vllm.model_executor.layers.quantization.compressed_tensors.triton_scaled_mm

is_weak_contiguous

is_weak_contiguous(x: Tensor)

Source code in vllm/model_executor/layers/quantization/compressed_tensors/triton_scaled_mm.py

def is_weak_contiguous(x: torch.Tensor):
    strides = x.stride()
    sizes = x.shape
    is_not_transpose = strides[0] == 1 and (strides[1] >= max(1, sizes[0]))
    is_transpose = strides[1] == 1 and (strides[0] >= max(1, sizes[1]))
    return is_transpose or is_not_transpose

scaled_mm_kernel

scaled_mm_kernel(
    a_ptr,
    b_ptr,
    scale_a_ptr,
    scale_b_ptr,
    c_ptr,
    bias_ptr,
    M,
    N,
    K,
    stride_am,
    stride_ak,
    stride_bk,
    stride_bn,
    stride_cm,
    stride_cn,
    ACCUMULATOR_DTYPE: constexpr,
    BLOCK_SIZE_M: constexpr,
    BLOCK_SIZE_N: constexpr,
    BLOCK_SIZE_K: constexpr,
    BLOCK_SIZE_SCALE_A: constexpr,
    BLOCK_SIZE_SCALE_B: constexpr,
)

Source code in vllm/model_executor/layers/quantization/compressed_tensors/triton_scaled_mm.py

@triton.jit
def scaled_mm_kernel(a_ptr, b_ptr, scale_a_ptr, scale_b_ptr, c_ptr, bias_ptr,
                     M, N, K, stride_am, stride_ak, stride_bk, stride_bn,
                     stride_cm, stride_cn, ACCUMULATOR_DTYPE: tl.constexpr,
                     BLOCK_SIZE_M: tl.constexpr, BLOCK_SIZE_N: tl.constexpr,
                     BLOCK_SIZE_K: tl.constexpr,
                     BLOCK_SIZE_SCALE_A: tl.constexpr,
                     BLOCK_SIZE_SCALE_B: tl.constexpr):
    pid = tl.program_id(axis=0)

    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)

    pid_m = pid // num_pid_n
    pid_n = pid % num_pid_n

    accumulator_dtype = ACCUMULATOR_DTYPE
    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N),
                           dtype=accumulator_dtype)

    # NOTE: Some tensor inputs are so large, they will cause int32 overflow
    # so it is necessary to use tl.int64 for all the offsets, else SEGV will
    # eventually occur.

    # Offsets and masks.
    offsets_am = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
    masks_am = offsets_am < M

    offsets_bn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)
    masks_bn = offsets_bn < N

    offsets_k = tl.arange(0, BLOCK_SIZE_K).to(tl.int64)
    offsets_a = (stride_am * offsets_am[:, None] +
                 stride_ak * offsets_k[None, :])
    offsets_b = (stride_bk * offsets_k[:, None] +
                 stride_bn * offsets_bn[None, :])

    # NOTE: BLOCK_SIZE_SCALE_A could be 1 or BLOCK_SIZE_M, so need to create
    # appropriate offsets and masks for each case. Same goes for
    # BLOCK_SIZE_SCALE_B.
    offsets_scale_am = (tl.arange(0, BLOCK_SIZE_SCALE_A) +
                        (BLOCK_SIZE_SCALE_A > 1) * pid_m * BLOCK_SIZE_M)
    masks_scale_am = offsets_scale_am < M

    offsets_scale_bn = (tl.arange(0, BLOCK_SIZE_SCALE_B) +
                        (BLOCK_SIZE_SCALE_B > 1) * pid_n * BLOCK_SIZE_N)
    masks_scale_bn = offsets_scale_bn < N

    a_ptrs = a_ptr + offsets_a
    b_ptrs = b_ptr + offsets_b

    scale_a_ptrs = scale_a_ptr + offsets_scale_am
    scale_b_ptrs = scale_b_ptr + offsets_scale_bn

    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K)):
        masks_k = offsets_k < K
        masks_a = masks_am[:, None] & masks_k[None, :]
        a = tl.load(a_ptrs, mask=masks_a)

        masks_b = masks_k[:, None] & masks_bn[None, :]
        b = tl.load(b_ptrs, mask=masks_b)

        # Accumulate results.
        accumulator = tl.dot(a, b, accumulator, out_dtype=accumulator_dtype)

        offsets_k += BLOCK_SIZE_K
        a_ptrs += BLOCK_SIZE_K * stride_ak
        b_ptrs += BLOCK_SIZE_K * stride_bk

    # Apply scale at end.
    masks_scale_a = masks_scale_am[:, None] & (tl.arange(0, 1) < 1)[:, None]
    scale_a = tl.load(scale_a_ptrs[:, None], masks_scale_a)
    # Need to broadcast to the appropriate size, if scale_a is already
    # (BLOCK_SIZE_M, 1) then it will broadcast to its own shape. Same goes
    # for scale_b below.
    scale_a = scale_a.broadcast_to((BLOCK_SIZE_M, 1))
    accumulator = scale_a * accumulator.to(tl.float32)

    masks_scale_b = masks_scale_bn[:, None] & (tl.arange(0, 1) < 1)[None, :]
    scale_b = tl.load(scale_b_ptrs[:, None], masks_scale_b)
    scale_b = scale_b.broadcast_to((BLOCK_SIZE_N, 1))
    accumulator = scale_b.T * accumulator.to(tl.float32)

