Re: [PATCH] concurrent inode allocation for ext2 against 2.5.64

Andrew Morton (akpm@digeo.com)
Sat, 15 Mar 2003 14:37:18 -0800


William Lee Irwin III <wli@holomorphy.com> wrote:
>
> On Sun, Mar 16, 2003 at 12:01:38AM +0300, Alex Tomas wrote:
> > here is the patch for ext2 concurrent inode allocation. should be applied
> > on top of previous concurrent-balloc patch. tested on dual p3 for several
> > hours of stress-test + fsck. hope someone test it on big iron ;)
>
> 32x/48GB NUMA-Q
>
> Throughput 257.986 MB/sec 128 procs
> dbench 128 95.36s user 4833.06s system 2832% cpu 2:53.97 total
>
> vma samples %-age symbol name
> c01dc9ac 4532033 21.4566 .text.lock.dec_and_lock
> c0169c0b 3835802 18.1603 .text.lock.dcache
> c0106ff4 1741849 8.24666 default_idle

Looks like it's gone nuts when 128 processes all try to close lots of
files at the same time.

One possible reason for this leaping out is that all the instances are now
achieving more uniform runtimes. You can tell that by comparing the dbench
dots.

-
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/