lkml.org 
[lkml]   [2003]   [Dec]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: bad performance on 2.4.23
Marcelo Tosatti wrote:

> About the high numbers on -ac and -aa:
>
> -ac includes rmap and the drop_behind() logic (I just posted the patch
> against 2.4.23 to lkml). I believe its the reason for the read slowdowns
> reported on lkml.
>
> -aa includes this patch which will increase the max readahead
> significantly. Mind trying it?

Sorry. I don't have any free machine to do this kind of tests. :(

>>In 'Sequential Reads' and 'Sequential Writes', 'Maximum Latency' is _too much high_
>>
>>2.4.23-pre4 8192 4096 256 5.10 6.83% 430.551 1602091.53 0.35501 0.31585 75
>>2.4.23-rc1 8192 4096 256 5.04 6.94% 432.486 1937701.66 0.33408 0.29884 73

> Now this is really odd. When did it started happening?

It looks like between .20 and .23, included.

Test was done by Randy Hron <rwhron@earthlink.net> over a OSDL server:
4x700 mhz Pentium III Xeons with 1MB cache 3.75 GB RAM
DAC960 Fiber channel to SCSI disks in RAID5 configuration
more details at http://home.earthlink.net/~rwhron/kernel/bigbox.html

what is very strange is that in two different tests and kernels
(2.4.23-pre4 and2.4.23-rc1) 'Maximum Latency' is very high with
256 threads.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:59    [W:0.054 / U:0.264 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site