1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-15 10:04:59 +00:00

LibJS: Stop converting between Object <-> IteratorRecord all the time

This patch makes IteratorRecord an Object. Although it's not exposed to
author code, this does allow us to store it in a VM register.

Now that we can store it in a VM register, we don't need to convert it
back and forth between IteratorRecord and Object when accessing it from
bytecode.

The big win here is avoiding 3 [[Get]] accesses on every iteration step
of for..of loops. There are also a bunch of smaller efficiencies gained.

20% speed-up on this microbenchmark:

    function go(a) {
        for (const p of a) {
        }
    }
    const a = [];
    a.length = 1_000_000;
    go(a);
This commit is contained in:
Andreas Kling 2023-12-07 10:44:41 +01:00
parent 4966c083df
commit 4699c81fc1
23 changed files with 226 additions and 144 deletions

View file

@ -1749,16 +1749,11 @@ Bytecode::CodeGenerationErrorOr<void> YieldExpression::generate_bytecode(Bytecod
// 5. Let iterator be iteratorRecord.[[Iterator]].
auto iterator_register = generator.allocate_register();
auto iterator_identifier = generator.intern_identifier("iterator");
generator.emit_get_by_id(iterator_identifier);
generator.emit<Bytecode::Op::Store>(iterator_register);
generator.emit<Bytecode::Op::GetObjectFromIteratorRecord>(iterator_register, iterator_record_register);
// Cache iteratorRecord.[[NextMethod]] for use in step 7.a.i.
auto next_method_register = generator.allocate_register();
auto next_method_identifier = generator.intern_identifier("next");
generator.emit<Bytecode::Op::Load>(iterator_record_register);
generator.emit_get_by_id(next_method_identifier);
generator.emit<Bytecode::Op::Store>(next_method_register);
generator.emit<Bytecode::Op::GetNextMethodFromIteratorRecord>(next_method_register, iterator_record_register);
// 6. Let received be NormalCompletion(undefined).
// See get_received_completion_type_and_value above.