Re: [BENCHMARKS] 2.5.70 for 4 filesystems

From: Hans Reiser (reiser@namesys.com)
Date: Mon Jun 02 2003 - 13:47:37 EST


You should expect V3 to be slow at quad CPU smp benchmarks because
balancing is giant locked. V4 fixes this with fine grained locking, we
hope to release at Linux Tag in July. Fortunately quad CPU boxes will
not be economical compared to dual CPU boxes until V4 has been out for a
while....;-) V4 CPU usage and general performance has gotten a lot
better, we need to put a new snapshot on our website.....

Andi Kleen wrote:

>rwhron@earthlink.net writes:
>
>
>> --------------- Sequential ----------
>> ----- Create ----- ---- Delete ----
>> /sec %CPU Eff /sec %CPU Eff
>>2.5.70-reiserfs 7584 86.7 8751 2628 37.3 7038
>>2.5.70-xfs 1710 39.3 4347 2053 28.3 7247
>>2.5.70-ext2 150 99.0 151 60883 100.0 6088
>>2.5.70-ext3 119 95.0 126 26319 87.7 3002
>>
>>
>
>It's quite surprising that reiserfs is so slow at deletion. In my
>normal experience reiserfs rm -rf is much faster than anything else
>(e.g. with a big rm -rf on an ext2 you have a chance to ctrl-c still,
>on reiserfs no such chance; XFS is really slow at this). Perhaps this
>is some 2.5 regression? Do you have 2.4 comparison numbers?
>
>-Andi
>
>-
>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/
>
>
>
>

-- 
Hans

- 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/



This archive was generated by hypermail 2b29 : Sat Jun 07 2003 - 22:00:17 EST