                                use_fp8_w8a8=self.use_fp8_w8a8,
+                                use_int8_w8a8=self.use_int8_w8a8,
+                                use_int8_w8a16=self.use_int8_w8a16,
+                                use_int4_w4a16=self.use_int4_w4a16,
+                                use_int4_w4a8= self.use_int4_w4a8,
+                                per_channel_quant=self.per_act_token_quant,
+                                block_shape=self.block_shape)
+
+
+def modular_triton_fused_moe(
+    use_fp8_w8a8: bool,
+    use_int8_w8a8: bool,
+    use_int8_w8a16: bool,
+    use_int4_w4a16: bool,
+    use_int4_w4a8: bool,
+    per_act_token_quant: bool,
+    block_shape: Optional[List[int]] = None,
+) -> mk.FusedMoEModularKernel:
+    return mk.FusedMoEModularKernel(
+        MoEPrepareAndFinalizeNoEP(),
+        TritonExperts(
+            use_fp8_w8a8=use_fp8_w8a8,
+            use_int8_w8a8=use_int8_w8a8,
+            use_int8_w8a16=use_int8_w8a16,
+            use_int4_w4a16=use_int4_w4a16,
+            use_int4_w4a8=use_int4_w4a8,
+            per_act_token_quant=per_act_token_quant,
+            block_shape=block_shape,
+        ),
+    )

vllm/model_executor/layers/quantization/__init__.py

@@ -33,6 +33,7 @@ QuantizationMethods = Literal[
     "ipex",
     "quark",
     "moe_wna16",
+    "groupwise-quant",
     "torchao",
     "auto-round",
     "rtn",
@@ -120,6 +121,7 @@ def get_quantization_config(quantization: str) -> type[QuantizationConfig]:
     from .blockwise_int8 import BlockInt8Config
     from .slimquant_w4a8 import SlimQuantW4A8Int8Config
     from .slimquant_w4a8_marlin import SlimQuantW4A8Int8MarlinConfig
+    from .groupwise_quant import GroupwiseQuantConfig
 
     method_to_config: dict[str, type[QuantizationConfig]] = {
         "aqlm": AQLMConfig,
@@ -152,6 +154,7 @@ def get_quantization_config(quantization: str) -> type[QuantizationConfig]:
         "auto-round": AutoRoundConfig,
         "rtn": RTNConfig,
         "blockwise_int8": BlockInt8Config,
+        "groupwise-quant": GroupwiseQuantConfig,
         "slimquant_w4a8":SlimQuantW4A8Int8Config,
         "slimquant_w4a8_marlin":SlimQuantW4A8Int8MarlinConfig,
     }

vllm/model_executor/layers/quantization/groupwise_quant.py

@@ -3,18 +3,16 @@ from typing import Any, Callable, Dict, List, Optional
 
 import torch
 from torch.nn.parameter import Parameter
-
+from vllm.distributed import get_tensor_model_parallel_rank, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.fused_moe import (FusedMoE, FusedMoEMethodBase,
                                                   FusedMoeWeightScaleSupported)
-from vllm.model_executor.layers.fused_moe.optimus_moe import (  # noqa: F401
-    optimus_moe_int8)
 from vllm.model_executor.layers.linear import (LinearBase, LinearMethodBase,
                                                UnquantizedLinearMethod)
 from vllm.model_executor.layers.quantization.base_config import (
     QuantizationConfig, QuantizeMethodBase)
+#from vllm.model_executor.layers.fused_moe.optimus_moe import (  # noqa: F401
+#    optimus_moe_int8)
 from vllm.model_executor.utils import set_weight_attrs
-from vllm.utils import direct_register_custom_op
-
 
