* cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
@ 2007-11-08 7:34 Darren Hart
2007-11-08 13:56 ` Gregory Haskins
2007-11-13 13:40 ` Jaswinder Singh
0 siblings, 2 replies; 10+ messages in thread
From: Darren Hart @ 2007-11-08 7:34 UTC (permalink / raw)
To: RT; +Cc: Thomas Gleixner, Steven Rostedt, Gregory Haskins
[-- Attachment #1: Type: text/plain, Size: 2424 bytes --]
Greg,
Here are the results I promised you. Sorry it took so long. I don't think
they are particularly interesting, perhaps more iterations, say 1000000?
If there are different options you would like to see let me know, I have all
5 kernels sitting on the machine and can keep it for a couple days. I
attached a simple bar chart to quickly summarize the data.
o testing 2.6.23.1-rt7 to 2.6.23.1-rt11 for ghaskins et all
o load used was building 2.6.23.1 make -j 16
o system:
- 8 Opterons @ 1.6 GHz
- x86_64
- 2 GB RAM
# uname -r
2.6.23.1-rt7
IDLE
# ./cyclictest -n -i 10000 -l 10000 -p 95
3.04 8.17 4.65 1/231 8854
T: 0 ( 8852) P:95 I:10000 C: 10000 Min: 5 Act: 7 Avg: 6 Max: 18
LOADED
# ./cyclictest -n -i 10000 -l 10000 -p 95
15.69 5.85 2.19 17/311 19488
T: 0 ( 5111) P:95 I:10000 C: 10000 Min: 7 Act: 16 Avg: 18 Max: 44
# uname -r
2.6.23.1-rt8
IDLE
# ./cyclictest -n -i 10000 -l 10000 -p 95
0.90 5.76 4.89 1/231 27569
T: 0 (27566) P:95 I:10000 C: 10000 Min: 5 Act: 6 Avg: 6 Max: 14
LOADED
# ./cyclictest -n -i 10000 -l 10000 -p 95
15.18 5.56 2.03 13/303 19336
T: 0 ( 4836) P:95 I:10000 C: 10000 Min: 7 Act: 19 Avg: 19 Max: 42
# uname -r
2.6.23.1-rt9
IDLE
# ./cyclictest -n -i 10000 -l 10000 -p 95
2.52 4.05 1.95 1/231 19749
T: 0 (19746) P:95 I:10000 C: 10000 Min: 6 Act: 8 Avg: 6 Max: 14
LOADED
# ./cyclictest -n -i 10000 -l 10000 -p 95
15.81 5.84 2.15 17/310 19573
T: 0 ( 4955) P:95 I:10000 C: 10000 Min: 8 Act: 21 Avg: 17 Max: 45
# uname -r
2.6.23.1-rt10
IDLE
# ./cyclictest -n -i 10000 -l 10000 -p 95
1.26 3.51 1.95 4/231 21269
T: 0 (21267) P:95 I:10000 C: 10000 Min: 6 Act: 7 Avg: 7 Max: 16
LOADED
# ./cyclictest -n -i 10000 -l 10000 -p 95
# 15.33 5.60 2.04 21/311 19269
T: 0 ( 4900) P:95 I:10000 C: 10000 Min: 7 Act: 20 Avg: 18 Max: 35
# uname -r
2.6.23.1-rt11
IDLE
# ./cyclictest -n -i 10000 -l 10000 -p 95
3.40 8.32 4.66 1/231 10118
T: 0 (10116) P:95 I:10000 C: 10000 Min: 5 Act: 7 Avg: 6 Max: 12
LOADED
# ./cyclictest -n -i 10000 -l 10000 -p 9515.68 5.74 2.10 18/313 19181
T: 0 ( 4963) P:95 I:10000 C: 10000 Min: 7 Act: 14 Avg: 18 Max: 40
--
Darren Hart
IBM Linux Technology Center
Realtime Linux Team
[-- Attachment #2: plot.png --]
[-- Type: image/png, Size: 16798 bytes --]
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-08 7:34 cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11) Darren Hart
@ 2007-11-08 13:56 ` Gregory Haskins
2007-11-08 15:56 ` Darren Hart
2007-11-13 13:30 ` Jaswinder Singh
2007-11-13 13:40 ` Jaswinder Singh
1 sibling, 2 replies; 10+ messages in thread
From: Gregory Haskins @ 2007-11-08 13:56 UTC (permalink / raw)
To: Darren Hart, RT; +Cc: Steven Rostedt, Thomas Gleixner
[-- Attachment #1: Type: text/plain, Size: 3383 bytes --]
>>> On Thu, Nov 8, 2007 at 2:34 AM, in message
<200711072334.40216.dvhltc@us.ibm.com>, Darren Hart <dvhltc@us.ibm.com> wrote:
> Greg,
>
> Here are the results I promised you.
Darren,
First off, thanks a bunch for running through those tests!
> I don't think they are particularly interesting
They *are* interesting in the sense that peaks of 40ish for an 8-way are excellent IMO ;) But yeah, i see what you mean; theres no interesting difference between them.
