<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - [JSC] implement async functions proposal"
href="https://bugs.webkit.org/show_bug.cgi?id=156147#c97">Comment # 97</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - [JSC] implement async functions proposal"
href="https://bugs.webkit.org/show_bug.cgi?id=156147">bug 156147</a>
from <span class="vcard"><a class="email" href="mailto:caitp@igalia.com" title="Caitlin Potter (:caitp) <caitp@igalia.com>"> <span class="fn">Caitlin Potter (:caitp)</span></a>
</span></b>
<pre>The dromaeo results for the latest patch are about 3-4% regressions for each test. With the branch prediction adjustment, there's a 5-10% improvement (If there's a command to dump results of run-perf-tests as text, I'll provide all the numbers).
With the command `run-jsc-benchmarks --outer 10 --octane --benchmark "closure|jquery" baseline:~/git/WebKit2/WebKitBuild/Release/DumpRenderTree asyncawait:~/git/WebKit/WebKitBuild/Release/DumpRenderTree`, results are:
baseline asyncawait
closure 0.30989+-0.00551 0.30340+-0.00295 might be 1.0214x faster
jquery 3.65550+-0.01427 ! 3.78112+-0.03262 ! definitely 1.0344x slower
<geometric> 1.06425+-0.01004 ? 1.07103+-0.00613 ? might be 1.0064x slower
---
The jquery results might look a little worse than previously described, but the fact that Dromaeo consistently does so much better with the branch predictions added goes a long way in my opinion.
Ensuring that `m_runtimeFlags` is packed efficiently and accessed in cache more often will probably see some improvement, but hard to guess how much</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
</ul>
</body>
</html>