GithubStar/ncnn
1
0
Fork 0
mirror of https://github.com/Tencent/ncnn.git synced 2026-10-18 13:42:18 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
master
ncnn/tests/test_gemm_oom.cpp
T

518 lines
17 KiB
C++
Raw Permalink Blame History

// Copyright 2025 Tencent
// SPDX-License-Identifier: BSD-3-Clause
#include "testutil.h"
static int test_gemm_oom(int M, int N, int K, int transA, int transB, int output_transpose, int constantA, int constantB, int output_N1M)
{
ncnn::ParamDict pd;
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, 1);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, -1);
pd.set(11, output_N1M);
pd.set(14, output_transpose);
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? (output_N1M ? ncnn::Mat(M, 1, K) : ncnn::Mat(M, K)) : (output_N1M ? ncnn::Mat(K, 1, M) : ncnn::Mat(K, M)));
if (constantB) weights.push_back(transB ? (output_N1M ? ncnn::Mat(K, 1, N) : ncnn::Mat(K, N)) : (output_N1M ? ncnn::Mat(N, 1, K) : ncnn::Mat(N, K)));
std::vector<ncnn::Mat> a;
if (!constantA) a.push_back(transA ? (output_N1M ? ncnn::Mat(M, 1, K) : ncnn::Mat(M, K)) : (output_N1M ? ncnn::Mat(K, 1, M) : ncnn::Mat(K, M)));
if (!constantB) a.push_back(transB ? (output_N1M ? ncnn::Mat(K, 1, N) : ncnn::Mat(K, N)) : (output_N1M ? ncnn::Mat(N, 1, K) : ncnn::Mat(N, K)));
for (size_t i = 0; i < weights.size(); i++)
{
Randomize(weights[i]);
}
for (size_t i = 0; i < a.size(); i++)
{
Randomize(a[i]);
}
int ret = test_layer_oom("Gemm", pd, weights, a);
if (ret != 0)
{
fprintf(stderr, "test_gemm_oom failed M=%d N=%d K=%d transA=%d transB=%d output_transpose=%d constantA=%d constantB=%d output_N1M=%d\n", M, N, K, transA, transB, output_transpose, constantA, constantB, output_N1M);
}
return ret;
}
static int test_gemm_bias_oom(int M, int N, int K, const ncnn::Mat& C, float alpha, float beta, int transA, int transB, int output_transpose, int constantA, int constantB, int constantC)
{
int broadcast_type_C = 0;
if (C.dims == 1 && C.w == 1)
{
// scalar
broadcast_type_C = 0;
}
if (C.dims == 1 && C.w == M)
{
// M
// auto broadcast from h to w is the ncnn-style convention
broadcast_type_C = 1;
}
if (C.dims == 1 && C.w == N)
{
// N
broadcast_type_C = 4;
}
if (C.dims == 2 && C.w == 1 && C.h == M)
{
// Mx1
broadcast_type_C = 2;
}
if (C.dims == 2 && C.w == N && C.h == M)
{
// MxN
broadcast_type_C = 3;
}
if (C.dims == 2 && C.w == N && C.h == 1)
{
// 1xN
broadcast_type_C = 4;
}
ncnn::ParamDict pd;
pd.set(0, alpha);
pd.set(1, beta);
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, constantC);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, broadcast_type_C);
pd.set(14, output_transpose);
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? ncnn::Mat(M, K) : ncnn::Mat(K, M));
if (constantB) weights.push_back(transB ? ncnn::Mat(K, N) : ncnn::Mat(N, K));
if (constantC) weights.push_back(C);
std::vector<ncnn::Mat> a;
if (!constantA) a.push_back(transA ? ncnn::Mat(M, K) : ncnn::Mat(K, M));
if (!constantB) a.push_back(transB ? ncnn::Mat(K, N) : ncnn::Mat(N, K));
if (!constantC) a.push_back(C);
for (size_t i = 0; i < weights.size(); i++)
{
