On Wed, Aug 27, 2008 at 11:23:11PM +0000, Kevin Rolfes wrote: > Package: linux-image-2.6.18-6-486 > Version: 2.6.18.dfsg.1-22etch2 > Severity: minor > > > I think this is minor, but it caught my attention and I thought the > package maintainer might want to know about it. Upgrading this > package from 2.6.18.dfsg.1-22 to 2.6.18.dfsg.1-22etch2 caused the > bogomips benchmark to drop from 99.07 to 20.03 on a Vortex86SX > based computer. > > Downgrading this package back to 2.6.18.dfsg-1-22 (using > "apt-get install linux-image-2.6.18-6-486=2.6.18.dfsg.1-22"), > and changing nothing else, restored the bogomips rating back to 99.07. > > I do not know if the computer actually ran slower. I didn't notice > a difference while using it, it was just the bogomips number that > caught my attention. If additional testing would be helpful, please > suggest something that would run on a minimal (command-line only) Etch > installation. > > The computer is an Ebox 2300SX, which is a tiny x86 computer that > contains a Vortex86SX processor running at 300MHz. (More CPU info at > http://www.dmp.com.tw/tech/vortex86sx/) > I installed Debian Etch 4.0.r4 which came with version > 2.6.18.dfsg.1-22 of the package linux-image-2.6.18-6-486. During > startup, the computer reported 99.07 BogoMIPS which was in line > with other reported ratings (i.e. 1/3 the processor clock). > > After the installation, I added security.debian.org to > /etc/apt/sources.list and used apt-get to upgrade the installed > packages. According to the time stamps of the installed files > and the log files, it was the next reboot that showed the > reduced BogoMIPS. > > Here is /proc/cpuinfo with the rolled back (2.6.18.dfsg.1-22) > version of linux-images: > > processor : 0 > vendor_id : unknown > cpu family : 4 > model : 0 > model name : 486 > stepping : unknown > fdiv_bug : no > hlt_bug : no > f00f_bug : no > coma_bug : no > fpu : no > fpu_exception : no > cpuid level : -1 > wp : yes > flags : > bogomips : 99.07
Does this error still occur with more recent kernel versions? If you're running Etch, could you try to reproduce this bug with the 2.6.24 based kernel added in 4.0r4? http://packages.qa.debian.org/l/linux-2.6.24.html Cheers, Moritz -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
