[lkml]   [2005]   [Jul]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [patch 2.6.13-rc3a] i386: inline restore_fpu
-=> Kenneth Parrish wrote to All <=-

KP> 2.6.13-rc3-git5 lags -git4 on nbench's bitfield test, a part of its
KP> memory index, eg typical runs
KP> -git5
KP> BITFIELD : 1.1589e+07 : 1.99 : 0.42
KP> -git4
KP> BITFIELD : 2.1892e+07 : 3.76 : 0.78

KP> lmbench has slower context switch numbers for -git5.

i've been experimentally adding -funswitch-loops and -fpeel loops to the
default -march=i586 in the ../arch/i386/Makefile

cflags-$(CONFIG_M586) += -march=i586 -funswitch-loops -fpeel-loops

based on some tests with gcc 3.4.3,4 and the cyrix mii cpu, but
with v2.6.13-rc3-git5 carelessly put these flags on the next,
cflags-$(CONFIG_M586TSC) line, and so they would not have been used.
i shall retest after recompiling to confirm.
yeah, a couple nbench runs show similar overall results as -git4
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-07-25 04:48    [W:0.026 / U:6.712 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site