On Mon, Mar 14, 2011 at 4:29 PM, Jacob Meuser <[email protected]>
wrote:
>
> i386
>
>  32-bit mfi frames (current code)
>   32m23.04s         33m50.57s         10m49.27s
>   32m31.09s         33m48.24s         10m51.21s
>
>  64-bit mfi frames (with patch)
>   32m29.76s         33m36.95s         11m3.20s
>   32m41.15s         33m38.55s         11m7.87s

> so, the question is, is the impact on i386 enough to warrant using
> 32-bit frames on 32-bit platforms?  if so, should this be decided
> at runtime or compile time?

I don't think so.  It's under 1% in the worst case.  That's a very
small difference, and the fact that user time got *better* when it
should have remained the same makes me think the difference is within
the noise.

Reply via email to