    # Convert to output format.
    c = accumulator.to(c_ptr.type.element_ty)

    # Add bias, it's already in output format, so add it after conversion.
    if bias_ptr:
        offsets_bias = offsets_bn
        bias_ptrs = bias_ptr + offsets_bias
        bias_mask = offsets_bias < N
        bias = tl.load(bias_ptrs, bias_mask)
        c += bias

    # Save output
    offs_cm = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M).to(tl.int64)
    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N).to(tl.int64)
    offs_cm = offs_cm.to(tl.int64)
    offs_cn = offs_cn.to(tl.int64)
    c_ptrs = (c_ptr + stride_cm * offs_cm[:, None] +
              stride_cn * offs_cn[None, :])
    c_mask = (offs_cm[:, None] < M) & (offs_cn[None, :] < N)

    tl.store(c_ptrs, c, mask=c_mask)

triton_scaled_mm

triton_scaled_mm(
    input: Tensor,
    weight: Tensor,
    scale_a: Tensor,
    scale_b: Tensor,
    out_dtype: type[dtype],
    bias: Optional[Tensor] = None,
    block_size_m: int = 32,
    block_size_n: int = 32,
    block_size_k: int = 32,
    use_heuristic=True,
) -> Tensor

Source code in vllm/model_executor/layers/quantization/compressed_tensors/triton_scaled_mm.py

def triton_scaled_mm(input: torch.Tensor,
                     weight: torch.Tensor,
                     scale_a: torch.Tensor,
                     scale_b: torch.Tensor,
                     out_dtype: type[torch.dtype],
                     bias: Optional[torch.Tensor] = None,
                     block_size_m: int = 32,
                     block_size_n: int = 32,
                     block_size_k: int = 32,
                     use_heuristic=True) -> torch.Tensor:
    M, K = input.shape
    N = weight.shape[1]

    assert N > 0 and K > 0 and M > 0
    assert weight.shape[0] == K
    assert input.dtype == weight.dtype

    scale_a = scale_a.reshape(-1, 1) if scale_a.dim() <= 1 else scale_a
    scale_b = scale_b.reshape(-1, 1) if scale_b.dim() <= 1 else scale_b

    assert scale_a.dtype == scale_b.dtype and scale_a.is_floating_point()
    assert scale_a.shape[1] == 1 and (scale_a.shape[0] == 1
                                      or scale_a.shape[0] == M)
    assert scale_b.shape[1] == 1 and (scale_b.shape[0] == 1
                                      or scale_b.shape[0] == N)
    assert out_dtype.is_floating_point
    assert bias is None or bias.is_floating_point()
    assert is_weak_contiguous(input)
    assert is_weak_contiguous(weight)

    grid = lambda META: (triton.cdiv(M, META['BLOCK_SIZE_M']) * triton.cdiv(
        N, META['BLOCK_SIZE_N']), )

    result = torch.empty((M, N), dtype=out_dtype, device=input.device)

    has_scalar = lambda x: x.shape[0] == 1 and x.shape[1] == 1

    if use_heuristic:
        is_small_N = N < 8192
        next_power_of_2_M = max(32, triton.next_power_of_2(M))
        if next_power_of_2_M <= 32:
            tile_shape = (64, 64, 256) if is_small_N else (64, 128, 256)
        elif next_power_of_2_M <= 64:
            tile_shape = (64, 64, 256)
        elif next_power_of_2_M <= 128:
            tile_shape = (64, 128, 128)
        else:
            tile_shape = (128, 128, 128)

    block_size_m, block_size_n, block_size_k = tile_shape

    block_size_sa = 1 if has_scalar(scale_a) else block_size_m
    block_size_sb = 1 if has_scalar(scale_b) else block_size_n

    accumulator_dtype = tl.float32 if input.is_floating_point() else tl.int32

    # A = input, B = weight, C = result
    # A = M x K, B = K x N, C = M x N
    scaled_mm_kernel[grid](input,
                           weight,
                           scale_a,
                           scale_b,
                           result,
                           bias,
                           M,
                           N,
                           K,
                           input.stride(0),
                           input.stride(1),
                           weight.stride(0),
                           weight.stride(1),
                           result.stride(0),
                           result.stride(1),
                           accumulator_dtype,
                           BLOCK_SIZE_M=block_size_m,
                           BLOCK_SIZE_N=block_size_n,
                           BLOCK_SIZE_K=block_size_k,
                           BLOCK_SIZE_SCALE_A=block_size_sa,
                           BLOCK_SIZE_SCALE_B=block_size_sb)

    return result.to(out_dtype)