"symphony/vscode:/vscode.git/clone" did not exist on "6c6572013f4f24d4a9d034d8db36d2e2606c0cc7"
example_deepgemm_fp8_2xAcc.py 6.63 KB
Newer Older
1
2
3
4
from typing import Tuple

import torch
import tilelang.testing
5
import tilelang
6
7
8
import tilelang.language as T
from tilelang.utils.tensor import map_torch_type

Taoyu Zhu's avatar
Taoyu Zhu committed
9
tilelang.testing.set_random_seed(42)
10
11


12
@tilelang.jit
13
14
15
16
def tl_gemm(
    M,
    N,
    K,
17
    block_N,
18
19
20
21
22
    in_dtype,
    out_dtype,
    accum_dtype,
):
    assert in_dtype in [
23
24
        "float8_e4m3",
    ], "Currently only float8_e4m3 is supported"
25
26
27
28
29
    assert out_dtype in [
        "bfloat16",
        "float32",
    ], "Currently only float16 and float32 are supported"

30
31
32
    group_size = 128
    block_M = 128
    block_K = 128
33
34

    A_shape = (M, K)
35
    Scales_A_shape = (M, T.ceildiv(K, group_size))
36
    B_shape = (N, K)
37
    Scales_B_shape = (T.ceildiv(N, group_size), T.ceildiv(K, group_size))
38
39
40
41
42
43
    A_shared_shape = (block_M, block_K)
    B_shared_shape = (block_N, block_K)
    C_shared_shape = (block_M, block_N)

    @T.prim_func
    def main(
44
45
46
47
48
            A: T.Tensor(A_shape, in_dtype),
            B: T.Tensor(B_shape, in_dtype),
            C: T.Tensor((M, N), out_dtype),
            scales_a: T.Tensor(Scales_A_shape, "float32"),
            scales_b: T.Tensor(Scales_B_shape, "float32"),
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
    ):
        with T.Kernel(T.ceildiv(N, block_N), T.ceildiv(M, block_M), threads=128) as (bx, by):

            A_shared = T.alloc_shared(A_shared_shape, in_dtype)
            B_shared = T.alloc_shared(B_shared_shape, in_dtype)
            C_shared = T.alloc_shared(C_shared_shape, out_dtype)
            Scale_C_shared = T.alloc_shared((block_M), "float32")
            C_local = T.alloc_fragment(C_shared_shape, accum_dtype)
            C_local_accum = T.alloc_fragment(C_shared_shape, accum_dtype)

            # Improve L2 Cache
            T.use_swizzle(panel_size=10)

            T.clear(C_local)
            T.clear(C_local_accum)
            K_iters = T.ceildiv(K, block_K)
            for k in T.Pipelined(K_iters, num_stages=4):
                # Load A into shared memory
                T.copy(A[by * block_M, k * block_K], A_shared)
                # Load B into shared memory
                T.copy(B[bx * block_N, k * block_K], B_shared)
                # Load scale into shared memory
71
                Scale_B = scales_b[bx * block_N // group_size, k]
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
                for i in T.Parallel(block_M):
                    Scale_C_shared[i] = scales_a[by * block_M + i, k] * Scale_B

                T.gemm(A_shared, B_shared, C_local, transpose_B=True)
                # Promote to enable 2xAcc
                for i, j in T.Parallel(block_M, block_N):
                    C_local_accum[i, j] += C_local[i, j] * Scale_C_shared[i]
                T.clear(C_local)
            # TMA store
            T.copy(C_local_accum, C_shared)
            T.copy(C_shared, C[by * block_M, bx * block_N])

    return main


def ceildiv(a, b):
    return (a + b - 1) // b


def per_token_cast_to_fp8(x: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
    assert x.dim() == 2 and x.size(1) % 128 == 0
    m, n = x.shape
    x_view = x.view(m, -1, 128)
    x_amax = x_view.abs().float().amax(dim=2).view(m, -1).clamp(1e-4)
    return (x_view * (448.0 / x_amax.unsqueeze(2))).to(torch.float8_e4m3fn).view(
        m, n), (x_amax / 448.0).view(m, -1)


