yuzu/xbyak
1
0
Fork 0
mirror of https://github.com/herumi/xbyak synced 2024-11-20 16:06:14 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

use xed 2024.11.04 and tests were reinstated for AVX10.2 rev. 2

Browse source
This commit is contained in:
MITSUNARI Shigeo 2024-11-11 15:05:12 +09:00
parent a05a63f870
commit c6ecb77823
3 changed files with 17 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
.github/workflows/main.yml vendored
View file

@ -19,11 +19,14 @@ jobs:
steps:
- uses: actions/checkout@v4
- run: apt -y update
- run: apt -y install g++-multilib libboost-dev make nasm yasm wget xz-utils python3
- run: apt -y install g++-multilib libboost-dev make nasm yasm wget python3 #xz-utils
- run: make test
- run: make -C sample CXXFLAGS="-DXBYAK_NO_EXCEPTION"
- run: |
cd test
wget https://downloadmirror.intel.com/831748/sde-external-9.44.0-2024-08-22-lin.tar.xz
tar xvf sde-external-9.44.0-2024-08-22-lin.tar.xz
env XED=sde-external-9.44.0-2024-08-22-lin/xed64 make xed_test
#wget https://downloadmirror.intel.com/831748/sde-external-9.44.0-2024-08-22-lin.tar.xz
#tar xvf sde-external-9.44.0-2024-08-22-lin.tar.xz
wget https://github.com/herumi/xed-bin/raw/refs/heads/main/xed
chmod +x ./xed
./xed -version
env XED=./xed make xed_test

3
test/Makefile
View file

@ -60,8 +60,7 @@ apx: apx.cpp $(XBYAK_INC)
avx10_test: avx10_test.cpp $(XBYAK_INC)
$(CXX) $(CFLAGS) avx10_test.cpp -o $@ -DXBYAK64
#TEST_FILES=old.txt new-ymm.txt bf16.txt comp.txt misc.txt convert.txt minmax.txt saturation.txt
TEST_FILES=old.txt new-ymm.txt bf16.txt misc.txt convert.txt minmax.txt saturation.txt
TEST_FILES=old.txt new-ymm.txt bf16.txt comp.txt misc.txt convert.txt minmax.txt saturation.txt
xed_test:
@set -e; \
for target in $(addprefix avx10/, $(TEST_FILES)); do \

18
test/avx10/bf16.txt
View file

@ -113,17 +113,17 @@ vfpclasspbf16(k7|k5, zword_b[rax+128], 13);
vcomsbf16(xm2, xm3);
vcomsbf16(xm2, ptr[rax+128]);
//vgetexppbf16(xm1|k3, xmm2);
//vgetexppbf16(xm1|k3, ptr[rax+128]);
//vgetexppbf16(xm1|k3, ptr_b[rax+128]);
vgetexppbf16(xm1|k3, xmm2);
vgetexppbf16(xm1|k3, ptr[rax+128]);
vgetexppbf16(xm1|k3, ptr_b[rax+128]);
//vgetexppbf16(ym1|k3, ymm2);
//vgetexppbf16(ym1|k3, ptr[rax+128]);
//vgetexppbf16(ym1|k3, ptr_b[rax+128]);
vgetexppbf16(ym1|k3, ymm2);
vgetexppbf16(ym1|k3, ptr[rax+128]);
vgetexppbf16(ym1|k3, ptr_b[rax+128]);
//vgetexppbf16(zm1|k3, zmm2);
//vgetexppbf16(zm1|k3, ptr[rax+128]);
//vgetexppbf16(zm1|k3, ptr_b[rax+128]);
vgetexppbf16(zm1|k3, zmm2);
vgetexppbf16(zm1|k3, ptr[rax+128]);
vgetexppbf16(zm1|k3, ptr_b[rax+128]);
vgetmantpbf16(xm1|k3, xmm2, 3);
vgetmantpbf16(xm1|k3, ptr[rax+128], 5);