* vm benchmarking
@ 2004-02-24 11:40 Andrew Morton
2004-02-24 12:01 ` Nikita Danilov
2004-02-25 10:20 ` Nick Piggin
0 siblings, 2 replies; 6+ messages in thread
From: Andrew Morton @ 2004-02-24 11:40 UTC (permalink / raw)
To: Nick Piggin; +Cc: linux-mm
I took the various patches in -mm for a quick ride. Dual CPU, mem=64m,
`time make -j4 vmlinux':
2.4.25 2:57.34 2:45.62
up to blk_congestion_wait-return-remaining 5:41.52 5:56.37
up to vmscan-remove-priority
up to kswapd-throttling-fixes 7:44.53
up to vm-dont-rotate-active-list 6:29.23
up to vm-dont-rotate-active-list-padding
up to vm-lru-info 9:28.47 6:14.70 5:11.99
up to vm-shrink-zone
up to vm-shrink-zone-div-by-0-fix 9:13.21 8:17.29
up to vm-tune-throttle 7:39.89
up to shrink_slab-for-all-zones 7:06.27
up to zone-balancing-fix 7:46.15
up to zone-balancing-batching
up to zone-balancing-batching-fix 4:44.76 4:27.02 4:05.56 4:31.66 4:06.76
Based on this, and on your totally contradictory benchmarking, all I'm
prepared to say is that kswapd-throttling-fixes slows things down, and we
don't know why. The rest appears to be worth zilch. Possibly they slow
some things down as much as they speed other things up.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: vm benchmarking
2004-02-24 11:40 vm benchmarking Andrew Morton
@ 2004-02-24 12:01 ` Nikita Danilov
2004-02-24 12:07 ` Andrew Morton
2004-02-25 10:20 ` Nick Piggin
1 sibling, 1 reply; 6+ messages in thread
From: Nikita Danilov @ 2004-02-24 12:01 UTC (permalink / raw)
To: Andrew Morton; +Cc: Nick Piggin, linux-mm
Andrew Morton writes:
> I took the various patches in -mm for a quick ride. Dual CPU, mem=64m,
> `time make -j4 vmlinux':
>
>
> 2.4.25 2:57.34 2:45.62
>
> up to blk_congestion_wait-return-remaining 5:41.52 5:56.37
> up to vmscan-remove-priority
> up to kswapd-throttling-fixes 7:44.53
> up to vm-dont-rotate-active-list 6:29.23
> up to vm-dont-rotate-active-list-padding
> up to vm-lru-info 9:28.47 6:14.70 5:11.99
> up to vm-shrink-zone
> up to vm-shrink-zone-div-by-0-fix 9:13.21 8:17.29
> up to vm-tune-throttle 7:39.89
> up to shrink_slab-for-all-zones 7:06.27
> up to zone-balancing-fix 7:46.15
> up to zone-balancing-batching
> up to zone-balancing-batching-fix 4:44.76 4:27.02 4:05.56 4:31.66 4:06.76
Can you clarify what these numbers mean?
Nikita.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: vm benchmarking
2004-02-24 12:01 ` Nikita Danilov
@ 2004-02-24 12:07 ` Andrew Morton
0 siblings, 0 replies; 6+ messages in thread
From: Andrew Morton @ 2004-02-24 12:07 UTC (permalink / raw)
To: Nikita Danilov; +Cc: piggin, linux-mm
Nikita Danilov <Nikita@Namesys.COM> wrote:
>
> Andrew Morton writes:
> > I took the various patches in -mm for a quick ride. Dual CPU, mem=64m,
> > `time make -j4 vmlinux':
> >
> >
> > 2.4.25 2:57.34 2:45.62
> >
> > up to blk_congestion_wait-return-remaining 5:41.52 5:56.37
> > up to vmscan-remove-priority
> > up to kswapd-throttling-fixes 7:44.53
> > up to vm-dont-rotate-active-list 6:29.23
> > up to vm-dont-rotate-active-list-padding
> > up to vm-lru-info 9:28.47 6:14.70 5:11.99
> > up to vm-shrink-zone
> > up to vm-shrink-zone-div-by-0-fix 9:13.21 8:17.29
> > up to vm-tune-throttle 7:39.89
> > up to shrink_slab-for-all-zones 7:06.27
> > up to zone-balancing-fix 7:46.15
> > up to zone-balancing-batching
> > up to zone-balancing-batching-fix 4:44.76 4:27.02 4:05.56 4:31.66 4:06.76
>
> Can you clarify what these numbers mean?
