Alan Cox wrote:
>
> > Yep. I didn't notice this until I started updating my local
> > repository to sync to the last few 2.3.99 prepatches, and I am finding
> > that a large P4 operation (which hits the disk a lot, much like "cvs
> > import") takes several times longer than it did under 2.2.14.
>
> IDE or SCSI. Also how much memory do you have and is the maching paging
> or showing different vm behaviour (if you cant tell subjectively look for
> differences in vmstat 5 output during the operation)
Just finished a test.
Hardware: Dual P-II, Asus P2B mobo, 128MB RAM, one 10GB IDE driver,
rtl-8139a NIC
Software: Vanilla kernel 2.2.1[45], Linux-Mandrake 7.1 beta
The test: Boot, then in parallel:
cvs import of 2.2.16-pre4
cvs import of 2.3.99-pre5
Netscape web browsing and e-mail
.config: http://gtf.org/garzik/vmstat/dot.config.txt.gz
vmstat for 2.2.14: http://gtf.org/garzik/vmstat/vmstat.2.2.14.txt.gz
vmstat for 2.2.15: http://gtf.org/garzik/vmstat/vmstat.2.2.15.txt.gz
Quick analysis:
* The length of time between processes in the "r" or "w" columns appears
longer in 2.2.15
* A -lot- more stuff is swapped out, for the exact same usage pattern.
* The total memory used by disk buffers is pitifully small by comparison
-- Jeff Garzik | Liberty is always dangerous, but Building 1024 | it is the safest thing we have. MandrakeSoft, Inc. | -- Harry Emerson Fosdick- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:24 EST