Re: [BENCHMARK] I/O regression after 2.6.0-test5

[Nick Piggin <piggin@cyberone.com.au>]

Andrew Morton wrote:

>Nick Piggin <piggin@cyberone.com.au> wrote:
>
>>
>>
>>rwhron@earthlink.net wrote:
>>
>>
>>>There was about a 50% regression in jobs/minute in AIM7
>>>database workload on quad P3 Xeon.  The CPU time has not
>>>gone up, so the extra run time is coming from something
>>>else.  (I/O or I/O scheduler?)
>>>
>>>tiobench sequential reads has a significant regression too.
>>>
>>>Regression appears unrelated to filesystem type.
>>>
>>>dbench was not affected.
>>>
>>>The AIM7 was run on ext2.
>>>
>>>
>>Yeah I'd say its all due to the IO scheduler. There is a problem
>>I'm thinking about how to fix - its the likely cause of this too.
>>
>>
>
>What change do you think it was due to?
>
>

I was thinking: [PATCH] fix AS crappy performance

(It still doesn't work properly)

>It's rather strange that test6 is slow but test6-mm is not: generally the
>IO scheduler regressions enter -mm first ;)
>

But if test6-mm isn't slow then maybe it is due to something else