go/sqlite3
1
0
Fork 0
mirror of https://github.com/ncruces/go-sqlite3.git synced 2026-01-12 05:59:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

Issue #274.

Browse Source
This commit is contained in:
Nuno Cruces
2025-05-07 01:37:52 +01:00
parent d3973b23e3
commit fd165ce724
3 changed files with 651 additions and 317 deletions
Download Patch File Download Diff File Expand all files Collapse all files

BIN
sqlite3/libc/libc.wasm
View File

Binary file not shown.

828
sqlite3/libc/libc.wat
View File

File diff suppressed because it is too large Load Diff

140
sqlite3/libc/string.h
View File

@@ -334,21 +334,17 @@ char *strrchr(const char *s, int c) {
__attribute__((weak))
size_t strspn(const char *s, const char *c) {
#ifndef _REENTRANT
static // Avoid the stack for builds without threads.
#endif
char byteset[UCHAR_MAX + 1];
// Set limit to the largest possible valid v128_t pointer.
// Unsigned modular arithmetic gives the correct result
// unless memory size is zero, in which case all pointers are invalid.
const v128_t *const limit =
(v128_t *)(__builtin_wasm_memory_size(0) * PAGESIZE) - 1;
const v128_t *w = (v128_t *)s;
const char *const a = s;
if (!c[0]) return 0;
if (!c[1]) {
// Set limit to the largest possible valid v128_t pointer.
// Unsigned modular arithmetic gives the correct result
// unless memory size is zero, in which case all pointers are invalid.
const v128_t *const limit =
(v128_t *)(__builtin_wasm_memory_size(0) * PAGESIZE) - 1;
const v128_t *w = (v128_t *)s;
const v128_t wc = wasm_i8x16_splat(*c);
while (w <= limit) {
const v128_t cmp = wasm_i8x16_eq(wasm_v128_load(w), wc);
@@ -361,39 +357,113 @@ size_t strspn(const char *s, const char *c) {
}
// Baseline algorithm.
s = (char *)w;
while (*s == *c) s++;
for (s = (char *)w; *s == *c; s++);
return s - a;
}
memset(byteset, 0, sizeof(byteset));
// Keeping byteset[0] = 0 avoids the next loop needing that check.
while (*c && (byteset[*(unsigned char *)c] = 1)) c++;
#if __OPTIMIZE__ && !__OPTIMIZE_SIZE__
#pragma unroll 4
#endif
while (byteset[*(unsigned char *)s]) s++;
return s - a;
// http://0x80.pl/notesen/2018-10-18-simd-byte-lookup.html
typedef unsigned char u8x16
__attribute__((__vector_size__(16), __aligned__(16)));
u8x16 bitmap07 = {};
u8x16 bitmap8f = {};
for (; *c; c++) {
unsigned lo_nibble = *(unsigned char *)c % 16;
unsigned hi_nibble = *(unsigned char *)c / 16;
bitmap07[lo_nibble] |= 1 << (hi_nibble - 0);
bitmap8f[lo_nibble] |= 1 << (hi_nibble - 8);
// Terminator IS NOT on the bitmap.
}
for (; w <= limit; w++) {
const v128_t lo_nibbles = wasm_v128_load(w) & wasm_u8x16_const_splat(0xf);
const v128_t hi_nibbles = wasm_u8x16_shr(wasm_v128_load(w), 4);
const v128_t bitmask_lookup =
wasm_u8x16_const(1, 2, 4, 8, 16, 32, 64, 128, //
1, 2, 4, 8, 16, 32, 64, 128);
const v128_t bitmask = wasm_i8x16_swizzle(bitmask_lookup, hi_nibbles);
const v128_t bitsets = wasm_v128_bitselect(
wasm_i8x16_swizzle(bitmap07, lo_nibbles),
wasm_i8x16_swizzle(bitmap8f, lo_nibbles),