 class GroupwiseQuantConfig(QuantizationConfig):
     """Config class for Groupwise Quantization.
@@ -159,92 +157,7 @@ class GroupwiseQuantLinearMethod(LinearMethodBase):
         has_num_experts = any("num_experts" in name for name in layer_keys)
         if not has_num_experts:
             layer.register_parameter("num_experts", None)
-
-        if self.quant_config.weight_bits == 4:
-            assert input_size_per_partition % self.quant_config.group_size == 0
-            assert output_size_per_partition % self.quant_config.pack_factor == 0
-            if num_experts:
-                weight_shape = [
-                    num_experts, input_size_per_partition,
-                    output_size_per_partition // self.quant_config.pack_factor
-                ]
-                scale_shape = [
-                    num_experts,
-                    input_size_per_partition // self.quant_config.group_size,
-                    output_size_per_partition,
-                ]
-            else:
-                weight_shape = [
-                    input_size_per_partition,
-                    output_size_per_partition // self.quant_config.pack_factor
-                ]
-                scale_shape = [
-                    input_size_per_partition // self.quant_config.group_size,
-                    output_size_per_partition,
-                ]
-            qweight = Parameter(
-                torch.empty(
-                    *weight_shape,
-                    dtype=torch.int32,
-                ),
-                requires_grad=False,
-            )
-            scales = Parameter(
-                torch.empty(
-                    *scale_shape,
-                    dtype=params_dtype,
-                ),
-                requires_grad=False,
-            )
-            zeros = Parameter(
-                torch.empty(
-                    *scale_shape,
-                    dtype=params_dtype,
-                ),
-                requires_grad=False,
-            )
-
-            if num_experts:
-                set_weight_attrs(
-                    qweight, {
-                        "input_dim": 1,
-                        "output_dim": 2,
-                        "packed_dim": 2,
-                        "pack_factor": self.quant_config.pack_factor,
-                    })
-                set_weight_attrs(scales, {
-                    "input_dim": 1,
-                    "output_dim": 2,
-                })
-                set_weight_attrs(zeros, {
-                    "input_dim": 1,
-                    "output_dim": 2,
-                })
-            else:
-                set_weight_attrs(
-                    qweight, {
-                        "input_dim": 0,
-                        "output_dim": 1,
-                        "packed_dim": 1,
-                        "pack_factor": self.quant_config.pack_factor,
-                    })
-                set_weight_attrs(scales, {
-                    "input_dim": 0,
-                    "output_dim": 1,
-                })
-                set_weight_attrs(zeros, {
-                    "input_dim": 0,
-                    "output_dim": 1,
-                })
-
-            layer.register_parameter("qweight", qweight)
-            set_weight_attrs(qweight, extra_weight_attrs)
-            layer.register_parameter("scales", scales)
-            set_weight_attrs(scales, extra_weight_attrs)
-            layer.register_parameter("zeros", zeros)
-            set_weight_attrs(zeros, extra_weight_attrs)
-
-        elif self.quant_config.weight_bits == 8:
+        if self.quant_config.weight_bits == 8:
             assert input_size_per_partition % self.quant_config.group_size == 0
             if num_experts:
                 weight_shape = [
@@ -300,68 +213,6 @@ class GroupwiseQuantLinearMethod(LinearMethodBase):
                     "output_dim": 1,
                 })
 
