AIM9 from 2.6.22 to 2.6.29

AIM9 from 2.6.22 to 2.6.29

[Christoph Lameter]

Here is a list of AIM9 results for all kernels between 2.6.22 2.6.29:

Significant regressions:

creat-clo
page_test
brk_test
exec_test
fork_test (!!)
shell_*
fifo_test
pipe_cpy

Significant improvements:

signal_test
tcp_test
udp_test

Kernel		2.6.22	2.6.23	2.6.24	2.6.25	2.6.26	2.6.27	2.6.28	2.6.29
==============================================================================
add_double	116.97	116.97	116.93	116.97	116.93	116.97	116.93	116.96
add_float	175.41	175.41	175.41	175.47	175.37	175.37	175.40	175.44
add_long	90.21	90.21	90.23	90.20	90.20	90.23	90.20	90.24
add_int		90.24	90.21	90.23	90.24	90.17	90.20	90.20	90.21
add_short	225.52	225.49	225.57	225.52	225.53	225.49	225.46	225.60
creat-clo	488.77	474.00	484.93	478.81	446.78	471.17	451.50	429.66
page_test	314.30	310.93	262.43	258.98	273.51	273.97	256.10	242.09
brk_test	316.59	314.83	307.03	273.77	290.53	278.27	282.84	272.11
jmp_test	34515	34513	34503	34518	34508	34525	34506	34503
signal_test	701.40	697.93	867.33	860.33	824.67	803.77	746.43	740.52
exec_test	220.96	232.47	211.00	201.33	201.80	205.46	201.97	200.63
fork_test	111.20	105.76	110.97	55.40	52.23	52.98	51.77	47.95
link_test	2713.66	2715.93	2770.50	2716.79	2710.53	2717.66	2737.43	2728.79
disk_rr		61.69	60.38	63.51	63.18	61.63	61.88	62.61	60.83
disk_rw		53.08	51.82	53.78	53.30	52.48	52.87	53.22	51.83
disk_rd		542.02	542.42	606.90	602.93	546.95	558.55	554.42	560.95
disk_wrt	76.17	74.68	78.17	77.58	76.42	76.81	78.01	75.18
disk_cp		61.18	60.23	62.68	63.38	62.01	62.56	62.75	60.59
sync_disk_rw	5.84	5.91	5.74	5.76	5.66	5.51	5.47	5.56
sync_disk_wrt	1.52	1.56	1.51	1.52	0.09	0.09	0.09	1.46
sync_disk_cp	1.52	1.55	1.50	1.51	0.09	0.10	0.10	1.46
disk_src	1845.70	1844.33	1852.70	1870.57	1838.69	1861.63	1888.07	1830.12
div_double	180.47	180.38	180.47	180.41	180.41	180.37	180.41	180.41
div_float	180.38	180.47	180.41	180.50	180.41	180.43	180.43	180.41
div_long	112.73	112.72	112.80	112.70	112.77	112.73	112.70	112.70
div_int		412.26	412.23	412.37	412.33	412.33	412.16	412.33	412.26
div_short	390.10	389.77	390.00	389.90	388.90	389.90	388.54	390.10
fun_cal		1027.69	1027.59	1027.90	1027.69	1027.36	1027.67	1027.77	1027.59
fun_cal1	1007.50	1007.43	1007.60	1007.70	1007.60	1007.57	1091.60	1091.74
fun_cal2	1013.26	1018.16	1072.24	1012.97	1011.23	1011.06	1012.20	1010.96
fun_cal15	318.16	395.50	317.83	322.63	317.60	350.50	323.79	397.00
sieve		31.32	31.31	31.32	31.34	31.31	31.33	31.31	31.32
mul_double	105.23	105.23	105.27	105.23	105.27	105.23	105.23	105.23
mul_float	131.60	131.56	131.56	131.60	131.52	131.56	131.52	131.51
mul_long	2719.39	2711.36	2714.80	2719.26	2719.07	2719.50	2719.17	2717.96
mul_int		5006.00	5007.13	5006.53	5007.17	5004.60	5004.90	5005.40	5005.56
mul_short	4005.90	4006.00	4006.50	4006.93	4040.55	4006.07	4004.03	4010.66
num_rtns_1	5737.62	5754.18	5737.87	5754.45	5755.27	5757.10	5752.43	5750.77
new_raph	16799	16803	16805	16434	16799	16796	16805	16799
trig_rtns	188.60	188.71	188.60	188.74	183.50	188.50	188.90	188.00
matrix_rtns	75809	75847	74889	74951	75830	75867	75916	70044
array_rtns	100.50	100.50	100.50	100.60	100.50	100.53	100.57	100.50
string_rtns	64.58	64.78	64.69	64.58	64.49	64.55	64.35	64.58
mem_rtns_1	234.12	236.11	236.37	236.25	233.60	236.69	237.73	235.95
mem_rtns_2	14093	14204	14098	14203	14126	14100	14179	next
sort_rtns_1	284.20	283.51	283.77	284.11	284.71	284.87	284.17	285.24
misc_rtns_1	4459.75	4608.23	4621.33	4652.40	4615.70	4625.30	4431.52	4501.30
dir_rtns_1	242.04	250.48	247.50	247.12	248.98	250.75	251.77	250.60
shell_rtns_1	272.80	289.80	266.63	254.15	252.10	251.57	253.32	252.85
shell_rtns_2	272.88	290.41	266.70	254.80	251.83	253.42	253.62	255.61
shell_rtns_3	273.24	290.27	266.71	251.70	253.52	252.78	253.75	254.78
series_1	157573	157432	157527	157333	157182	157566	157523	157353
shared_memory	12040	12233	10198	9652	9526	11991	11972	11982
tcp_test	2779	2801.70	2789	3404.27	3615.97	3294.03	3287.70	3112.96
udp_test	4717	4754.18	4881.60	5933.09	6114.89	6036.09	5510.03	5823.66
fifo_test	14463	14203	14650	14454	12951	13199	13135	13150
stream_pipe	13003	13105	13202	12514	13144	12887	12767	12930
dgram_pipe	12060	12124	12332	11729	12219	12067	11982	12094
pipe_cpy	17166	17396	18061	17318	15643	15794	15392	15784
ram_copy	201765	201054	201576	201421	201500	201063	201389	201351

