mirrors/bun.sh
1
0
Fork 0
mirror of https://github.com/oven-sh/bun synced 2026-02-10 19:08:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

faster Buffer.byteLength("latin1")

Browse Source
This commit is contained in:
Jarred Sumner
2023-02-16 17:17:29 -08:00
parent 5007c6b218
commit 56b75dbac3
1 changed files with 28 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
src/string_immutable.zig
View File

@@ -1729,48 +1729,40 @@ pub fn elementLengthLatin1IntoUTF8(comptime Type: type, latin1_: Type) usize {
const latin1_last = latin1.ptr + latin1.len;
if (latin1.ptr != latin1_last) {
const wrapped_len = latin1.len - (latin1.len % ascii_vector_size);
// reference the pointer directly because it improves codegen
var ptr = latin1.ptr;
const latin1_vec_end = ptr + wrapped_len;
while (ptr != latin1_vec_end) {
const vec: AsciiVector = ptr[0..ascii_vector_size].*;
if (@reduce(.Max, vec) > 127) {
const Int = u64;
const size = @sizeOf(Int);
const bytes = [2]Int{
@bitCast(Int, ptr[0..size].*) & 0x8080808080808080,
@bitCast(Int, ptr[size .. 2 * size].*) & 0x8080808080808080,
};
total_non_ascii_count += @popCount(bytes[0]) + @popCount(bytes[1]);
if (comptime Environment.enableSIMD) {
const wrapped_len = latin1.len - (latin1.len % ascii_vector_size);
const latin1_vec_end = ptr + wrapped_len;
while (ptr != latin1_vec_end) {
const vec: AsciiVector = ptr[0..ascii_vector_size].*;
const cmp = vec & @splat(ascii_vector_size, @as(u8, 0x80));
total_non_ascii_count += @reduce(.Add, cmp);
ptr += ascii_vector_size;
}
} else {
while (@ptrToInt(ptr + 8) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u64, ptr[0..8].*) & 0x8080808080808080;
total_non_ascii_count += @popCount(bytes);
ptr += 8;
}
ptr += ascii_vector_size;
}
if (@ptrToInt(ptr + 4) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u32, ptr[0..4].*) & 0x80808080;
total_non_ascii_count += @popCount(bytes);
ptr += 4;
}
if (@ptrToInt(ptr + 8) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u64, ptr[0..8].*) & 0x8080808080808080;
total_non_ascii_count += @popCount(bytes);
ptr += 8;
}
if (@ptrToInt(ptr + 4) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u32, ptr[0..4].*) & 0x80808080;
total_non_ascii_count += @popCount(bytes);
ptr += 4;
}
if (@ptrToInt(ptr + 2) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u16, ptr[0..2].*) & 0x8080;
total_non_ascii_count += @popCount(bytes);
ptr += 2;
if (@ptrToInt(ptr + 2) < @ptrToInt(latin1_last)) {
if (comptime Environment.allow_assert) std.debug.assert(@ptrToInt(ptr) <= @ptrToInt(latin1_last) and @ptrToInt(ptr) >= @ptrToInt(latin1_.ptr));
const bytes = @bitCast(u16, ptr[0..2].*) & 0x8080;
total_non_ascii_count += @popCount(bytes);
ptr += 2;
}
}
while (ptr != latin1_last) {