-            layer.register_parameter("qweight", qweight)
-            set_weight_attrs(qweight, extra_weight_attrs)
-            layer.register_parameter("scales", scales)
-            set_weight_attrs(scales, extra_weight_attrs)
-        elif self.quant_config.weight_bits == 6:
-            assert input_size_per_partition % self.quant_config.group_size == 0
-            if num_experts:
-                weight_shape = [
-                    num_experts,
-                    output_size_per_partition,
-                    input_size_per_partition,
-                ]
-                scale_shape = [
-                    num_experts,
-                    input_size_per_partition // self.quant_config.group_size,
-                    output_size_per_partition,
-                ]
-            else:
-                weight_shape = [
-                    output_size_per_partition, input_size_per_partition
-                ]
-                scale_shape = [
-                    input_size_per_partition // self.quant_config.group_size,
-                    output_size_per_partition,
-                ]
-            qweight = Parameter(
-                torch.zeros(
-                    *weight_shape,
-                    device=
-                    "cpu",  # hack for fp6 weight is stored in float16, to avoid cuda oom
-                    dtype=torch.float16,
-                ),
-                requires_grad=False,
-            )
-
-            scales = Parameter(
-                torch.empty(
-                    *scale_shape,
-                    device="cuda",
-                    dtype=torch.float16,
-                ),
-                requires_grad=False,
-            )
-            if num_experts:
-                set_weight_attrs(qweight, {
-                    "input_dim": 2,
-                    "output_dim": 1,
-                })
-                set_weight_attrs(scales, {
-                    "input_dim": 1,
-                    "output_dim": 2,
-                })
-            else:
-                set_weight_attrs(qweight, {
-                    "input_dim": 1,
-                    "output_dim": 0,
-                })
-                set_weight_attrs(scales, {
-                    "input_dim": 0,
-                    "output_dim": 1,
-                })
-
             layer.register_parameter("qweight", qweight)
             set_weight_attrs(qweight, extra_weight_attrs)
             layer.register_parameter("scales", scales)
@@ -372,78 +223,18 @@ class GroupwiseQuantLinearMethod(LinearMethodBase):
     def process_weights_after_loading(self, layer: torch.nn.Module) -> None:
         if not hasattr(layer, "qweight"):
             return
-
-        if self.quant_config.weight_bits == 4:
-            num_experts = layer.num_experts
-            qweight = layer.qweight
-            zeros = layer.zeros
-            scales = layer.scales
-            if num_experts:
-                qscales_list = []
-                for i in range(num_experts):
-                    qweight_processed, qscales = torch.ops.Optimus.GemmInt4GroupQuantWeight(
-                        qweight[i], zeros[i], scales[i] + zeros[i],
-                        self.quant_config.group_size)
-                    qweight[i].copy_(qweight_processed)
-                    qscales_list.append(qscales)
-                qscales = Parameter(torch.stack(qscales_list),
-                                    requires_grad=False)
-                layer.register_parameter("qscales", qscales)
-            else:
-                qweight_processed, qscales = torch.ops.Optimus.GemmInt4GroupQuantWeight(
-                    qweight, zeros, scales + zeros,
-                    self.quant_config.group_size)
-                qweight.copy_(qweight_processed)
-                qscales = Parameter(qscales, requires_grad=False)
-                layer.register_parameter("qscales", qscales)
-            layer._parameters.pop("zeros")
-            layer._parameters.pop("scales")
-        elif self.quant_config.weight_bits == 8:
+        if self.quant_config.weight_bits == 8:
             num_experts = layer.num_experts
             qweight = layer.qweight
-
             if num_experts:
                 for i in range(num_experts):
-                    qweight[i].copy_(
-                        torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
-                            qweight[i].t().contiguous(), torch.int8))
+                    qweight[i].copy_(qweight[i].contiguous())
             else:
-                qweight.copy_(
-                    torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
-                        qweight.t().contiguous(), torch.int8))
-        elif self.quant_config.weight_bits == 6:
-            num_experts = layer.num_experts
-            qweight = layer.qweight
-            layer._parameters.pop("qweight")
-            assert qweight.shape[-1] % 8 == 0
-            if num_experts:
-                qweight_processed = torch.empty(qweight.shape[0],
-                                                qweight.shape[1],
-                                                qweight.shape[2] * 6 // 8,
-                                                dtype=torch.uint8,
-                                                device="cuda")
-            else:
-                qweight_processed = torch.empty(qweight.shape[0],
-                                                qweight.shape[1] * 6 // 8,
-                                                dtype=torch.uint8,
-                                                device="cuda")
-            if num_experts:
-                for i in range(num_experts):
-                    qweight_processed[
-                        i] = torch.ops.Optimus.fp6_preprocess_weight(
-                            qweight[i].cpu()).cuda()
-                qweight_processed = Parameter(qweight_processed,
-                                              requires_grad=False)
-                layer.register_parameter("qweight", qweight_processed)
-            else:
-                qweight_processed = Parameter(
-                    torch.ops.Optimus.fp6_preprocess_weight(
-                        qweight.cpu()).cuda(),
-                    requires_grad=False)
-                layer.register_parameter("qweight", qweight_processed)
+                qweight.copy_(qweight.contiguous())         
         else:
             raise NotImplementedError
         
+
     def apply(self,
             layer: torch.nn.Module,
             x: torch.Tensor,
@@ -451,78 +242,59 @@ class GroupwiseQuantLinearMethod(LinearMethodBase):
             residual: Optional[torch.Tensor] = None,
             output: Optional[torch.Tensor] = None,
             expert_idx: Optional[int] = None) -> torch.Tensor:
-        if self.quant_config.weight_bits == 4:
-            qweight = layer.qweight
-            qscales = layer.qscales
-            num_experts = layer.num_experts
-            if num_experts:
-                assert expert_idx is not None, "expert_idx is None"
-                qweight = qweight[expert_idx]
-                qscales = qscales[expert_idx]
-            out = torch.ops.vllm.optimus_gemm_int4_group(x,
-                                          qweight,
-                                          qscales,
-                                          bias,
-                                          None, # Placeholder for a fifth argument that is None
-                                          out=output)
-            if residual is not None:
-                out += residual
-            return out
-        elif self.quant_config.weight_bits == 8:
+        if self.quant_config.weight_bits == 8:
             qweight = layer.qweight
             scales = layer.scales
             num_experts = layer.num_experts
+
             if num_experts:
                 assert expert_idx is not None, "expert_idx is None"
                 qweight = qweight[expert_idx]
                 scales = scales[expert_idx]
 
