[cfe-commits] [patch] change libc++'s string hash function to cityhash64

[Chandler Carruth]

While I tend to agree with you on the size being worth the performance, I
wonder...

On Thu, Dec 8, 2011 at 5:16 PM, Craig Silverstein <csilvers at google.com>wrote:

> SIZE (-O2):
> murmur2: 31841
> cityhash64: 36082    // 13.3% bigger
>

It would be nice to break this down into the constant overhead, the
overhead per-call, and finally the overhead per-type. Essentially, try with
1, 2, and 3 loops, see how it grows; then try with 1, 2, and 3 types? Maybe
that's too much work?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20111208/278d8182/attachment.html>