1
0
mirror of https://github.com/opencv/opencv.git synced 2026-07-29 23:33:05 +04:00

Merge pull request #9753 from tomoaki0705:universalMatmul

* add accuracy test and performance check for matmul
  * add performance tests for transform and dotProduct
  * add test Core_TransformLargeTest for 8u version of transform

* remove raw SSE2/NEON implementation from matmul.cpp
  * use universal intrinsic instead of raw intrinsic
  * remove unused templated function
  * add v_matmuladd which multiply 3x3 matrix and add 3x1 vector
  * add v_rotate_left/right in universal intrinsic
  * suppress intrinsic on some function and platform
  * add pure SW implementation of new universal intrinsics
  * add test for new universal intrinsics

* core: prevent memory access after the end of buffer

* fix perf tests
This commit is contained in:
Tomoaki Teshima
2017-11-20 21:56:53 +09:00
committed by Alexander Alekhin
parent 2674c6b5e0
commit 3cbe60cca2
9 changed files with 563 additions and 374 deletions
+8
View File
@@ -33,6 +33,7 @@ TEST(hal_intrin, uint8x16) {
.test_pack_u<1>().test_pack_u<2>().test_pack_u<3>().test_pack_u<8>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<8>().test_extract<15>()
.test_rotate<0>().test_rotate<1>().test_rotate<8>().test_rotate<15>()
;
}
@@ -54,6 +55,7 @@ TEST(hal_intrin, int8x16) {
.test_pack<1>().test_pack<2>().test_pack<3>().test_pack<8>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<8>().test_extract<15>()
.test_rotate<0>().test_rotate<1>().test_rotate<8>().test_rotate<15>()
;
}
@@ -81,6 +83,7 @@ TEST(hal_intrin, uint16x8) {
.test_pack_u<1>().test_pack_u<2>().test_pack_u<7>().test_pack_u<16>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<4>().test_extract<7>()
.test_rotate<0>().test_rotate<1>().test_rotate<4>().test_rotate<7>()
;
}
@@ -107,6 +110,7 @@ TEST(hal_intrin, int16x8) {
.test_pack<1>().test_pack<2>().test_pack<7>().test_pack<16>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<4>().test_extract<7>()
.test_rotate<0>().test_rotate<1>().test_rotate<4>().test_rotate<7>()
;
}
@@ -132,6 +136,7 @@ TEST(hal_intrin, uint32x4) {
.test_pack<1>().test_pack<2>().test_pack<15>().test_pack<32>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<2>().test_extract<3>()
.test_rotate<0>().test_rotate<1>().test_rotate<2>().test_rotate<3>()
.test_transpose()
;
}
@@ -155,6 +160,7 @@ TEST(hal_intrin, int32x4) {
.test_pack<1>().test_pack<2>().test_pack<15>().test_pack<32>()
.test_unpack()
.test_extract<0>().test_extract<1>().test_extract<2>().test_extract<3>()
.test_rotate<0>().test_rotate<1>().test_rotate<2>().test_rotate<3>()
.test_float_cvt32()
.test_float_cvt64()
.test_transpose()
@@ -170,6 +176,7 @@ TEST(hal_intrin, uint64x2) {
.test_shift<1>().test_shift<8>()
.test_logic()
.test_extract<0>().test_extract<1>()
.test_rotate<0>().test_rotate<1>()
;
}
@@ -180,6 +187,7 @@ TEST(hal_intrin, int64x2) {
.test_shift<1>().test_shift<8>()
.test_logic()
.test_extract<0>().test_extract<1>()
.test_rotate<0>().test_rotate<1>()
;
}