Randomize(weights[i]);
}
for (size_t i = 0; i < a.size(); i++)
{
Randomize(a[i]);
}
int ret = test_layer_oom("Gemm", pd, weights, a);
if (ret != 0)
{
fprintf(stderr, "test_gemm_bias_oom failed M=%d N=%d K=%d C.dims=%d C=(%d %d %d) alpha=%f beta=%f transA=%d transB=%d output_transpose=%d constantA=%d constantB=%d constantC=%d\n", M, N, K, C.dims, C.w, C.h, C.c, alpha, beta, transA, transB, output_transpose, constantA, constantB, constantC);
}
return ret;
}
static int test_gemm_0(int M, int N, int K)
{
return 0
|| test_gemm_oom(M, N, K, 0, 0, 0, 0, 0, 0)
|| test_gemm_oom(M, N, K, 1, 0, 1, 0, 0, 0)
|| test_gemm_oom(M, N, K, 0, 1, 1, 1, 0, 0)
|| test_gemm_oom(M, N, K, 1, 1, 0, 1, 0, 1)
|| test_gemm_oom(M, N, K, 0, 0, 0, 0, 1, 0)
|| test_gemm_oom(M, N, K, 0, 1, 1, 0, 1, 0)
|| test_gemm_oom(M, N, K, 1, 0, 1, 1, 1, 0)
|| test_gemm_oom(M, N, K, 1, 1, 0, 1, 1, 1);
}
static int test_gemm_1(int M, int N, int K)
{
return 0
|| test_gemm_bias_oom(M, N, K, RandomMat(1), 2.1f, 0.5f, 0, 0, 0, 0, 0, 0)
|| test_gemm_bias_oom(M, N, K, RandomMat(M), 3.1f, 0.6f, 0, 1, 0, 1, 0, 0)
|| test_gemm_bias_oom(M, N, K, RandomMat(1, M), 4.1f, 0.7f, 1, 0, 1, 0, 1, 0)
|| test_gemm_bias_oom(M, N, K, RandomMat(N, M), 5.1f, 0.8f, 1, 1, 1, 1, 1, 0)
|| test_gemm_bias_oom(M, N, K, RandomMat(N, 1), 2.1f, 0.5f, 0, 0, 0, 0, 0, 1)
|| test_gemm_bias_oom(M, N, K, RandomMat(N), 3.1f, 0.6f, 0, 1, 0, 1, 0, 1)
|| test_gemm_bias_oom(M, N, K, RandomMat(M), 3.1f, 0.6f, 0, 1, 0, 0, 1, 1)
|| test_gemm_bias_oom(M, N, K, RandomMat(N, M), 5.1f, 0.8f, 1, 1, 1, 1, 1, 1);
}
#if NCNN_WEIGHT_QUANT
static int test_gemm_wq_int8_oom(int M, int N, int K, int block_size, int input_scale, int output_transpose, int dynamic_C = 0)
{
const int block_count = (K + block_size - 1) / block_size;
const int block_size_code = block_size == 32 ? 0 : block_size == 64 ? 1 : 2;
ncnn::ParamDict pd;
pd.set(0, 1.7f);
pd.set(1, 0.3f);
pd.set(2, 0);
pd.set(3, 1);
pd.set(4, 0);
pd.set(5, 1);
pd.set(6, dynamic_C ? 0 : 1);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, dynamic_C ? 3 : 4);
pd.set(14, output_transpose);
pd.set(18, 800 + input_scale * 10 + block_size_code);
std::vector<ncnn::Mat> weights;
weights.push_back(RandomS8Mat(K, N));
if (!dynamic_C)
weights.push_back(RandomMat(N));
weights.push_back(RandomMat(block_count, N, 10.f, 20.f));
if (input_scale)
weights.push_back(RandomMat(K, 0.5f, 1.5f));
std::vector<ncnn::Mat> a(1, RandomMat(K, M));
if (dynamic_C)
a.push_back(RandomMat(N, M));
int ret = test_layer_oom("Gemm", pd, weights, a, 1, TEST_LAYER_ENABLE_THREADING);
if (ret != 0)
{
fprintf(stderr, "test_gemm_wq_int8_oom failed M=%d N=%d K=%d block_size=%d input_scale=%d output_transpose=%d dynamic_C=%d\n", M, N, K, block_size, input_scale, output_transpose, dynamic_C);
}
return ret;
}
#endif // NCNN_WEIGHT_QUANT
#if NCNN_INT8
static int test_gemm_int8_oom(int M, int N, int K, int transA, int transB, int output_elemtype, int output_transpose, int constantA, int constantB, int output_N1M)
{
ncnn::ParamDict pd;
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, 1);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, -1);
