On 12/14/12 1:48 AM, Justin Lebar wrote:
FWIW, I once tried changing all of our atomic string refcounting to non-atomic operations and could not eke out a performance (or stability) difference on x64. This was despite the fact that I was able to generate profiles where the atomic string refcounting showed up as taking a few percentage points. I think bz reproduced this somewhat surprising result more recently.
What I found was that changing the refcounting on nsStringBuffer to not be atomic didn't obviously help. Neither did inlining those addref/release calls (they're out-of-line right now). But doing both together _did_ help a good bit, for my microbenchmark. Also x64 for that test. And _very_ microbenchmarky...
-Boris _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