>
Duration of `time make -j4 vmlinux', minutes:seconds
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: vm benchmarking
2004-02-25 10:20 ` Nick Piggin
@ 2004-02-24 23:43 ` Andrew Morton
2004-02-25 10:50 ` Nick Piggin
0 siblings, 1 reply; 6+ messages in thread
From: Andrew Morton @ 2004-02-24 23:43 UTC (permalink / raw)
To: Nick Piggin; +Cc: linux-mm
Nick Piggin <piggin@cyberone.com.au> wrote:
>
> My machine doesn't touch swap at make -j4 with mem=64m. It is
> dual CPU with a SMP kernel but I was using maxcpus=1.
It is light-to-moderate paging.
> It compiles 2.4.21 with gcc-3.3.3 I think (I can tell you when I
> get home).
gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)
This is a 2.4.19 defconfig build.
> I can't explain your results. Maybe you have other stuff running.
Only `vmstat 1'.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: vm benchmarking
2004-02-24 11:40 vm benchmarking Andrew Morton
2004-02-24 12:01 ` Nikita Danilov
@ 2004-02-25 10:20 ` Nick Piggin
2004-02-24 23:43 ` Andrew Morton
1 sibling, 1 reply; 6+ messages in thread
From: Nick Piggin @ 2004-02-25 10:20 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-mm
Andrew Morton wrote:
>I took the various patches in -mm for a quick ride. Dual CPU, mem=64m,
>`time make -j4 vmlinux':
>
>
>2.4.25 2:57.34 2:45.62
>
>up to blk_congestion_wait-return-remaining 5:41.52 5:56.37
>up to vmscan-remove-priority
>up to kswapd-throttling-fixes 7:44.53
>up to vm-dont-rotate-active-list 6:29.23
>up to vm-dont-rotate-active-list-padding
>up to vm-lru-info 9:28.47 6:14.70 5:11.99
>up to vm-shrink-zone
>up to vm-shrink-zone-div-by-0-fix 9:13.21 8:17.29
>up to vm-tune-throttle 7:39.89
>up to shrink_slab-for-all-zones 7:06.27
>up to zone-balancing-fix 7:46.15
>up to zone-balancing-batching
>up to zone-balancing-batching-fix 4:44.76 4:27.02 4:05.56 4:31.66 4:06.76
>
>
>Based on this, and on your totally contradictory benchmarking, all I'm
>prepared to say is that kswapd-throttling-fixes slows things down, and we
>don't know why. The rest appears to be worth zilch. Possibly they slow
>some things down as much as they speed other things up.
>
>
My machine doesn't touch swap at make -j4 with mem=64m. It is
dual CPU with a SMP kernel but I was using maxcpus=1.
It compiles 2.4.21 with gcc-3.3.3 I think (I can tell you when I
get home).
I can't explain your results. Maybe you have other stuff running.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: vm benchmarking
2004-02-24 23:43 ` Andrew Morton
@ 2004-02-25 10:50 ` Nick Piggin
0 siblings, 0 replies; 6+ messages in thread
From: Nick Piggin @ 2004-02-25 10:50 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-mm
Andrew Morton wrote:
>Nick Piggin <piggin@cyberone.com.au> wrote:
>
>>My machine doesn't touch swap at make -j4 with mem=64m. It is
>>dual CPU with a SMP kernel but I was using maxcpus=1.
>>
>
>It is light-to-moderate paging.
>
>
>>It compiles 2.4.21 with gcc-3.3.3 I think (I can tell you when I
>>get home).
>>
>
>gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)
>
>This is a 2.4.19 defconfig build.
>
>
So it should be pretty similar to what I've been doing.
>>I can't explain your results. Maybe you have other stuff running.
>>
>
>Only `vmstat 1'.
>
>
That shouldn't hurt. Maybe running two CPUs is a problem. I'd better
try that.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-02-25 10:50 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-02-24 11:40 vm benchmarking Andrew Morton
2004-02-24 12:01 ` Nikita Danilov
2004-02-24 12:07 ` Andrew Morton
2004-02-25 10:20 ` Nick Piggin
2004-02-24 23:43 ` Andrew Morton
2004-02-25 10:50 ` Nick Piggin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox