mirror of
https://github.com/opencv/opencv.git
synced 2026-07-21 19:33:03 +04:00
Kumataro
1f674dcdb4
Close https://github.com/opencv/opencv/issues/27413 ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake
47 lines
1.3 KiB
C++
47 lines
1.3 KiB
C++
#if (defined __GNUC__ && (defined __arm__ || defined __aarch64__)) || (defined _MSC_VER && (defined _M_ARM64 || defined _M_ARM64EC))
|
|
#include <stdio.h>
|
|
#include "arm_neon.h"
|
|
|
|
/*#if defined __clang__
|
|
#pragma clang attribute push (__attribute__((target("bf16"))), apply_to=function)
|
|
#elif defined GCC
|
|
#pragma GCC push_options
|
|
#pragma GCC target("armv8.2-a", "bf16")
|
|
#endif*/
|
|
bfloat16x8_t vld1q_as_bf16(const float* src)
|
|
{
|
|
float32x4_t s0 = vld1q_f32(src), s1 = vld1q_f32(src + 4);
|
|
return vcombine_bf16(vcvt_bf16_f32(s0), vcvt_bf16_f32(s1));
|
|
}
|
|
|
|
void vprintreg(const char* name, const float32x4_t& r)
|
|
{
|
|
float data[4];
|
|
vst1q_f32(data, r);
|
|
printf("%s: (%.2f, %.2f, %.2f, %.2f)\n",
|
|
name, data[0], data[1], data[2], data[3]);
|
|
}
|
|
|
|
void test()
|
|
{
|
|
const float src1[] = { 1.f, 2.f, 3.f, 4.f, 5.f, 6.f, 7.f, 8.f };
|
|
const float src2[] = { 1.f, 3.f, 6.f, 10.f, 15.f, 21.f, 28.f, 36.f };
|
|
bfloat16x8_t s1 = vld1q_as_bf16(src1), s2 = vld1q_as_bf16(src2);
|
|
float32x4_t d = vbfdotq_f32(vdupq_n_f32(0.f), s1, s2);
|
|
vprintreg("(s1[0]*s2[0] + s1[1]*s2[1], ... s1[6]*s2[6] + s1[7]*s2[7])", d);
|
|
}
|
|
/*#if defined __clang__
|
|
#pragma clang attribute pop
|
|
#elif defined GCC
|
|
#pragma GCC pop_options
|
|
#endif*/
|
|
#else
|
|
#error "BF16 is not supported"
|
|
#endif
|
|
|
|
int main()
|
|
{
|
|
test();
|
|
return 0;
|
|
}
|