def per_block_cast_to_fp8(x: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
    assert x.dim() == 2
    m, n = x.shape
    x_padded = torch.zeros(
        ceildiv(m, 128) * 128, ceildiv(n, 128) * 128, dtype=x.dtype, device=x.device)
    x_padded[:m, :n] = x
    x_view = x_padded.view(-1, 128, x_padded.size(1) // 128, 128)
    x_amax = x_view.abs().float().amax(dim=(1, 3), keepdim=True).clamp(1e-4)
    x_scaled = (x_view * (448.0 / x_amax)).to(torch.float8_e4m3fn)
    return x_scaled.view_as(x_padded)[:m, :n].contiguous(), (x_amax / 448.0).view(
        x_view.size(0), x_view.size(2))


def ref_deepgemm_fp8(A_fp8, B_fp8, A_scale, B_scale, out_dtype):
    # A_scale: (M, K//128)       ==>   (M//128, K//128, 128)
    # B_scale: (N//128, K//128)  ==>   (N//128, K//128, 128)
    # A_fp8: (M, K)
    # B_fp8: (N, K)
    # out_dtype: float16 or float32
    # return C: (M, N)
    M, N, K = A_fp8.shape[0], B_fp8.shape[0], A_fp8.shape[1]
    A_scales = A_scale.view(M // 128, 128, K // 128).permute(0, 2, 1)
    B_scales = B_scale.repeat_interleave(128, dim=1).view(N // 128, K // 128, 128)
    C = torch.zeros(M, N, device="cuda", dtype=out_dtype)
    c_acc = torch.zeros(128, 128, device="cuda", dtype=torch.float32)
    for i in range(ceildiv(M, 128)):
        for j in range(ceildiv(N, 128)):
            c_acc.zero_()
            for k in range(ceildiv(K, 128)):
                c = torch._scaled_mm(
                    A_fp8[i * 128:(i + 1) * 128, k * 128:(k + 1) * 128],
                    B_fp8[j * 128:(j + 1) * 128, k * 128:(k + 1) * 128].T,
                    scale_a=A_scales[i, k].view(128, 1).contiguous(),
                    scale_b=B_scales[j, k].view(1, 128).contiguous(),
                    out_dtype=torch.bfloat16)
                c_acc += c.to(torch.float32)
            C[i * 128:(i + 1) * 128, j * 128:(j + 1) * 128] = c_acc.to(out_dtype)
    return C


def calc_diff(x, y):
    x, y = x.double(), y.double()
    denominator = (x * x + y * y).sum()
    sim = 2 * (x * y).sum() / denominator
    return 1 - sim


Taoyu Zhu's avatar
Taoyu Zhu committed
147
def assert_tl_gemm_correctness(M, N, K, block_N, in_dtype, out_dtype, accum_dtype):
148
    kernel = tl_gemm(M, N, K, block_N, in_dtype, out_dtype, accum_dtype)
149
    src_code = kernel.get_kernel_source()
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164

    # src_code is the generated cuda source
    assert src_code is not None

    in_dtype = map_torch_type(in_dtype)
    out_dtype = map_torch_type(out_dtype)
    accum_dtype = map_torch_type(accum_dtype)

    A = torch.randn(M, K).to(torch.bfloat16).cuda()
    B = torch.randn(N, K).to(torch.bfloat16).cuda()
    A_fp8, A_scale = per_token_cast_to_fp8(A.clone())
    B_fp8, B_scale = per_block_cast_to_fp8(B.clone())

    C = torch.zeros(M, N, device="cuda", dtype=out_dtype)

165
    kernel(A_fp8, B_fp8, C, A_scale, B_scale)
166
167
168
169
170
171
    # Get Reference Result
    ref_c = ref_deepgemm_fp8(A_fp8, B_fp8, A_scale, B_scale, out_dtype)
    diff = calc_diff(C, ref_c)
    print(f"diff: {diff}")
    assert diff < 1e-3

172
173
    profiler = kernel.get_profiler()
    latency = profiler.do_bench(warmup=25)
174
175
176
177
178
179
180
    # Ensure that the latency is not None
    assert latency is not None
    print(f"latency: {latency} ms")
    tflops = 2 * M * N * K / latency / 1e9
    print(f"tflops: {tflops}")


181
def main():
182
    assert_tl_gemm_correctness(1024, 1024, 8192, 128, "float8_e4m3", "bfloat16", "float32")
183
184


185
if __name__ == "__main__":
186
    for dtype in ["float8_e4m3"]:
187
        for out_dtype in ["bfloat16", "float32"]:
188
189
            for block_N in [16, 32, 64, 128]:
                assert_tl_gemm_correctness(1024, 1024, 8192, block_N, dtype, out_dtype, "float32")