Do you have any comparative data from runs in -rt1 or earlier kernels?
> perhaps more iterations, say 1000000?
We generally load the system up a lot more (e.g. "make -j 128") or let it run for a long time with the make in a loop (or both). For instance, here is the last data I gathered (with plot attached). This was for a single pass of "make mrproper; make allmodconfig; time make -j 128"
23.1-rt7
-----------------
real 10m49.312s
user 49m59.674s
sys 16m15.404s
26.38 76.02 56.11 1/304 21301
T: 0 ( 5959) P:90 I:100 C:7500381 Min: 2 Act: 2 Avg: 4 Max: 78
T: 1 ( 5960) P:89 I:200 C:3750191 Min: 2 Act: 3 Avg: 5 Max: 91
T: 2 ( 5961) P:88 I:300 C:2500127 Min: 2 Act: 3 Avg: 4 Max: 66
T: 3 ( 5962) P:87 I:400 C:1875096 Min: 2 Act: 5 Avg: 6 Max: 106
T: 4 ( 5963) P:86 I:500 C:1500077 Min: 2 Act: 6 Avg: 5 Max: 92
T: 5 ( 5964) P:85 I:600 C:1250064 Min: 2 Act: 5 Avg: 6 Max: 93
T: 6 ( 5965) P:84 I:700 C:1071483 Min: 2 Act: 3 Avg: 6 Max: 78
T: 7 ( 5966) P:83 I:800 C: 937548 Min: 2 Act: 5 Avg: 6 Max: 97
23.1-rt10
-----------------
real 10m39.046s
user 56m31.662s
sys 12m51.158s
7.64 63.88 62.70 1/291 30774
T: 0 (15455) P:90 I:100 C:8001856 Min: 2 Act: 3 Avg: 4 Max: 65
T: 1 (15456) P:89 I:200 C:4000928 Min: 2 Act: 3 Avg: 4 Max: 46
T: 2 (15457) P:88 I:300 C:2667286 Min: 2 Act: 3 Avg: 5 Max: 54
T: 3 (15458) P:87 I:400 C:2000464 Min: 2 Act: 3 Avg: 5 Max: 46
T: 4 (15459) P:86 I:500 C:1600372 Min: 2 Act: 3 Avg: 5 Max: 46
T: 5 (15460) P:85 I:600 C:1333643 Min: 2 Act: 5 Avg: 5 Max: 75
T: 6 (15461) P:84 I:700 C:1143123 Min: 2 Act: 5 Avg: 6 Max: 52
T: 7 (15462) P:83 I:800 C:1000232 Min: 2 Act: 4 Avg: 5 Max: 66
23.1-rt11
-----------------
real 10m41.610s
user 55m37.065s
sys 13m16.394s
101.92 101.56 61.65 1/308 21349
T: 0 ( 6028) P:90 I:100 C:6584414 Min: 2 Act: 3 Avg: 4 Max: 55
T: 1 ( 6029) P:89 I:200 C:3292207 Min: 2 Act: 6 Avg: 5 Max: 58
T: 2 ( 6030) P:88 I:300 C:2194805 Min: 2 Act: 5 Avg: 5 Max: 54
T: 3 ( 6031) P:87 I:400 C:1646104 Min: 2 Act: 5 Avg: 5 Max: 79
T: 4 ( 6032) P:86 I:500 C:1316883 Min: 2 Act: 5 Avg: 5 Max: 60
T: 5 ( 6033) P:85 I:600 C:1097403 Min: 2 Act: 5 Avg: 6 Max: 45
T: 6 ( 6034) P:84 I:700 C: 940631 Min: 2 Act: 3 Avg: 6 Max: 45
T: 7 ( 6035) P:83 I:800 C: 823052 Min: 2 Act: 5 Avg: 6 Max: 47
As Steven Rostedt pointed out on IRC, numbers coming from me are suspect ;). But we can at least use them to see if you can repro similar results.
Thanks again!
Regards,
-Greg
[-- Attachment #2: Portable Network Graphics Format --]
[-- Type: image/png, Size: 5130 bytes --]
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-08 13:56 ` Gregory Haskins
@ 2007-11-08 15:56 ` Darren Hart
2007-11-13 13:30 ` Jaswinder Singh
1 sibling, 0 replies; 10+ messages in thread
From: Darren Hart @ 2007-11-08 15:56 UTC (permalink / raw)
To: Gregory Haskins; +Cc: RT, Steven Rostedt, Thomas Gleixner
On Thursday 08 November 2007 05:56:07 Gregory Haskins wrote:
> >>> On Thu, Nov 8, 2007 at 2:34 AM, in message
>
> <200711072334.40216.dvhltc@us.ibm.com>, Darren Hart <dvhltc@us.ibm.com>
wrote:
> > Greg,
> >
> > Here are the results I promised you.
>
> Darren,
> First off, thanks a bunch for running through those tests!
>
> > I don't think they are particularly interesting
>
> They *are* interesting in the sense that peaks of 40ish for an 8-way are
> excellent IMO ;) But yeah, i see what you mean; theres no interesting
> difference between them.
