- From: Takeshi Yoshino <notifications@github.com>
- Date: Mon, 13 Apr 2015 01:33:40 -0700
- To: whatwg/streams <streams@noreply.github.com>
Received on Monday, 13 April 2015 08:34:09 UTC
@wanderview @domenic In resolve() definition in benchmark.js, setTimeout(fn, 0) is called for each operation (function added by suite.add()). It seems Chrome always has 4-5ms delay for it, and therefore it's hiding actual performance difference (seeing ~200 ops for everything). This setTimeout can be omitted as far as the number of cycle is small. I commented it out (and implemented URLSearchParams) and ran the test on Chrome 41.0.2272.118 (64-bit) on Linux. ``` Testing 1 chunks of 1 bytes per operation. sync x 35,994 ops/sec ±1.99% (84 runs sampled) Testing 32 chunks of 1 bytes per operation. sync x 28,294 ops/sec ±0.98% (83 runs sampled) Testing 1 chunks of 1 bytes per operation. promise x 21,132 ops/sec ±1.94% (81 runs sampled) Testing 32 chunks of 1 bytes per operation. promise x 2,628 ops/sec ±0.86% (87 runs sampled) ``` --- Reply to this email directly or view it on GitHub: https://github.com/whatwg/streams/issues/320#issuecomment-92271111
Received on Monday, 13 April 2015 08:34:09 UTC