externals: Update xbyak to v7.05

Merge commit 'fdf626b74f'
This commit is contained in:
Merry 2024-01-30 00:36:49 +00:00
commit 213fe7a452
62 changed files with 5179 additions and 2059 deletions

View file

@ -30,7 +30,7 @@ else
endif
ifeq ($(BIT),64)
TARGET += test64 bf64 memfunc64 test_util64 jmp_table64
TARGET += test64 bf64 memfunc64 test_util64 jmp_table64 zero_upper ccmp no_flags
ifeq ($(BOOST_EXIST),1)
TARGET += calc64 #calc2_64
endif
@ -103,6 +103,18 @@ profiler: profiler.cpp ../xbyak/xbyak_util.h
$(CXX) $(CFLAGS) profiler.cpp -o $@
profiler-vtune: profiler.cpp ../xbyak/xbyak_util.h
$(CXX) $(CFLAGS) profiler.cpp -o $@ -DXBYAK_USE_VTUNE -I /opt/intel/vtune_amplifier/include/ -L /opt/intel/vtune_amplifier/lib64 -ljitprofiling -ldl
zero_upper: zero_upper.cpp $(XBYAK_INC)
$(CXX) $(CFLAGS) zero_upper.cpp -o $@
test_zero_upper: zero_upper
sde -future -- ./zero_upper
ccmp: ccmp.cpp $(XBYAK_INC)
$(CXX) $(CFLAGS) ccmp.cpp -o $@
test_ccmp: ccmp
sde -future -- ./ccmp
no_flags: no_flags.cpp $(XBYAK_INC)
$(CXX) $(CFLAGS) no_flags.cpp -o $@
test_no_flags: no_flags
sde -future -- ./no_flags
clean:
rm -rf $(TARGET) profiler profiler-vtune
@ -122,7 +134,7 @@ toyvm : toyvm.cpp $(XBYAK_INC)
static_buf: static_buf.cpp $(XBYAK_INC)
static_buf64: static_buf.cpp $(XBYAK_INC)
test_util : test_util.cpp $(XBYAK_INC) ../xbyak/xbyak_util.h
test_util2 : test_util.cpp $(XBYAK_INC) ../xbyak/xbyak_util.h
test_util64 : test_util.cpp $(XBYAK_INC) ../xbyak/xbyak_util.h
jmp_table: jmp_table.cpp $(XBYAK_INC)
jmp_table64: jmp_table.cpp $(XBYAK_INC)
memfd: memfd.cpp $(XBYAK_INC)

68
externals/xbyak/sample/ccmp.cpp vendored Normal file
View file

@ -0,0 +1,68 @@
/*
An example of ccmp
> g++ ccmp.cpp -I ../xbyak
> sde -future -- ./a.out
*/
#include <stdio.h>
#include <xbyak/xbyak.h>
#include <xbyak/xbyak_util.h>
using namespace Xbyak;
struct Code1 : Xbyak::CodeGenerator {
Code1()
{
Xbyak::util::StackFrame sf(this, 2);
const auto& p1 = sf.p[0];
const auto& p2 = sf.p[1];
int dfv = 0;
cmp(p1, 3);
ctesta(p2, 1, dfv); // eflags = (p1 > 3) ? ((p2 & 1) == 0) : dfv;
setz(al|T_zu);
}
};
struct Code2 : Xbyak::CodeGenerator {
Code2()
{
Xbyak::util::StackFrame sf(this, 3);
const auto& p1 = sf.p[0];
const auto& p2 = sf.p[1];
const auto& p3 = sf.p[2];
int dfv = 0;
cmp(p1, 1);
ccmpe(p2, 2, dfv); // eflags = p1==1 ? p2==2 : dfv;
ccmpe(p3, 3, dfv); // eflags = (p1==1 && p2==2) ? p3==3 : dfv;
setz(al|T_zu); // p1==1 && p2==2 && p3==3
}
};
int main()
try
{
{
puts("(p1 > 3) && ((p2 & 1) == 0)");
Code1 c;
auto f = c.getCode<int (*)(int, int)>();
for (int p1 = 2; p1 < 5; p1++) {
for (int p2 = 0; p2 < 3; p2++) {
printf("p1=%d p2=%d ret=%d (%d)\n", p1, p2, f(p1, p2), p1 > 3 && ((p2&1) == 0));
}
}
}
{
puts("p1 == 1 && p2 == 2 && p3 == 3");
Code2 c;
auto f = c.getCode<int (*)(int, int, int)>();
for (int p1 = 0; p1 < 3; p1++) {
for (int p2 = 1; p2 < 4; p2++) {
for (int p3 = 2; p3 < 5; p3++) {
printf("p1=%d p2=%d p3=%d ret=%d (%d)\n", p1, p2, p3, f(p1, p2, p3), p1==1 && p2==2 && p3==3);
}
}
}
}
} catch (std::exception& e) {
printf("ERR %s\n", e.what());
}