+            group_size = self.quant_config.group_size
+            input_size_per_partition = qweight.size(0)  
+            output_size_per_partition = qweight.size(1)
+
+            qweight = qweight.to(torch.float32) 
+            scales = scales.to(torch.float32)
+
+            scales_expanded = scales.repeat_interleave(group_size, dim=0)
+            scales_expanded = scales_expanded[:input_size_per_partition]
+
+            weight = qweight * scales_expanded
+
+            weight = weight.to(x.dtype)
             if residual is not None:
                 assert output is None or output is residual
-                out = torch.ops.vllm.optimus_fp_aintb_gemm(x,
-                                            qweight,
-                                            torch.int8, # Placeholder for dtype argument
-                                            scales,
-                                            residual,
-                                            out=residual)
-                if bias is not None:
-                    out += bias
+                if get_tensor_model_parallel_world_size() > 1 and get_tensor_model_parallel_rank() != 0:
+                    beta = 0.0
                 else:
-                out = torch.ops.vllm.optimus_fp_aintb_gemm(x,
-                                            qweight,
-                                            torch.int8, # Placeholder for dtype argument
-                                            scales,
-                                            bias,
-                                            out=output)
-
-            return out
-        elif self.quant_config.weight_bits == 6:
-            qweight = layer.qweight
-            scales = layer.scales
-            num_experts = layer.num_experts
-            if num_experts:
-                assert expert_idx is not None, "expert_idx is None"
-                qweight = qweight[expert_idx]
-                scales = scales[expert_idx]
-
-            if x.dtype != torch.bfloat16:
-                if output is None:
-                    output = torch.empty(x.shape[0],
-                                         qweight.shape[0],
-                                         device=x.device,
-                                         dtype=torch.bfloat16)
+                    beta = 1.0
+                if x.dim() == 2:
+                    torch.addmm(residual, x, weight.t(), beta=beta, out=residual)
+                elif x.dim() >= 3:
+                    hx = x.size(-1)
+                    hr = residual.size(-1)
+                    torch.addmm(residual.view(-1, hr),
+                                x.view(-1, hx),
+                                weight.t(),
+                                beta=beta,
+                                out=residual.view(-1, hr))
                 else:
-                    output = output.to(torch.bfloat16)
-            out = torch.ops.vllm.optimus_fp6_linear(x,
-                                       qweight,
-                                       scales,
-                                       4, # Placeholder for fp6_format_code
-                                       out=output)
+                    raise AssertionError(f"unrecognized tensor dimensions: {x.dim()}")
                 if bias is not None:
-                out += bias
-            if residual is not None:
-                out += residual
-            return out
+                    residual += bias
+                return residual
+            else:
+                if output is not None:
+                    if bias is not None:  # always separate bias add when output is provided
+                        torch.matmul(x, weight.t(), out=output)
+                        output.add_(bias)
+                        return output
+                    return torch.matmul(x, weight.t(), out=output)
+                else:
+                    return torch.nn.functional.linear(x, weight.t(), bias)
         else:
             raise NotImplementedError
         
@@ -603,18 +375,19 @@ class GroupwiseInt8MoeMethod(FusedMoEMethodBase):
 
     def process_weights_after_loading(self, layer: torch.nn.Module) -> None:
         # Process weights similar to GroupwiseQuantLinearMethod for 8-bit case
-        num_experts = layer.w13_weight.shape[0]
+        # num_experts = layer.w13_weight.shape[0]
         
-        for expert in range(num_experts):
-            # Preprocess w13 weight (gate and up combined)
-            layer.w13_weight[expert].copy_(
-                torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
-                    layer.w13_weight[expert].t().contiguous(), torch.int8))
+        # for expert in range(num_experts):
+        #     # Preprocess w13 weight (gate and up combined)
+        #     layer.w13_weight[expert].copy_(
+        #         torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
+        #             layer.w13_weight[expert].t().contiguous(), torch.int8))
             
