mirror of
https://github.com/denoland/deno.git
synced 2024-12-25 08:39:09 -05:00
02187966c1
V8's JIT can do a better job knowing the argument count and also enable
fast call path (in future).
This also lets us call async ops without `opAsync`:
```js
const { ops } = Deno.core;
await ops.op_void_async();
```
this patch: 4405286 ops/sec
main: 3508771 ops/sec
|
||
|---|---|---|
| .. | ||
| benches | ||
| Cargo.toml | ||
| js_runtime.rs | ||
| lib.rs | ||
| profiling.rs | ||
| README.md | ||
Benching utility for deno_core op system
Example:
use deno_bench_util::bench_or_profile;
use deno_bench_util::bencher::{benchmark_group, Bencher};
use deno_bench_util::bench_js_sync};
use deno_core::op_sync;
use deno_core::serialize_op_result;
use deno_core::Extension;
use deno_core::JsRuntime;
use deno_core::Op;
use deno_core::OpState;
fn setup() -> Vec<Extension> {
let custom_ext = Extension::builder()
.ops(vec![
("op_nop", |state, _| {
Op::Sync(serialize_op_result(Ok(9), state))
}),
])
.build();
vec![
// deno_{ext}::init(...),
custom_ext,
]
}
fn bench_op_nop(b: &mut Bencher) {
bench_js_sync(b, r#"Deno.core.ops.op_nop();"#, setup);
}
benchmark_group!(benches, bench_op_nop);
bench_or_profile!(benches);