pd.set(11, output_N1M);
pd.set(13, output_elemtype);
pd.set(14, output_transpose);
pd.set(18, 2); // int8_scale_term
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? RandomS8Mat(M, K) : RandomS8Mat(K, M));
if (constantB) weights.push_back(transB ? RandomS8Mat(K, N) : RandomS8Mat(N, K));
if (constantA) weights.push_back(RandomMat(M, 10.f, 20.f));
if (constantB) weights.push_back(RandomMat(1, 10.f, 20.f));
std::vector<ncnn::Mat> a;
if (!constantA)
{
a.push_back(transA ? (output_N1M ? RandomMat(M, 1, K) : RandomMat(M, K)) : (output_N1M ? RandomMat(K, 1, M) : RandomMat(K, M)));
}
if (!constantB)
{
a.push_back(transB ? (output_N1M ? RandomMat(K, 1, N) : RandomMat(K, N)) : (output_N1M ? RandomMat(N, 1, K) : RandomMat(N, K)));
}
int ret = test_layer_oom("Gemm", pd, weights, a);
if (ret != 0)
{
fprintf(stderr, "test_gemm_int8_oom failed M=%d N=%d K=%d transA=%d transB=%d output_elemtype=%d output_transpose=%d constantA=%d constantB=%d output_N1M=%d\n", M, N, K, transA, transB, output_elemtype, output_transpose, constantA, constantB, output_N1M);
}
return ret;
}
static int test_gemm_int8_bias_oom(int M, int N, int K, const ncnn::Mat& C, float alpha, float beta, int transA, int transB, int output_elemtype, int output_transpose, int constantA, int constantB, int constantC)
{
int broadcast_type_C = 0;
if (C.dims == 1 && C.w == 1)
{
// scalar
broadcast_type_C = 0;
}
if (C.dims == 1 && C.w == M)
{
// M
// auto broadcast from h to w is the ncnn-style convention
broadcast_type_C = 1;
}
if (C.dims == 1 && C.w == N)
{
// N
broadcast_type_C = 4;
}
if (C.dims == 2 && C.w == 1 && C.h == M)
{
// Mx1
broadcast_type_C = 2;
}
if (C.dims == 2 && C.w == N && C.h == M)
{
// MxN
broadcast_type_C = 3;
}
if (C.dims == 2 && C.w == N && C.h == 1)
{
// 1xN
broadcast_type_C = 4;
}
ncnn::ParamDict pd;
pd.set(0, alpha);
pd.set(1, beta);
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, constantC);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, broadcast_type_C);
pd.set(13, output_elemtype);
pd.set(14, output_transpose);
pd.set(18, 2); // int8_scale_term
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? RandomS8Mat(M, K) : RandomS8Mat(K, M));
if (constantB) weights.push_back(transB ? RandomS8Mat(K, N) : RandomS8Mat(N, K));
if (constantC) weights.push_back(C);
if (constantA) weights.push_back(RandomMat(M, 10.f, 20.f));
if (constantB) weights.push_back(RandomMat(1, 10.f, 20.f));
std::vector<ncnn::Mat> a;
if (!constantA)
{
a.push_back(transA ? RandomMat(M, K) : RandomMat(K, M));
}
if (!constantB)
{
a.push_back(transB ? RandomMat(K, N) : RandomMat(N, K));
}
if (!constantC) a.push_back(C);
int ret = test_layer_oom("Gemm", pd, weights, a);
if (ret != 0)
{
fprintf(stderr, "test_gemm_int8_bias_oom failed M=%d N=%d K=%d C.dims=%d C=(%d %d %d) alpha=%f beta=%f transA=%d transB=%d output_elemtype=%d output_transpose=%d constantA=%d constantB=%d constantC=%d\n", M, N, K, C.dims, C.w, C.h, C.c, alpha, beta, transA, transB, output_elemtype, output_transpose, constantA, constantB, constantC);
}
return ret;
}