wasm_i8x16_lt(hi_nibbles, wasm_u8x16_const_splat(8)));
const v128_t cmp = wasm_i8x16_eq(bitsets & bitmask, bitmask);
if (!wasm_i8x16_all_true(cmp)) {
size_t ctz = __builtin_ctz(~wasm_i8x16_bitmask(cmp));
return (char *)w + ctz - s;
}
}
// Baseline algorithm.
for (s = (char *)w;; s++) {
const unsigned lo_nibble = *(unsigned char *)s & 0xf;
const unsigned hi_nibble = *(unsigned char *)s >> 4;
const unsigned bitmask = 1 << (hi_nibble & 0x7);
const unsigned bitset =
hi_nibble < 8 ? bitmap07[lo_nibble] : bitmap8f[lo_nibble];
if ((bitset & bitmask) == 0) return s - a;
}
}
__attribute__((weak))
size_t strcspn(const char *s, const char *c) {
#ifndef _REENTRANT
static // Avoid the stack for builds without threads.
#endif
char byteset[UCHAR_MAX + 1];
const char *const a = s;
if (!c[0] || !c[1]) return __strchrnul(s, *c) - s;
memset(byteset, 0, sizeof(byteset));
// Setting byteset[0] = 1 avoids the next loop needing that check.
while ((byteset[*(unsigned char *)c] = 1) && *c) c++;
#if __OPTIMIZE__ && !__OPTIMIZE_SIZE__
#pragma unroll 4
#endif
while (!byteset[*(unsigned char *)s]) s++;
return s - a;
// Set limit to the largest possible valid v128_t pointer.
// Unsigned modular arithmetic gives the correct result
// unless memory size is zero, in which case all pointers are invalid.
const v128_t *const limit =
(v128_t *)(__builtin_wasm_memory_size(0) * PAGESIZE) - 1;
const v128_t *w = (v128_t *)s;
const char *const a = s;
// http://0x80.pl/notesen/2018-10-18-simd-byte-lookup.html
typedef unsigned char u8x16
__attribute__((__vector_size__(16), __aligned__(16)));
u8x16 bitmap07 = {};
u8x16 bitmap8f = {};
for (;; c++) {
unsigned lo_nibble = *(unsigned char *)c % 16;
unsigned hi_nibble = *(unsigned char *)c / 16;
bitmap07[lo_nibble] |= 1 << (hi_nibble - 0);
bitmap8f[lo_nibble] |= 1 << (hi_nibble - 8);
if (!*c) break; // Terminator IS on the bitmap.
}
for (; w <= limit; w++) {
const v128_t lo_nibbles = wasm_v128_load(w) & wasm_u8x16_const_splat(0xf);
const v128_t hi_nibbles = wasm_u8x16_shr(wasm_v128_load(w), 4);
const v128_t bitmask_lookup =
wasm_u8x16_const(1, 2, 4, 8, 16, 32, 64, 128, //
1, 2, 4, 8, 16, 32, 64, 128);
const v128_t bitmask = wasm_i8x16_swizzle(bitmask_lookup, hi_nibbles);
const v128_t bitsets = wasm_v128_bitselect(
wasm_i8x16_swizzle(bitmap07, lo_nibbles),
wasm_i8x16_swizzle(bitmap8f, lo_nibbles),
wasm_i8x16_lt(hi_nibbles, wasm_u8x16_const_splat(8)));
const v128_t cmp = wasm_i8x16_eq(bitsets & bitmask, bitmask);
if (wasm_v128_any_true(cmp)) {
size_t ctz = __builtin_ctz(wasm_i8x16_bitmask(cmp));
return (char *)w + ctz - s;
}
}
// Baseline algorithm.
for (s = (char *)w;; s++) {
const unsigned lo_nibble = *(unsigned char *)s & 0xf;
const unsigned hi_nibble = *(unsigned char *)s >> 4;
const unsigned bitmask = 1 << (hi_nibble & 0x7);
const unsigned bitset =
hi_nibble < 8 ? bitmap07[lo_nibble] : bitmap8f[lo_nibble];
if (bitset & bitmask) return s - a;
}
}
// Given the above SIMD implementations,