From: Anca Emanuel on
Linux 2.6.24 Through Linux 2.6.33 Benchmarks

http://www.phoronix.com/scan.php?page=article&item=linux_2624_2633&num=1

quote:
"This test system was using the EXT3 file-system and we found its
TPC-B database performance to improve by many times between the Linux
2.6.29 and 2.6.30 kernel releases. In fact, the Linux 2.6.30 kernel
was 770% faster than its predecessor was and it remained that way with
the Linux 2.6.31 kernel and then regressed only slightly with the
Linux 2.6.32 kernel. With the new Linux 2.6.33 kernel, however, the
PostgreSQL performance atop the EXT3 file-system has fallen off a
cliff. Not only has the PostgreSQL performance lost its gains made
during the 2.6.30 development cycle, but also with Linux 2.6.33, it is
actually much slower than it was in any of the pre-2.6.30 releases."

http://www.phoronix.com/data/img/results/linux_2624_2633/2.png
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Paolo Ornati on
On Wed, 10 Mar 2010 20:34:12 +0200
Anca Emanuel <anca.emanuel(a)gmail.com> wrote:

> In fact, the Linux 2.6.30 kernel
> was 770% faster than its predecessor was and it remained that way with
> the Linux 2.6.31 kernel and then regressed only slightly with the
> Linux 2.6.32 kernel. With the new Linux 2.6.33 kernel, however, the
> PostgreSQL performance atop the EXT3 file-system has fallen off a
> cliff.

The extra performance was just a "bug":

http://www.mail-archive.com/pgsql-performance(a)postgresql.org/msg34841.html

"[This change] is required for safe behavior with volatile write caches
on drives. You could mount with -o nobarrier and [the performance drop]
would go away, but a sequence like write->fsync->lose power->reboot may
well find your file without the data that you synced, if the drive had
write caches enabled. If you know you have no write cache, or that it
is safely battery backed, then you can mount with -o nobarrier, and not
incur this penalty."

--
Paolo Ornati
Linux 2.6.33-00001-gbaac35c on x86_64
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/