static int test_gemm_int8_fp16s_oom(int M, int N, int K, int transA, int transB, int output_elemtype, int output_transpose, int constantA, int constantB, int output_N1M)
{
ncnn::ParamDict pd;
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, 1);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, -1);
pd.set(11, output_N1M);
pd.set(13, output_elemtype);
pd.set(14, output_transpose);
pd.set(18, 2); // int8_scale_term
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? RandomS8Mat(M, K) : RandomS8Mat(K, M));
if (constantB) weights.push_back(transB ? RandomS8Mat(K, N) : RandomS8Mat(N, K));
if (constantA) weights.push_back(RandomMat(M, 10.f, 20.f));
if (constantB) weights.push_back(RandomMat(1, 10.f, 20.f));
std::vector<ncnn::Mat> a;
if (!constantA)
{
a.push_back(transA ? (output_N1M ? RandomMat(M, 1, K) : RandomMat(M, K)) : (output_N1M ? RandomMat(K, 1, M) : RandomMat(K, M)));
}
if (!constantB)
{
a.push_back(transB ? (output_N1M ? RandomMat(K, 1, N) : RandomMat(K, N)) : (output_N1M ? RandomMat(N, 1, K) : RandomMat(N, K)));
}
ncnn::Option opt;
opt.num_threads = 1;
opt.use_packing_layout = true;
opt.use_fp16_packed = false;
opt.use_fp16_storage = true;
opt.use_fp16_arithmetic = false;
opt.use_bf16_storage = false;
float epsilon = 0.001;
int ret = test_layer_oom_opt("Gemm", pd, weights, opt, a, 1, epsilon);
if (ret != 233 && ret != 0)
{
fprintf(stderr, "test_gemm_int8_fp16s_oom failed M=%d N=%d K=%d transA=%d transB=%d output_elemtype=%d output_transpose=%d constantA=%d constantB=%d output_N1M=%d\n", M, N, K, transA, transB, output_elemtype, output_transpose, constantA, constantB, output_N1M);
return ret;
}
return 0;
}
static int test_gemm_2(int M, int N, int K)
{
return 0
|| test_gemm_int8_oom(M, N, K, 0, 0, 0, 0, 0, 0, 0)
|| test_gemm_int8_oom(M, N, K, 1, 0, 0, 1, 0, 0, 0)
|| test_gemm_int8_oom(M, N, K, 0, 1, 1, 1, 1, 0, 0)
|| test_gemm_int8_oom(M, N, K, 1, 1, 0, 0, 1, 0, 1)
|| test_gemm_int8_oom(M, N, K, 0, 0, 2, 0, 0, 1, 0)
|| test_gemm_int8_oom(M, N, K, 0, 1, 0, 1, 0, 1, 0)
|| test_gemm_int8_oom(M, N, K, 1, 0, 0, 1, 1, 1, 0)
|| test_gemm_int8_oom(M, N, K, 1, 1, 3, 0, 1, 1, 1);
}
static int test_gemm_3(int M, int N, int K)
{
return 0
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(1), 2.1f, 0.5f, 0, 0, 0, 0, 0, 0, 0)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(M), 3.1f, 0.6f, 0, 1, 0, 0, 1, 0, 0)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(1, M), 4.1f, 0.7f, 1, 0, 1, 1, 0, 1, 0)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(N, M), 5.1f, 0.8f, 1, 1, 0, 1, 1, 1, 0)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(N, 1), 2.1f, 0.5f, 0, 0, 3, 0, 0, 0, 1)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(N), 3.1f, 0.6f, 0, 1, 0, 0, 1, 0, 1)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(M), 3.1f, 0.6f, 0, 1, 0, 0, 0, 1, 1)
|| test_gemm_int8_bias_oom(M, N, K, RandomMat(N, M), 5.1f, 0.8f, 1, 1, 2, 1, 1, 1, 1);
}
static int test_gemm_4(int M, int N, int K)
{
return 0
|| test_gemm_int8_fp16s_oom(M, N, K, 0, 0, 0, 0, 0, 0, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 1, 0, 0, 1, 0, 0, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 0, 1, 3, 1, 1, 0, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 1, 1, 0, 0, 1, 0, 1)
