Update zstd_lazy.c
switch to unaligned load as I don't know if buffer will always be aligned to 32 bytes, and compilers aside from MSVC might actually use aligned loads
This commit is contained in:
@@ -928,7 +928,7 @@ typedef struct {
|
||||
|
||||
static ZSTD_Vec256 ZSTD_Vec256_read(const void* const ptr) {
|
||||
ZSTD_Vec256 v;
|
||||
v.v = _mm256_load_si256((const __m256i*)ptr);
|
||||
v.v = _mm256_loadu_si256((const __m256i*)ptr);
|
||||
return v;
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user