-            # Preprocess w2 weight (down)
-            layer.w2_weight[expert].copy_(
-                torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
-                    layer.w2_weight[expert].t().contiguous(), torch.int8))
+        #     # Preprocess w2 weight (down)
+        #     layer.w2_weight[expert].copy_(
+        #         torch.ops.Optimus.FpAIntBPreprocessWeightGPU(
+        #             layer.w2_weight[expert].t().contiguous(), torch.int8))
+        pass
         
     def apply(
         self,
@@ -632,95 +405,48 @@ class GroupwiseInt8MoeMethod(FusedMoEMethodBase):
         scoring_func: str = "softmax",
         e_score_correction_bias: Optional[torch.Tensor] = None,
         apply_router_weight_on_input: bool = False,
+        enable_eplb: bool = False,
+        expert_load_view: Optional[torch.Tensor] = None,
         activation: str = "silu",
+        use_nn_moe: Optional[bool] = False,
+        routed_scaling_factor: Optional[float] = None,
+        use_fused_gate: Optional[bool] = False,
+        logical_to_physical_map: Optional[torch.Tensor] = None,
+        logical_replica_count: Optional[torch.Tensor] = None,
+        **_            
     ) -> torch.Tensor:
-        return torch.ops.vllm.optimus_moe_int8(
+        if enable_eplb:
+            raise NotImplementedError(
+                "EPLB not supported for `ExpertsInt8MoEMethod` yet.")
+
+        from vllm.model_executor.layers.fused_moe.fused_moe_step3vw8a16 import fused_experts
+
+        topk_weights, topk_ids = FusedMoE.select_experts(
             hidden_states=x,
             router_logits=router_logits,
-            w1=layer.w13_weight,
-            w2=layer.w2_weight,
-            w1_scale=layer.w13_weight_scale,
-            w2_scale=layer.w2_weight_scale,
+            use_grouped_topk=use_grouped_topk,
             top_k=top_k,
-            global_num_experts=global_num_experts,
-            norm_expert_weight=renormalize,
+            renormalize=renormalize,
+            topk_group=topk_group,
+            num_expert_group=num_expert_group,
+            custom_routing_function=custom_routing_function,
+            scoring_func=scoring_func,
+            e_score_correction_bias=e_score_correction_bias)
+#复用bf16 moe逻辑
+        return fused_experts(
+            x,
+            layer.w13_weight,
+            layer.w2_weight,
+            topk_weights=topk_weights,
+            topk_ids=topk_ids,
+            inplace=True,
             activation=activation,
-        )
+            use_nn_moe = True,
+            global_num_experts=global_num_experts,
+            apply_router_weight_on_input=apply_router_weight_on_input,
+            expert_map=expert_map,
+            w1_scale=layer.w13_weight_scale,
+            w2_scale=layer.w2_weight_scale,
+            use_step3v_w8a16 = True)
 
