* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#include "gtest/gtest.h"
#include "tikicpulib.h"
#include "kernel_operator.h"
using namespace AscendC;
#include "utils.h"
#include "scalar_sub.h"
template<class T>
void GmToUb(LocalTensor<T>& local, T* gm, int size) {
for (int i = 0; i < size; i++) {
local.SetValue(i, gm[i]);
}
}
template<class T>
void UbToGm(T* gm, LocalTensor<T>& local, int size) {
for (int i = 0; i < size; i++) {
gm[i] = local.GetValue(i);
}
}
TEST(TestApiSubs, Test_scalar_latter) {
int a = 2, b = 32;
auto *x = (half*)AscendC::GmAlloc(sizeof(half) * a * b);
auto *y = (half*)AscendC::GmAlloc(sizeof(half) * a * b);
half expect[a][b];
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
x[i * b + j] = (double)(i + 1);
expect[i][j] = (double)((i + 1) - 2.0);
}
}
auto kernel = [](int a, int b, half *x, half *y) {
half constant_x = (double)(2.0);
uint32_t cal_cnt = a * b;
TPipe tpipe;
TBuf<TPosition::VECCALC> xbuf, ybuf, tmp;
tpipe.InitBuffer(xbuf, sizeof(half) * a * b);
tpipe.InitBuffer(ybuf, sizeof(half) * a * b);
tpipe.InitBuffer(tmp, 8 * 1024);
auto l_x = xbuf.Get<half>();
auto l_y = ybuf.Get<half>();
auto l_tmp = tmp.Get<uint8_t>();
GmToUb(l_x, x, a * b);
GmToUb(l_y, y, a * b);
Subs<half>(l_y, l_x, constant_x, l_tmp, cal_cnt);
UbToGm(y, l_y, a * b);
};
AscendC::SetKernelMode(KernelMode::AIV_MODE);
ICPU_RUN_KF(kernel, 1, a, b, x, y);
int diff_count = 0;
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
auto diff = (double)(y[i*b + j] - expect[i][j]);
if (diff < -1e-5 || diff > 1e-5) {
diff_count++;
}
}
}
EXPECT_EQ(diff_count, 0);
}
TEST(TestApiSubs, Test_scalar_front) {
int a = 2, b = 4;
auto *x = (half*)AscendC::GmAlloc(sizeof(half) * a * b);
auto *y = (half*)AscendC::GmAlloc(sizeof(half) * a * b);
half expect[a][b];
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
x[i * b + j] = (double)(i + 1);
expect[i][j] = (double)(2.0 - (i + 1));
}
}
auto kernel = [](int a, int b, half *x, half *y) {
half constant_x = (double)(2.0);
uint32_t cal_cnt = 2 * 4;
TPipe tpipe;
TBuf<TPosition::VECCALC> xbuf, ybuf, tmp;
tpipe.InitBuffer(xbuf, sizeof(half) * a * b);
tpipe.InitBuffer(ybuf, sizeof(half) * a * b);
tpipe.InitBuffer(tmp, 8 * 1024);
auto l_x = xbuf.Get<half>();
auto l_y = ybuf.Get<half>();
auto l_tmp = tmp.Get<uint8_t>();
GmToUb(l_x, x, a * b);
GmToUb(l_y, y, a * b);
Subs<half, false>(l_y, l_x, constant_x, l_tmp, cal_cnt);
UbToGm(y, l_y, a * b);
};
AscendC::SetKernelMode(KernelMode::AIV_MODE);
ICPU_RUN_KF(kernel, 1, a, b, x, y);
int diff_count = 0;
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
auto diff = (double)(y[i*b + j] - expect[i][j]);
if (diff < -1e-5 || diff > 1e-5) {
diff_count++;
}
}
}
EXPECT_EQ(diff_count, 0);
}
TEST(TestApiSubs, Test_scalar_front_float) {
int calc_size = 3840;
auto *x = (float*)AscendC::GmAlloc(sizeof(float) * calc_size);
auto *y = (float*)AscendC::GmAlloc(sizeof(float) * calc_size);
float expect[calc_size];
for (int i = 0; i < calc_size; i++) {
x[i] = (float)(i + 1);
expect[i] = (float)(4000.0 - (i + 1));
}
auto kernel = [](float *x, float *y) {
float constant_x = (float)(4000.0);
uint32_t cal_cnt = 3840;
TPipe tpipe;
TBuf<TPosition::VECCALC> xbuf, ybuf, tmp;
tpipe.InitBuffer(xbuf, sizeof(float) * cal_cnt);
tpipe.InitBuffer(ybuf, sizeof(float) * cal_cnt);
tpipe.InitBuffer(tmp, 8 * 1024);
auto l_x = xbuf.Get<float>();
auto l_y = ybuf.Get<float>();
auto l_tmp = tmp.Get<uint8_t>();
GmToUb(l_x, x, cal_cnt);
GmToUb(l_y, y, cal_cnt);
Subs<float, false>(l_y, l_x, constant_x, l_tmp, cal_cnt);
UbToGm(y, l_y, cal_cnt);
};
AscendC::SetKernelMode(KernelMode::AIV_MODE);
ICPU_RUN_KF(kernel, 1, x, y);
int diff_count = 0;
for (int i = 0; i < calc_size; i++) {
auto diff = (float)(y[i] - expect[i]);
if (diff < -1e-5 || diff > 1e-5) {
diff_count++;
}
}
EXPECT_EQ(diff_count, 0);
}
TEST(TestApiSubs, Test_scalar_both) {
auto kernel = [](int a, int b, half *y) {
half constant_x = (double)(2.0);
half constant_y = (double)(1.0);
uint32_t cal_cnt = 2 * 4;
TPipe tpipe;
TBuf<TPosition::VECCALC> ybuf, tmp;
tpipe.InitBuffer(ybuf, sizeof(half) * a * b);
tpipe.InitBuffer(tmp, 8 * 1024);
auto l_y = ybuf.Get<half>();
auto l_tmp = tmp.Get<uint8_t>();
GmToUb(l_y, y, a * b);
Subs<half>(l_y, constant_x, constant_y);
UbToGm(y, l_y, a * b);
};
int a = 2, b = 4;
auto *y = (half*)AscendC::GmAlloc(sizeof(half) * a * b);
half expect[a][b];
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
expect[i][j] = (double)(2.0 - (1.0));
}
}
AscendC::SetKernelMode(KernelMode::AIV_MODE);
ICPU_RUN_KF(kernel, 1, a, b, y);
int diff_count = 0;
for (int i = 0; i < a; i++) {
for (int j = 0; j < b; j++) {
auto diff = (double)(y[i*b + j] - expect[i][j]);
if (diff < -1e-5 || diff > 1e-5) {
diff_count++;
}
}
}
EXPECT_EQ(diff_count, 0);
}