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx_vnni waitpkg clflushopt cldemote movdiri movdir64b
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b serialize aeskle wide_kl keylocker keylocker_wide

2
externals/xbyak/sample/cpuid/arl.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize avx_vnni_int8 avx_ne_convert avx_ifma cmpccxadd sha512 sm3 sm4 avx_vnni_int16 aeskle wide_kl keylocker keylocker_wide

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl avx512_vnni clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl avx512_vnni clflushopt clwb

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi clflushopt

View file

@ -1,11 +1,25 @@
#!/bin/bash
UPDATE=0
if [ $# -eq 1 ]; then
UPDATE=1
fi
if [ $UPDATE == 1 ]; then
echo "UPDATE"
fi
make -C ../ test_util64
cpus=(p4p mrm pnr nhm wsm snb ivb hsw bdw slt slm glm glp tnt skl cnl icl skx clx cpx icx knl knm tgl adl spr)
cpus=(p4p mrm pnr nhm wsm snb ivb hsw bdw slt slm glm glp tnt skl cnl icl skx clx cpx icx knl knm tgl adl mtl rpl spr gnr srf grr arl lnl)
for cpu in ${cpus[@]} ; do
echo $cpu
~/bin/sde -$cpu -- ../test_util64 -cpuid > tmp.txt
diff tmp.txt $cpu.txt
if [ $UPDATE == 1 ]; then
~/bin/sde -$cpu -- ../test_util64 -cpuid > $cpu.txt
else
~/bin/sde -$cpu -- ../test_util64 -cpuid > tmp.txt
diff $cpu.txt tmp.txt
fi
done

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl avx512_vnni avx512_bf16 clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl avx512_vnni avx512_bf16 clflushopt clwb

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq enh_rep rdrand rdseed smap sha movbe clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq enh_rep rdrand rdseed smap sha movbe clflushopt

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq enh_rep rdrand rdseed smap sha movbe clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq enh_rep rdrand rdseed smap sha movbe clflushopt

2
externals/xbyak/sample/cpuid/gnr.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_bf16 amx(tile) amx(int8) amx(bf16) avx_vnni avx512_fp16 waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize amx_fp16 prefetchiti avx10

2
externals/xbyak/sample/cpuid/grr.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize avx_vnni_int8 avx_ne_convert avx_ifma rao-int cmpccxadd aeskle wide_kl keylocker keylocker_wide

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt enh_rep rdrand f16c movbe
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt enh_rep rdrand f16c movbe

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq clflushopt clwb

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq clflushopt clwb

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx enh_rep rdrand f16c
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx enh_rep rdrand f16c

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed prefetchwt1 f16c movbe avx512f avx512pf avx512er avx512cd
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed prefetchwt1 f16c movbe avx512f avx512pf avx512er avx512cd

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed prefetchwt1 f16c movbe avx512f avx512pf avx512er avx512cd avx512_4vnniw avx512_4fmaps avx512_vpopcntdq
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed prefetchwt1 f16c movbe avx512f avx512pf avx512er avx512cd avx512_4vnniw avx512_4fmaps avx512_vpopcntdq

2
externals/xbyak/sample/cpuid/lnl.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize avx_vnni_int8 avx_ne_convert avx_ifma cmpccxadd sha512 sm3 sm4 avx_vnni_int16 aeskle wide_kl keylocker keylocker_wide

