<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - STP TypedArray.subarray 5x slowdown compared to 9.1"
href="https://bugs.webkit.org/show_bug.cgi?id=156404#c3">Comment # 3</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - STP TypedArray.subarray 5x slowdown compared to 9.1"
href="https://bugs.webkit.org/show_bug.cgi?id=156404">bug 156404</a>
from <span class="vcard"><a class="email" href="mailto:ggaren@apple.com" title="Geoffrey Garen <ggaren@apple.com>"> <span class="fn">Geoffrey Garen</span></a>
</span></b>
<pre>(In reply to <a href="show_bug.cgi?id=156404#c2">comment #2</a>)
<span class="quote">> (In reply to <a href="show_bug.cgi?id=156404#c1">comment #1</a>)
> > Fascinating. That's a pretty serious regression. Keith, do you remember
> > what we've done with subarray?
> >
>
> This regression is not particularly surprising. The ES6 spec changes the
> behavior of the subarray function substantially. I expect that issue here is
> that the subarray function now uses species constructors, which require
> several property lookups. Namely, the constructor and the Symbol.species
> properties. Since the subarray function is in C++ (where we don't cache the
> property lookups) and the subarrays you are making are very small (4
> elements). The cost of the species construction will dominate the cost of
> the element copying.</span >
Why not perform the property accesses in JavaScript and then call a helper function to do the allocation?
I don't think we want to consider a 5X regression unsurprising.</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>