david.venhoff/QEMU-Nyx-fork
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

cutils: Add SSE4 version

Browse Source 
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This commit is contained in:
Paolo Bonzini 2016-09-13 17:04:52 +02:00
parent efad668245
commit 86444f084b
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
util/bufferiszero.c
View File

@ -113,6 +113,13 @@ ACCEL_BUFFER_ZERO(buffer_zero_sse2, 64, __m128i, SSE2_NONZERO)
#endif #endif
#ifdef CONFIG_AVX2_OPT #ifdef CONFIG_AVX2_OPT
#pragma GCC push_options
#pragma GCC target("sse4")
#include <smmintrin.h>
#define SSE4_NONZERO(X) !_mm_testz_si128((X), (X))
ACCEL_BUFFER_ZERO(buffer_zero_sse4, 64, __m128i, SSE4_NONZERO)
#pragma GCC pop_options
#pragma GCC push_options #pragma GCC push_options
#pragma GCC target("avx2") #pragma GCC target("avx2")
#include <immintrin.h> #include <immintrin.h>
@ -182,6 +189,9 @@ static bool select_accel_fn(const void *buf, size_t len)
if (len % 128 == 0 && ibuf % 32 == 0 && (cpuid_cache & CACHE_AVX2)) { if (len % 128 == 0 && ibuf % 32 == 0 && (cpuid_cache & CACHE_AVX2)) {
return buffer_zero_avx2(buf, len); return buffer_zero_avx2(buf, len);
} }
if (len % 64 == 0 && ibuf % 16 == 0 && (cpuid_cache & CACHE_SSE4)) {
return buffer_zero_sse4(buf, len);
}
#endif #endif
if (len % 64 == 0 && ibuf % 16 == 0 && (cpuid_cache & CACHE_SSE2)) { if (len % 64 == 0 && ibuf % 16 == 0 && (cpuid_cache & CACHE_SSE2)) {
return buffer_zero_sse2(buf, len); return buffer_zero_sse2(buf, len);