github_mirrors/vectorscan
3
0
Fork 1
mirror of https://github.com/VectorCamp/vectorscan.git synced 2025-06-28 16:41:01 +03:00
Code Issues Packages Projects Releases Wiki Activity

optimize case mask AND out of the loop

Browse Source
This commit is contained in:
Konstantinos Margaritis 2021-02-10 13:29:45 +02:00
parent be66cdb51d
commit d3e03ed88a
1 changed files with 12 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
src/hwlm/noodle_engine_sse.c
View File

@ -49,12 +49,8 @@ hwlm_error_t scanSingleShort(const struct noodTable *n, const u8 *buf,
if (!l) { if (!l) {
return HWLM_SUCCESS; return HWLM_SUCCESS;
} }
m128 v = zeroes128(); m128 mask128 = noCase ? caseMask : ones128();
// we don't have a clever way of doing this move yet m128 v = and128(load128(d), mask128);
memcpy(&v, d, l);
if (noCase) {
v = and128(v, caseMask);
}
// mask out where we can't match // mask out where we can't match
u32 mask = (0xFFFF >> (16 - l)); u32 mask = (0xFFFF >> (16 - l));
@ -76,11 +72,8 @@ hwlm_error_t scanSingleUnaligned(const struct noodTable *n, const u8 *buf,
DEBUG_PRINTF("start %zu end %zu offset %zu\n", start, end, offset); DEBUG_PRINTF("start %zu end %zu offset %zu\n", start, end, offset);
const size_t l = end - start; const size_t l = end - start;
m128 v = loadu128(d); m128 mask128 = noCase ? caseMask : ones128();
m128 v = and128(loadu128(d), mask128);
if (noCase) {
v = and128(v, caseMask);
}
u32 buf_off = start - offset; u32 buf_off = start - offset;
u32 mask = ((1 << l) - 1) << buf_off; u32 mask = ((1 << l) - 1) << buf_off;
@ -109,11 +102,8 @@ hwlm_error_t scanDoubleShort(const struct noodTable *n, const u8 *buf,
assert(l <= 32); assert(l <= 32);
DEBUG_PRINTF("d %zu\n", d - buf); DEBUG_PRINTF("d %zu\n", d - buf);
m128 v = zeroes128(); m128 mask128 = noCase ? caseMask : ones128();
memcpy(&v, d, l); m128 v = and128(load128(d), mask128);
if (noCase) {
v = and128(v, caseMask);
}
u32 z = movemask128(and128(lshiftbyte_m128(eq128(mask1, v), 1), u32 z = movemask128(and128(lshiftbyte_m128(eq128(mask1, v), 1),
eq128(mask2, v))); eq128(mask2, v)));
@ -137,11 +127,8 @@ hwlm_error_t scanDoubleUnaligned(const struct noodTable *n, const u8 *buf,
DEBUG_PRINTF("start %zu end %zu offset %zu\n", start, end, offset); DEBUG_PRINTF("start %zu end %zu offset %zu\n", start, end, offset);
size_t l = end - start; size_t l = end - start;
m128 v = loadu128(d); m128 mask128 = noCase ? caseMask : ones128();
m128 v = and128(loadu128(d), mask128);
if (noCase) {
v = and128(v, caseMask);
}
u32 z = movemask128(and128(lshiftbyte_m128(eq128(mask1, v), 1), u32 z = movemask128(and128(lshiftbyte_m128(eq128(mask1, v), 1),
eq128(mask2, v))); eq128(mask2, v)));
@ -164,9 +151,10 @@ hwlm_error_t scanSingleFast(const struct noodTable *n, const u8 *buf,
size_t end) { size_t end) {
const u8 *d = buf + start, *e = buf + end; const u8 *d = buf + start, *e = buf + end;
assert(d < e); assert(d < e);
m128 mask128 = noCase ? caseMask : ones128();
for (; d < e; d += 16) { for (; d < e; d += 16) {
m128 v = noCase ? and128(load128(d), caseMask) : load128(d); m128 v = and128(load128(d), mask128);
u32 z = movemask128(eq128(mask1, v)); u32 z = movemask128(eq128(mask1, v));
@ -186,9 +174,10 @@ hwlm_error_t scanDoubleFast(const struct noodTable *n, const u8 *buf,
const u8 *d = buf + start, *e = buf + end; const u8 *d = buf + start, *e = buf + end;
assert(d < e); assert(d < e);
m128 lastz1 = zeroes128(); m128 lastz1 = zeroes128();
m128 mask128 = noCase ? caseMask : ones128();
for (; d < e; d += 16) { for (; d < e; d += 16) {
m128 v = noCase ? and128(load128(d), caseMask) : load128(d); m128 v = and128(load128(d), mask128);
m128 z1 = eq128(mask1, v); m128 z1 = eq128(mask1, v);
m128 z2 = eq128(mask2, v); m128 z2 = eq128(mask2, v);
u32 z = movemask128(and128(palignr(z1, lastz1, 15), z2)); u32 z = movemask128(and128(palignr(z1, lastz1, 15), z2));