2
externals/xbyak/sample/cpuid/mtl.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b serialize aeskle wide_kl keylocker keylocker_wide

2
externals/xbyak/sample/cpuid/rpl.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b serialize aeskle wide_kl keylocker keylocker_wide

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe clflushopt

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap f16c movbe avx512f avx512dq avx512cd avx512bw avx512vl clflushopt clwb

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_bf16 avx512_vp2intersect amx(tile) amx(int8) amx(bf16) avx_vnni avx512_fp16 waitpkg clflushopt cldemote movdiri movdir64b
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_bf16 amx(tile) amx(int8) amx(bf16) avx_vnni avx512_fp16 waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize

2
externals/xbyak/sample/cpuid/srf.txt vendored Normal file
View file

@ -0,0 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt enh_rep rdrand adx rdseed smap sha f16c movbe gfni vaes vpclmulqdq avx_vnni waitpkg clflushopt cldemote clwb movdiri movdir64b uintr serialize avx_vnni_int8 avx_ne_convert avx_ifma cmpccxadd aeskle wide_kl keylocker keylocker_wide

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_vp2intersect clflushopt movdiri movdir64b
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_vp2intersect clflushopt clwb movdiri movdir64b aeskle wide_kl keylocker keylocker_wide

View file

@ -1,2 +0,0 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq avx fma avx2 bmi1 bmi2 lzcnt prefetchw enh_rep rdrand adx rdseed smap sha f16c movbe avx512f avx512dq avx512_ifma avx512cd avx512bw avx512vl avx512_vbmi avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq avx512_bf16 avx512_vp2intersect amx(tile) amx(int8) amx(bf16) avx_vnni avx512_fp16 waitpkg clflushopt cldemote movdiri movdir64b

View file

@ -1,2 +1,2 @@
vendor intel
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp osxsave(xgetvb) pclmulqdq enh_rep rdrand rdseed smap sha movbe clflushopt
mmx mmx2 cmov sse sse2 sse3 ssse3 sse41 sse42 popcnt aesni rdtscp xsave(xgetvb) osxsave pclmulqdq enh_rep rdrand rdseed smap sha movbe gfni clflushopt cldemote clwb

View file

@ -1,10 +0,0 @@
#!/bin/bash
make -C ../ test_util64
cpus=(p4p mrm pnr nhm wsm snb ivb hsw bdw slt slm glm glp tnt skl cnl icl skx clx cpx icx knl knm tgl adl spr)
for cpu in ${cpus[@]} ; do
echo $cpu
~/bin/sde -$cpu -- ../test_util64 -cpuid > $cpu.txt
done

25
externals/xbyak/sample/no_flags.cpp vendored Normal file
View file

@ -0,0 +1,25 @@
#include <stdio.h>
#include <xbyak/xbyak.h>
struct Code : Xbyak::CodeGenerator {
Code(bool nf) {
xor_(eax, eax); // CF = 0
mov(eax, -1);
if (nf) {
puts("no flags (with T_nf)");
add(eax|T_nf, eax, 1); // does not change CF
} else {
puts("change flags (without T_nf)");
add(eax, eax, 1); // CF = 1
}
adc(eax, 0); // eax = CF ? 1 : 0
ret();
}
};
int main() {
for (int i = 0; i < 2; i++) {
Code c(i);
printf("i=%d ret=%d\n", i, c.getCode<int(*)()>()());
}
}

