|
From: | Ian Rogers |
Subject: | Re: [Qemu-devel] Profiling Qemu for speed? |
Date: | Mon, 18 Apr 2005 15:29:02 +0100 |
User-agent: | Mozilla Thunderbird 0.8 (X11/20040913) |
Christian MICHON wrote:
could this be down to the tables used to find the translators/generators? are they constant? is it possible to make them amenable to feedback directed analysis? x86 is also an odd case, it could be that load-store architectures with more registers notice FDO more as the x86 can access the L1 caches similarly to a large register file. If the problem is gcc's feedback analysis then you could always try Intel's compiler. Regards,I did months ago gcc/FDO with a xp/lite installation as a "repetitive task" :) I did not improve the timings after all the effort.
Ian Rogers -- http://www.binarytranslator.org/
[Prev in Thread] | Current Thread | [Next in Thread] |