 
-# Wrapper and Fake Functions for Optimus::GemmInt4Group
-def optimus_gemm_int4_group(x: torch.Tensor, qweight: torch.Tensor,
-                            qscales: torch.Tensor,
-                            bias: Optional[torch.Tensor],
-                            out: Optional[torch.Tensor]) -> torch.Tensor:
-    return torch.ops.Optimus.GemmInt4Group(x, qweight, qscales, bias,
-                                           None, out=out)
-
-def optimus_gemm_int4_group_fake(x: torch.Tensor, qweight: torch.Tensor,
-                                 qscales: torch.Tensor,
-                                 bias: Optional[torch.Tensor],
-                                 out: Optional[torch.Tensor]) -> torch.Tensor:
-    output_shape = list(x.shape[:-1]) + [qscales.shape[-1]]
-    if out is not None:
-        return torch.empty(output_shape, dtype=x.dtype, device=x.device)
-    return torch.empty(output_shape, dtype=x.dtype, device=x.device)
-
-# Wrapper and Fake Functions for Optimus::FpAIntBGemm
-def optimus_fp_aintb_gemm(x: torch.Tensor, qweight: torch.Tensor,
-                          dtype_arg: torch.dtype, scales: torch.Tensor,
-                          bias_or_residual: Optional[torch.Tensor],
-                          out: Optional[torch.Tensor]) -> torch.Tensor:
-    return torch.ops.Optimus.FpAIntBGemm(x, qweight, dtype_arg, scales,
-                                         bias_or_residual, "identity",
-                                         out=out)
-
-def optimus_fp_aintb_gemm_fake(x: torch.Tensor, qweight: torch.Tensor,
-                               dtype_arg: torch.dtype, scales: torch.Tensor,
-                               bias_or_residual: Optional[torch.Tensor],
-                               out: Optional[torch.Tensor]) -> torch.Tensor:
-    output_shape = list(x.shape[:-1]) + [qweight.shape[-1]]
-    if out is not None:
-        return torch.empty(output_shape, dtype=x.dtype, device=x.device)
-    return torch.empty(output_shape, dtype=x.dtype, device=x.device)
-
-# Wrapper and Fake Functions for Optimus::fp6_linear
-def optimus_fp6_linear(x: torch.Tensor, qweight: torch.Tensor,
-                       scales: torch.Tensor, fp6_format_code: int,
-                       out: Optional[torch.Tensor]) -> torch.Tensor:
-    return torch.ops.Optimus.fp6_linear(x, qweight, scales, fp6_format_code,
-                                        out=out)
-
-def optimus_fp6_linear_fake(x: torch.Tensor, qweight: torch.Tensor,
-                            scales: torch.Tensor, fp6_format_code: int,
-                            out: Optional[torch.Tensor]) -> torch.Tensor:
-    output_channels = scales.shape[-1]
-    output_shape = list(x.shape[:-1]) + [output_channels]
-    output_dtype = x.dtype
-    if x.dtype != torch.bfloat16:
-        output_dtype = torch.bfloat16
-
-    if out is not None:
-        return torch.empty(output_shape, dtype=output_dtype, device=x.device)
-    return torch.empty(output_shape, dtype=output_dtype, device=x.device)
-
-
-direct_register_custom_op(
-    op_name="optimus_gemm_int4_group",
-    op_func=optimus_gemm_int4_group,
-    mutates_args=["out"],
-    fake_impl=optimus_gemm_int4_group_fake,
-)
-
-direct_register_custom_op(
-    op_name="optimus_fp_aintb_gemm",
-    op_func=optimus_fp_aintb_gemm,
-    mutates_args=["out", "bias_or_residual"],
-    fake_impl=optimus_fp_aintb_gemm_fake,
-)
-
-direct_register_custom_op(
-    op_name="optimus_fp6_linear",
-    op_func=optimus_fp6_linear,
-    mutates_args=["out"],
-    fake_impl=optimus_fp6_linear_fake,
-)
\ No newline at end of file

vllm/platforms/rocm.py

@@ -16,16 +16,13 @@ from vllm.utils import cuda_device_count_stateless
 
 from .interface import DeviceCapability, Platform, PlatformEnum, _Backend
 
-from vllm.utils import is_kme, SUPPORT_TC
+from vllm.utils import SUPPORT_TC
 
 if not SUPPORT_TC:
     os.environ['VLLM_USE_V1'] = '0'
     os.environ['VLLM_USE_FLASH_ATTN_PA'] = '0'
     os.environ['VLLM_USE_FLASH_MLA'] = '0'
     
-if is_kme:
-    os.environ['VLLM_USE_FLASH_ATTN_PA'] = '0'
-    
 
 if TYPE_CHECKING:
     from vllm.config import ModelConfig, VllmConfig
@@ -190,7 +187,7 @@ class RocmPlatform(Platform):
     device_control_env_var: str = "CUDA_VISIBLE_DEVICES"
 
     supported_quantization: list[str] = [
-        "awq", "gptq", "fp8", "compressed-tensors", "fbgemm_fp8", "gguf",
+        "awq", "gptq", "fp8", "compressed-tensors", "fbgemm_fp8", "gguf","groupwise-quant",
         "quark", "ptpc_fp8", "moe_wna16", "blockwise_int8","slimquant_w4a8","awq_marlin","slimquant_w4a8_marlin"
     ]
 
@@ -304,8 +301,6 @@ class RocmPlatform(Platform):
                 logger.info("flash_attn is not supported on NAVI GPUs.")
         else:
             logger.info("%s is not supported in AMD GPUs.", selected_backend)
-        if is_kme:
-            os.environ['VLLM_USE_TRITON_FLASH_ATTN'] = '1'
         logger.info("Using ROCmFlashAttention backend.")
         return "vllm.attention.backends.rocm_flash_attn.ROCmFlashAttentionBackend"  # noqa: E501