FazBrowse GitHub Viewer | Trending |
URL:
| Home
Tools: [Download Repo ZIP]   [Original HTTPS Page]

src: improve textEncoder encode performance with simdutf · nodejs/node@e62608b · GitHub

/ node Public

Commit e62608b

Browse files
mertcanaltin
authored and
aduh95
committed
src: improve textEncoder encode performance with simdutf
PR-URL: #61496 Refs: cloudflare/workerd#5448 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Daniel Lemire <daniel@lemire.me>
1 parent 915d105 commit e62608b

1 file changed

Lines changed: 84 additions & 11 deletions

File tree

‎src/encoding_binding.cc‎

Lines changed: 84 additions & 11 deletions
Original file line numberDiff line numberDiff line change
@@ -312,11 +312,12 @@ void BindingData::EncodeUtf8String(const FunctionCallbackInfo<Value>& args) {
312312
CHECK_GE(args.Length(), 1);
313313
CHECK(args[0]->IsString());
314314

315-
Local<String> str = args[0].As<String>();
316-
size_t length = str->Utf8LengthV2(isolate);
315+
Local<String> source = args[0].As<String>();
317316

318-
Local<ArrayBuffer> ab;
319-
{
317+
// For small strings, use the V8 path
318+
static constexpr int kSmallStringThreshold = 32;
319+
if (source->Length() <= kSmallStringThreshold) {
320+
size_t length = source->Utf8LengthV2(isolate);
320321
std::unique_ptr<BackingStore> bs = ArrayBuffer::NewBackingStore(
321322
isolate,
322323
length,
@@ -328,16 +329,88 @@ void BindingData::EncodeUtf8String(const FunctionCallbackInfo<Value>& args) {
328329
return;
329330
}
330331

331-
// We are certain that `data` is sufficiently large
332-
str->WriteUtf8V2(isolate,
333-
static_cast<char*>(bs->Data()),
334-
bs->MaxByteLength(),
335-
String::WriteFlags::kReplaceInvalidUtf8);
332+
source->WriteUtf8V2(isolate,
333+
static_cast<char*>(bs->Data()),
334+
bs->MaxByteLength(),
335+
String::WriteFlags::kReplaceInvalidUtf8);
336+
Local<ArrayBuffer> ab = ArrayBuffer::New(isolate, std::move(bs));
337+
args.GetReturnValue().Set(Uint8Array::New(ab, 0, length));
338+
return;
339+
}
336340

337-
ab = ArrayBuffer::New(isolate, std::move(bs));
341+
size_t length = source->Length();
342+
size_t utf8_length = 0;
343+
bool is_one_byte = source->IsOneByte();
344+
345+
if (is_one_byte) {
346+
// One-byte string (Latin1) - copy to buffer first, then process
347+
MaybeStackBuffer<uint8_t, MAX_SIZE_FOR_STACK_ALLOC> latin1_buffer(length);
348+
source->WriteOneByteV2(isolate, 0, length, latin1_buffer.out());
349+
350+
auto data = reinterpret_cast<const char*>(latin1_buffer.out());
351+
352+
// Check if it's pure ASCII - if so, we can just copy
353+
simdutf::result result = simdutf::validate_ascii_with_errors(data, length);
354+
if (result.error == simdutf::SUCCESS) {
355+
// Pure ASCII - direct copy
356+
std::unique_ptr<BackingStore> bs = ArrayBuffer::NewBackingStore(
357+
isolate, length, BackingStoreInitializationMode::kUninitialized);
358+
CHECK(bs);
359+
memcpy(bs->Data(), data, length);
360+
Local<ArrayBuffer> ab = ArrayBuffer::New(isolate, std::move(bs));
361+
args.GetReturnValue().Set(Uint8Array::New(ab, 0, length));
362+
return;
363+
}
364+
365+
// Latin1 with non-ASCII characters - need conversion
366+
utf8_length = simdutf::utf8_length_from_latin1(data, length);
367+
std::unique_ptr<BackingStore> bs = ArrayBuffer::NewBackingStore(
368+
isolate, utf8_length, BackingStoreInitializationMode::kUninitialized);
369+
CHECK(bs);
370+
[[maybe_unused]] size_t written = simdutf::convert_latin1_to_utf8(
371+
data, length, static_cast<char*>(bs->Data()));
372+
DCHECK_EQ(written, utf8_length);
373+
Local<ArrayBuffer> ab = ArrayBuffer::New(isolate, std::move(bs));
374+
args.GetReturnValue().Set(Uint8Array::New(ab, 0, utf8_length));
375+
return;
376+
}
377+
378+
// Two-byte string (UTF-16) - copy to buffer first
379+
MaybeStackBuffer<uint16_t, MAX_SIZE_FOR_STACK_ALLOC> utf16_buffer(length);
380+
source->WriteV2(isolate, 0, length, utf16_buffer.out());
381+
382+
auto data = reinterpret_cast<char16_t*>(utf16_buffer.out());
383+
384+
// Check for unpaired surrogates
385+
simdutf::result validation_result =
386+
simdutf::validate_utf16_with_errors(data, length);
387+
388+
if (validation_result.error == simdutf::SUCCESS) {
389+
// Valid UTF-16 - use the fast path
390+
utf8_length = simdutf::utf8_length_from_utf16(data, length);
391+
std::unique_ptr<BackingStore> bs = ArrayBuffer::NewBackingStore(
392+
isolate, utf8_length, BackingStoreInitializationMode::kUninitialized);
393+
CHECK(bs);
394+
[[maybe_unused]] size_t written = simdutf::convert_utf16_to_utf8(
395+
data, length, static_cast<char*>(bs->Data()));
396+
DCHECK_EQ(written, utf8_length);
397+
Local<ArrayBuffer> ab = ArrayBuffer::New(isolate, std::move(bs));
398+
args.GetReturnValue().Set(Uint8Array::New(ab, 0, utf8_length));
399+
return;
338400
}
339401

340-
args.GetReturnValue().Set(Uint8Array::New(ab, 0, length));
402+
// Invalid UTF-16 with unpaired surrogates - convert to well-formed in place
403+
simdutf::to_well_formed_utf16(data, length, data);
404+
405+
utf8_length = simdutf::utf8_length_from_utf16(data, length);
406+
std::unique_ptr<BackingStore> bs = ArrayBuffer::NewBackingStore(
407+
isolate, utf8_length, BackingStoreInitializationMode::kUninitialized);
408+
CHECK(bs);
409+
[[maybe_unused]] size_t written = simdutf::convert_utf16_to_utf8(
410+
data, length, static_cast<char*>(bs->Data()));
411+
DCHECK_EQ(written, utf8_length);
412+
Local<ArrayBuffer> ab = ArrayBuffer::New(isolate, std::move(bs));
413+
args.GetReturnValue().Set(Uint8Array::New(ab, 0, utf8_length));
341414
}
342415

343416
// Convert the input into an encoded string

0 commit comments

Comments
 (0)
Back | FazBrowse Home | New Git URL