Skip to content

compress: vectorize ZSTD_count() with SSE2 #4612

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

Open
rmilkowski wants to merge 2 commits into
base: dev
Choose a base branch
from
+19 −0
Open

compress: vectorize ZSTD_count() with SSE2 #4612

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2fb0de7
compress: vectorize ZSTD_count() with SSE2
rmilkowski Jan 15, 2026
5fb1974
compress: fix SSE2 cast-align warnings
rmilkowski Mar 6, 2026
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
19 changes: 19 additions & 0 deletions lib/compress/zstd_compress_internal.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -855,11 +855,30 @@ MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* co
{
const BYTE* const pStart = pIn;
const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1);
#if defined(ZSTD_ARCH_X86_SSE2)
const BYTE* const pInLimit16 = pInLimit - (sizeof(__m128i)-1);
#endif

if (pIn < pInLoopLimit) {
{ size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
if (diff) return ZSTD_NbCommonBytes(diff); }
pIn+=sizeof(size_t); pMatch+=sizeof(size_t);
#if defined(ZSTD_ARCH_X86_SSE2)
if ((size_t)(pInLimit - pIn) >= 32) {
while (pIn < pInLimit16) {
__m128i const matchVec = _mm_loadu_si128((const __m128i*)(const void*)pMatch);
__m128i const inVec = _mm_loadu_si128((const __m128i*)(const void*)pIn);
U32 const matchMask = (U32)_mm_movemask_epi8(_mm_cmpeq_epi8(matchVec, inVec));
if (matchMask != 0xFFFF) {
U32 const diffMask = ~matchMask & 0xFFFF;
pIn += ZSTD_countTrailingZeros32(diffMask);
return (size_t)(pIn - pStart);
}
pIn += sizeof(__m128i);
pMatch += sizeof(__m128i);
}
}
#endif
while (pIn < pInLoopLimit) {
size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn);
if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; }
Expand Down