|| test_gemm_int8_fp16s_oom(M, N, K, 0, 0, 2, 0, 0, 1, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 0, 1, 0, 1, 0, 1, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 1, 0, 0, 1, 1, 1, 0)
|| test_gemm_int8_fp16s_oom(M, N, K, 1, 1, 1, 0, 1, 1, 1);
}
#endif // NCNN_INT8
static int test_gemm_nt_oom(int M, int N, int K, int transA, int transB, int output_transpose, int constantA, int constantB)
{
ncnn::ParamDict pd;
pd.set(2, transA);
pd.set(3, transB);
pd.set(4, constantA);
pd.set(5, constantB);
pd.set(6, 1);
pd.set(7, M);
pd.set(8, N);
pd.set(9, K);
pd.set(10, -1);
pd.set(14, output_transpose);
std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? ncnn::Mat(M, K) : ncnn::Mat(K, M));
if (constantB) weights.push_back(transB ? ncnn::Mat(K, N) : ncnn::Mat(N, K));
std::vector<ncnn::Mat> a;
if (!constantA) a.push_back(transA ? ncnn::Mat(M, K) : ncnn::Mat(K, M));
if (!constantB) a.push_back(transB ? ncnn::Mat(K, N) : ncnn::Mat(N, K));
for (size_t i = 0; i < weights.size(); i++)
{
Randomize(weights[i]);
}
for (size_t i = 0; i < a.size(); i++)
{
Randomize(a[i]);
}
int ret = test_layer_oom("Gemm", pd, weights, a, 1, TEST_LAYER_ENABLE_THREADING);
if (ret != 0)
{
fprintf(stderr, "test_gemm_nt_oom failed M=%d N=%d K=%d transA=%d transB=%d output_transpose=%d constantA=%d constantB=%d\n", M, N, K, transA, transB, output_transpose, constantA, constantB);
}
return ret;
}
static int test_gemm_5(int M, int N, int K)
{
return 0
|| test_gemm_nt_oom(M, N, K, 0, 0, 0, 0, 0)
|| test_gemm_nt_oom(M, N, K, 0, 1, 0, 0, 0)
|| test_gemm_nt_oom(M, N, K, 1, 0, 1, 0, 0)
|| test_gemm_nt_oom(M, N, K, 1, 1, 1, 0, 0)
|| test_gemm_nt_oom(M, N, K, 0, 0, 1, 1, 0)
|| test_gemm_nt_oom(M, N, K, 0, 1, 1, 1, 0)
|| test_gemm_nt_oom(M, N, K, 1, 0, 0, 1, 0)
|| test_gemm_nt_oom(M, N, K, 1, 1, 0, 1, 0)
|| test_gemm_nt_oom(M, N, K, 0, 0, 0, 0, 1)
|| test_gemm_nt_oom(M, N, K, 0, 1, 1, 0, 1)
|| test_gemm_nt_oom(M, N, K, 1, 0, 0, 0, 1)
|| test_gemm_nt_oom(M, N, K, 1, 1, 1, 0, 1)
|| test_gemm_nt_oom(M, N, K, 0, 0, 1, 1, 1)
|| test_gemm_nt_oom(M, N, K, 0, 1, 0, 1, 1)
|| test_gemm_nt_oom(M, N, K, 1, 0, 1, 1, 1)
|| test_gemm_nt_oom(M, N, K, 1, 1, 0, 1, 1);
}
int main()
{
SRAND(7767517);
int mnk[][3] = {
{11, 12, 13},
{1, 2, 3},
{4, 1, 6},
{9, 8, 7}
};
int mnk_count = sizeof(mnk) / sizeof(int) / 3;
for (int i = 0; i < mnk_count; i++)
{
int M = mnk[i][0];
int N = mnk[i][1];
int K = mnk[i][2];
int ret = test_gemm_0(M, N, K) || test_gemm_1(M, N, K);
if (ret != 0)
return ret;
#if NCNN_INT8
int ret2 = test_gemm_2(M, N, K) || test_gemm_3(M, N, K) || test_gemm_4(M, N, K);
if (ret2 != 0)
return ret2;
#endif
int ret3 = test_gemm_5(M, N, K);
if (ret3 != 0)
return ret3;
}
#if NCNN_WEIGHT_QUANT
return 0
|| test_gemm_wq_int8_oom(3, 5, 65, 32, 0, 0)
|| test_gemm_wq_int8_oom(8, 17, 129, 128, 1, 1)
|| test_gemm_wq_int8_oom(7, 9, 67, 32, 1, 0, 1)
|| test_gemm_wq_int8_oom(16, 16, 35, 32, 1, 1, 1);
#else
return 0;
#endif
}
Reference in New Issue View Git Blame Copy Permalink