--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[KOSAKI Motohiro]

> Here is a list of AIM9 results for all kernels between 2.6.22 2.6.29:
> 
> Significant regressions:
> 
> creat-clo
> page_test

I'm interest to it.
How do I get AIM9 benchmark?

and, Can you compare CONFIG_UNEVICTABLE_LRU is y and n?


> brk_test
> exec_test
> fork_test (!!)
> shell_*
> fifo_test
> pipe_cpy
> 
> Significant improvements:
> 
> signal_test
> tcp_test
> udp_test


--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Christoph Lameter]

On Sat, 18 Apr 2009, KOSAKI Motohiro wrote:

> > Here is a list of AIM9 results for all kernels between 2.6.22 2.6.29:
> >
> > Significant regressions:
> >
> > creat-clo
> > page_test
>
> I'm interest to it.
> How do I get AIM9 benchmark?

Checkout reaim9 on sourceforge.
>
> and, Can you compare CONFIG_UNEVICTABLE_LRU is y and n?


Sure.

--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[KOSAKI Motohiro]

> On Sat, 18 Apr 2009, KOSAKI Motohiro wrote:
> 
> > > Here is a list of AIM9 results for all kernels between 2.6.22 2.6.29:
> > >
> > > Significant regressions:
> > >
> > > creat-clo
> > > page_test
> >
> > I'm interest to it.
> > How do I get AIM9 benchmark?
> 
> Checkout reaim9 on sourceforge.

sourceforge search engine don't search reaim9 ;)


http://sourceforge.net/search/?words=aim9&type_of_search=soft&pmode=0&words=reaim9&Search=Search


> > and, Can you compare CONFIG_UNEVICTABLE_LRU is y and n?
> 
> Sure.





--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Pekka Enberg]

Hi!

On Tue, Apr 21, 2009 at 4:20 AM, KOSAKI Motohiro
<kosaki.motohiro@jp.fujitsu.com> wrote:
>> On Sat, 18 Apr 2009, KOSAKI Motohiro wrote:
>>
>> > > Here is a list of AIM9 results for all kernels between 2.6.22 2.6.29:
>> > >
>> > > Significant regressions:
>> > >
>> > > creat-clo
>> > > page_test
>> >
>> > I'm interest to it.
>> > How do I get AIM9 benchmark?
>>
>> Checkout reaim9 on sourceforge.
>
> sourceforge search engine don't search reaim9 ;)
>
>
> http://sourceforge.net/search/?words=aim9&type_of_search=soft&pmode=0&words=reaim9&Search=Search

I can only find "reaim7" so maybe Christoph meant the aim9 suite here:

  http://aimbench.sourceforge.net/

--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Christoph Lameter]

On Tue, 21 Apr 2009, KOSAKI Motohiro wrote:

> sourceforge search engine don't search reaim9 ;)

It has an obscure name so the search wont find it.

http://sourceforge.net/project/showfiles.php?group_id=71019&package_id=79015&release_id=276799

or try

http://gentwo.org/benchmark

I renamed the tarballs to aim9 and aim7 so that you can see which is
which. Also includes Mel's vmregress stuff.


--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Christoph Lameter]

On Tue, 21 Apr 2009, Pekka Enberg wrote:

> > http://sourceforge.net/search/?words=aim9&type_of_search=soft&pmode=0&words=reaim9&Search=Search
>
> I can only find "reaim7" so maybe Christoph meant the aim9 suite here:
>
>   http://aimbench.sourceforge.net/

Oh they tried to get this sorted out. The problem is nowhere does AIM9
appear so its not discoverable via google or the sf net search engine.

--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Mel Gorman]

On Tue, Apr 21, 2009 at 10:32:51AM -0400, Christoph Lameter wrote:
> On Tue, 21 Apr 2009, KOSAKI Motohiro wrote:
> 
> > sourceforge search engine don't search reaim9 ;)
> 
> It has an obscure name so the search wont find it.
> 
> http://sourceforge.net/project/showfiles.php?group_id=71019&package_id=79015&release_id=276799
> 
> or try
> 
> http://gentwo.org/benchmark
> 
> I renamed the tarballs to aim9 and aim7 so that you can see which is
> which. Also includes Mel's vmregress stuff.
> 

Wow, that's probably unusable it's so old. I don't release vmregress any
more but I'm up to version 0.89-rc5

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab

--
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:"dont@kvack.org"> email@kvack.org </a>

Re: AIM9 from 2.6.22 to 2.6.29

[Christoph Lameter]

On Thu, 23 Apr 2009, Mel Gorman wrote:

> Wow, that's probably unusable it's so old. I don't release vmregress any
> more but I'm up to version 0.89-rc5

Well vmregress includes a nice AIM9diff script that I am using. Where can
we get the newest version of your vmregress?



--
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:"dont@kvack.org"> email@kvack.org </a>