lkml.org 
[lkml]   [1996]   [Dec]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: Mega Patch #8, or "Why stop now while we haven't broken it all"
Date

> What does make a difference is #define CLEVER_SMP_INVALIDATE
> at line 85 in include/asm/pgtable.h (asm-i386).

Yep

> For some reason context switch performance was 10% better at #6, against
> all the SMP 2.0 and 2.1 benchmarks I have run.

I've no idea why. Most odd

> The 33% improvement in process creation time has a direct impact on
> overall performance. Why isn't the tlb_flush optimization the default?

One mod at a time is the answer (it is btw in 2.0.27)

> The areas to get more from SMP are in context switch, process creation and
> syscall overhead improvements.

Looks like it, context switch will be hard to improve much without some
very very horribly clever FPU stuff (lazy saving with cross CPU call
to lazy save and copy FPU state to another CPU). I guess since everyone
seems to be FPU freaks with SMP boxes I ought try and do that.

Syscall looks like it ought to go much faster, even allowing for the
pulling of the cache line from cpu to cpu

Alan


\
 
 \ /
  Last update: 2005-03-22 13:38    [from the cache]
©2003-2011 Jasper Spaans