View file

@ -31,12 +31,14 @@ void putCPUinfo(bool onlyCpuidFeature)
{ Cpu::tSSSE3, "ssse3" },
{ Cpu::tSSE41, "sse41" },
{ Cpu::tSSE42, "sse42" },
{ Cpu::tSSE4a, "sse4a" },
{ Cpu::tPOPCNT, "popcnt" },
{ Cpu::t3DN, "3dn" },
{ Cpu::tE3DN, "e3dn" },
{ Cpu::tAESNI, "aesni" },
{ Cpu::tRDTSCP, "rdtscp" },
{ Cpu::tOSXSAVE, "osxsave(xgetvb)" },
{ Cpu::tXSAVE, "xsave(xgetvb)" },
{ Cpu::tOSXSAVE, "osxsave" },
{ Cpu::tPCLMULQDQ, "pclmulqdq" },
{ Cpu::tAVX, "avx" },
{ Cpu::tFMA, "fma" },
@ -86,8 +88,11 @@ void putCPUinfo(bool onlyCpuidFeature)
{ Cpu::tWAITPKG, "waitpkg" },
{ Cpu::tCLFLUSHOPT, "clflushopt" },
{ Cpu::tCLDEMOTE, "cldemote" },
{ Cpu::tCLWB, "clwb" },
{ Cpu::tMOVDIRI, "movdiri" },
{ Cpu::tMOVDIR64B, "movdir64b" },
{ Cpu::tUINTR, "uintr" },
{ Cpu::tSERIALIZE, "serialize" },
{ Cpu::tCLZERO, "clzero" },
{ Cpu::tAMX_FP16, "amx_fp16" },
{ Cpu::tAVX_VNNI_INT8, "avx_vnni_int8" },
@ -96,12 +101,25 @@ void putCPUinfo(bool onlyCpuidFeature)
{ Cpu::tRAO_INT, "rao-int" },
{ Cpu::tCMPCCXADD, "cmpccxadd" },
{ Cpu::tPREFETCHITI, "prefetchiti" },
{ Cpu::tSHA512, "sha512" },
{ Cpu::tSM3, "sm3" },
{ Cpu::tSM4, "sm4" },
{ Cpu::tAVX_VNNI_INT16, "avx_vnni_int16" },
{ Cpu::tAPX_F, "apx_f" },
{ Cpu::tAVX10, "avx10" },
{ Cpu::tAESKLE, "aeskle" },
{ Cpu::tWIDE_KL, "wide_kl" },
{ Cpu::tKEYLOCKER, "keylocker" },
{ Cpu::tKEYLOCKER_WIDE, "keylocker_wide" },
};
for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
if (cpu.has(tbl[i].type)) printf(" %s", tbl[i].str);
}
printf("\n");
if (onlyCpuidFeature) return;
if (cpu.has(Cpu::tAVX10)) {
printf("AVX10 version %d\n", cpu.getAVX10version());
}
if (cpu.has(Cpu::tPOPCNT)) {
const int n = 0x12345678; // bitcount = 13
const int ok = 13;
@ -127,7 +145,6 @@ void putCPUinfo(bool onlyCpuidFeature)
Core i7-3930K 6 2D
*/
cpu.putFamily();
if (!cpu.has(Cpu::tINTEL)) return;
for (unsigned int i = 0; i < cpu.getDataCacheLevels(); i++) {
printf("cache level=%u data cache size=%u cores sharing data cache=%u\n", i, cpu.getDataCacheSize(i), cpu.getCoresSharingDataCache(i));
}

48
externals/xbyak/sample/zero_upper.cpp vendored Normal file
View file

@ -0,0 +1,48 @@
/*
An example of T_zu (zero upper) flag
> g++ zero_upper.cpp -I ../xbyak
> sde -future -- ./a.out
*/
#include <stdio.h>
#include <xbyak/xbyak.h>
using namespace Xbyak;
struct Code : Xbyak::CodeGenerator {
Code(int mode)
{
mov(eax, 0x12345678);
cmp(eax, eax); // ZF=1
switch (mode) {
case 0: // imul
puts("imul");
imul(ax,ax, 0x1234);
break;
case 1: // imul+zu
puts("imul+zu");
imul(ax|T_zu, ax, 0x1234);
break;
case 2: // setz
puts("setz");
setz(al);
break;
case 3: // setz+zu
puts("setz+zu");
setz(al|T_zu);
break;
}
ret();
}
};
int main()
try
{
for (int mode = 0; mode < 4; mode++) {
Code c(mode);
auto f = c.getCode<int (*)()>();
printf("ret=%08x\n", f());
}
} catch (std::exception& e) {
printf("ERR %s\n", e.what());
}