Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Rewrite Universal Intrinsic code by using new API: Core module. #23980

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Aug 11, 2023
Merged

Rewrite Universal Intrinsic code by using new API: Core module. #23980

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
eccd0c4
Rewrite Universal Intrinsic code by using new API: Core module.
hanliutong Jul 12, 2023
71efd64
Update matmul.simd.hpp and enable v_reduce_sum for f64.
hanliutong Jul 13, 2023
7ddd22d
Rewrite Universal Intrinsic code by using new API
hanliutong Jul 18, 2023
59773b2
Revert dot product for 32s.
hanliutong Aug 2, 2023
4d31f0a
Revert block in lapack and remove unused code.
hanliutong Aug 4, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions modules/core/include/opencv2/core/hal/intrin_rvv_scalable.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -924,6 +924,9 @@ inline scalartype v_reduce_sum(const _Tpvec& a) \
return (scalartype)v_get0(res); \
}
OPENCV_HAL_IMPL_RVV_REDUCE_SUM_FP(v_float32, v_float32, vfloat32m1_t, float, f32, VTraits<v_float32>::vlanes())
#if CV_SIMD_SCALABLE_64F
OPENCV_HAL_IMPL_RVV_REDUCE_SUM_FP(v_float64, v_float64, vfloat64m1_t, float, f64, VTraits<v_float64>::vlanes())
#endif

#define OPENCV_HAL_IMPL_RVV_REDUCE(_Tpvec, func, scalartype, suffix, vl, red) \
inline scalartype v_reduce_##func(const _Tpvec& a) \
Expand Down
52 changes: 26 additions & 26 deletions modules/core/src/arithm.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1332,7 +1332,7 @@ struct InRange_SIMD
}
};

#if CV_SIMD
#if (CV_SIMD || CV_SIMD_SCALABLE)

template <>
struct InRange_SIMD<uchar>
Expand All @@ -1341,15 +1341,15 @@ struct InRange_SIMD<uchar>
uchar * dst, int len) const
{
int x = 0;
const int width = v_uint8::nlanes;
const int width = VTraits<v_uint8>::vlanes();

for (; x <= len - width; x += width)
{
v_uint8 values = vx_load(src1 + x);
v_uint8 low = vx_load(src2 + x);
v_uint8 high = vx_load(src3 + x);

v_store(dst + x, (values >= low) & (high >= values));
v_store(dst + x, v_and(v_ge(values, low), v_ge(high, values)));
}
vx_cleanup();
return x;
Expand All @@ -1363,15 +1363,15 @@ struct InRange_SIMD<schar>
uchar * dst, int len) const
{
int x = 0;
const int width = v_int8::nlanes;
const int width = VTraits<v_int8>::vlanes();

for (; x <= len - width; x += width)
{
v_int8 values = vx_load(src1 + x);
v_int8 low = vx_load(src2 + x);
v_int8 high = vx_load(src3 + x);

v_store((schar*)(dst + x), (values >= low) & (high >= values));
v_store((schar*)(dst + x), v_and(v_ge(values, low), v_ge(high, values)));
}
vx_cleanup();
return x;
Expand All @@ -1385,19 +1385,19 @@ struct InRange_SIMD<ushort>
uchar * dst, int len) const
{
int x = 0;
const int width = v_uint16::nlanes * 2;
const int width = VTraits<v_uint16>::vlanes() * 2;

for (; x <= len - width; x += width)
{
v_uint16 values1 = vx_load(src1 + x);
v_uint16 low1 = vx_load(src2 + x);
v_uint16 high1 = vx_load(src3 + x);

v_uint16 values2 = vx_load(src1 + x + v_uint16::nlanes);
v_uint16 low2 = vx_load(src2 + x + v_uint16::nlanes);
v_uint16 high2 = vx_load(src3 + x + v_uint16::nlanes);
v_uint16 values2 = vx_load(src1 + x + VTraits<v_uint16>::vlanes());
v_uint16 low2 = vx_load(src2 + x + VTraits<v_uint16>::vlanes());
v_uint16 high2 = vx_load(src3 + x + VTraits<v_uint16>::vlanes());

v_store(dst + x, v_pack((values1 >= low1) & (high1 >= values1), (values2 >= low2) & (high2 >= values2)));
v_store(dst + x, v_pack(v_and(v_ge(values1, low1), v_ge(high1, values1)), v_and(v_ge(values2, low2), v_ge(high2, values2))));
}
vx_cleanup();
return x;
Expand All @@ -1411,19 +1411,19 @@ struct InRange_SIMD<short>
uchar * dst, int len) const
{
int x = 0;
const int width = (int)v_int16::nlanes * 2;
const int width = (int)VTraits<v_int16>::vlanes() * 2;

for (; x <= len - width; x += width)
{
v_int16 values1 = vx_load(src1 + x);
v_int16 low1 = vx_load(src2 + x);
v_int16 high1 = vx_load(src3 + x);

v_int16 values2 = vx_load(src1 + x + v_int16::nlanes);
v_int16 low2 = vx_load(src2 + x + v_int16::nlanes);
v_int16 high2 = vx_load(src3 + x + v_int16::nlanes);
v_int16 values2 = vx_load(src1 + x + VTraits<v_int16>::vlanes());
v_int16 low2 = vx_load(src2 + x + VTraits<v_int16>::vlanes());
v_int16 high2 = vx_load(src3 + x + VTraits<v_int16>::vlanes());

v_store((schar*)(dst + x), v_pack((values1 >= low1) & (high1 >= values1), (values2 >= low2) & (high2 >= values2)));
v_store((schar*)(dst + x), v_pack(v_and(v_ge(values1, low1), v_ge(high1, values1)), v_and(v_ge(values2, low2), v_ge(high2, values2))));
}
vx_cleanup();
return x;
Expand All @@ -1437,19 +1437,19 @@ struct InRange_SIMD<int>
uchar * dst, int len) const
{
int x = 0;
const int width = (int)v_int32::nlanes * 2;
const int width = (int)VTraits<v_int32>::vlanes() * 2;

for (; x <= len - width; x += width)
{
v_int32 values1 = vx_load(src1 + x);
v_int32 low1 = vx_load(src2 + x);
v_int32 high1 = vx_load(src3 + x);

v_int32 values2 = vx_load(src1 + x + v_int32::nlanes);
v_int32 low2 = vx_load(src2 + x + v_int32::nlanes);
v_int32 high2 = vx_load(src3 + x + v_int32::nlanes);
v_int32 values2 = vx_load(src1 + x + VTraits<v_int32>::vlanes());
v_int32 low2 = vx_load(src2 + x + VTraits<v_int32>::vlanes());
v_int32 high2 = vx_load(src3 + x + VTraits<v_int32>::vlanes());

v_pack_store(dst + x, v_reinterpret_as_u16(v_pack((values1 >= low1) & (high1 >= values1), (values2 >= low2) & (high2 >= values2))));
v_pack_store(dst + x, v_reinterpret_as_u16(v_pack(v_and(v_ge(values1, low1), v_ge(high1, values1)), v_and(v_ge(values2, low2), v_ge(high2, values2)))));
}
vx_cleanup();
return x;
Expand All @@ -1463,20 +1463,20 @@ struct InRange_SIMD<float>
uchar * dst, int len) const
{
int x = 0;
const int width = (int)v_float32::nlanes * 2;
const int width = (int)VTraits<v_float32>::vlanes() * 2;

for (; x <= len - width; x += width)
{
v_float32 values1 = vx_load(src1 + x);
v_float32 low1 = vx_load(src2 + x);
v_float32 high1 = vx_load(src3 + x);

v_float32 values2 = vx_load(src1 + x + v_float32::nlanes);
v_float32 low2 = vx_load(src2 + x + v_float32::nlanes);
v_float32 high2 = vx_load(src3 + x + v_float32::nlanes);
v_float32 values2 = vx_load(src1 + x + VTraits<v_float32>::vlanes());
v_float32 low2 = vx_load(src2 + x + VTraits<v_float32>::vlanes());
v_float32 high2 = vx_load(src3 + x + VTraits<v_float32>::vlanes());

v_pack_store(dst + x, v_pack(v_reinterpret_as_u32(values1 >= low1) & v_reinterpret_as_u32(high1 >= values1),
v_reinterpret_as_u32(values2 >= low2) & v_reinterpret_as_u32(high2 >= values2)));
v_pack_store(dst + x, v_pack(v_and(v_reinterpret_as_u32(v_ge(values1, low1)), v_reinterpret_as_u32(v_ge(high1, values1))),
v_and(v_reinterpret_as_u32(v_ge(values2, low2)), v_reinterpret_as_u32(v_ge(high2, values2)))));
}
vx_cleanup();
return x;
Expand Down
Loading