>
> Do you have any comparative data from runs in -rt1 or earlier kernels?
Unfortunately no.
>
> > perhaps more iterations, say 1000000?
>
> We generally load the system up a lot more (e.g. "make -j 128") or let it
> run for a long time with the make in a loop (or both). For instance, here
> is the last data I gathered (with plot attached). This was for a single
> pass of "make mrproper; make allmodconfig; time make -j 128"
I loosely based my runs off the examples here:
http://rt.wiki.kernel.org/index.php/Cyclictest
If there are a better set of "start here for cyclictest results" I would
really like it if someone would add that to the cyclictest wiki page,
complete with command line options, etc. For instance I can see you ran -t 8
below, but how many loops? with what interval? with which timer mechanism?
Thanks,
Darren
>
> 23.1-rt7
> -----------------
> real 10m49.312s
> user 49m59.674s
> sys 16m15.404s
>
> 26.38 76.02 56.11 1/304 21301
>
> T: 0 ( 5959) P:90 I:100 C:7500381 Min: 2 Act: 2 Avg: 4 Max:
> 78 T: 1 ( 5960) P:89 I:200 C:3750191 Min: 2 Act: 3 Avg: 5 Max:
> 91 T: 2 ( 5961) P:88 I:300 C:2500127 Min: 2 Act: 3 Avg: 4
> Max: 66 T: 3 ( 5962) P:87 I:400 C:1875096 Min: 2 Act: 5 Avg:
> 6 Max: 106 T: 4 ( 5963) P:86 I:500 C:1500077 Min: 2 Act: 6
> Avg: 5 Max: 92 T: 5 ( 5964) P:85 I:600 C:1250064 Min: 2 Act:
> 5 Avg: 6 Max: 93 T: 6 ( 5965) P:84 I:700 C:1071483 Min: 2
> Act: 3 Avg: 6 Max: 78 T: 7 ( 5966) P:83 I:800 C: 937548 Min:
> 2 Act: 5 Avg: 6 Max: 97
>
> 23.1-rt10
> -----------------
>
> real 10m39.046s
> user 56m31.662s
> sys 12m51.158s
>
> 7.64 63.88 62.70 1/291 30774
>
> T: 0 (15455) P:90 I:100 C:8001856 Min: 2 Act: 3 Avg: 4 Max:
> 65 T: 1 (15456) P:89 I:200 C:4000928 Min: 2 Act: 3 Avg: 4 Max:
> 46 T: 2 (15457) P:88 I:300 C:2667286 Min: 2 Act: 3 Avg: 5
> Max: 54 T: 3 (15458) P:87 I:400 C:2000464 Min: 2 Act: 3 Avg:
> 5 Max: 46 T: 4 (15459) P:86 I:500 C:1600372 Min: 2 Act: 3
> Avg: 5 Max: 46 T: 5 (15460) P:85 I:600 C:1333643 Min: 2 Act:
> 5 Avg: 5 Max: 75 T: 6 (15461) P:84 I:700 C:1143123 Min: 2
> Act: 5 Avg: 6 Max: 52 T: 7 (15462) P:83 I:800 C:1000232 Min:
> 2 Act: 4 Avg: 5 Max: 66
>
> 23.1-rt11
> -----------------
>
> real 10m41.610s
> user 55m37.065s
> sys 13m16.394s
>
> 101.92 101.56 61.65 1/308 21349
>
> T: 0 ( 6028) P:90 I:100 C:6584414 Min: 2 Act: 3 Avg: 4 Max:
> 55 T: 1 ( 6029) P:89 I:200 C:3292207 Min: 2 Act: 6 Avg: 5 Max:
> 58 T: 2 ( 6030) P:88 I:300 C:2194805 Min: 2 Act: 5 Avg: 5
> Max: 54 T: 3 ( 6031) P:87 I:400 C:1646104 Min: 2 Act: 5 Avg:
> 5 Max: 79 T: 4 ( 6032) P:86 I:500 C:1316883 Min: 2 Act: 5
> Avg: 5 Max: 60 T: 5 ( 6033) P:85 I:600 C:1097403 Min: 2 Act:
> 5 Avg: 6 Max: 45 T: 6 ( 6034) P:84 I:700 C: 940631 Min: 2
> Act: 3 Avg: 6 Max: 45 T: 7 ( 6035) P:83 I:800 C: 823052 Min:
> 2 Act: 5 Avg: 6 Max: 47
>
> As Steven Rostedt pointed out on IRC, numbers coming from me are suspect
> ;). But we can at least use them to see if you can repro similar results.
>
> Thanks again!
>
> Regards,
> -Greg
--
Darren Hart
IBM Linux Technology Center
Realtime Linux Team
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-08 13:56 ` Gregory Haskins
2007-11-08 15:56 ` Darren Hart
@ 2007-11-13 13:30 ` Jaswinder Singh
2007-11-13 14:20 ` Steven Rostedt
2007-11-13 16:19 ` Darren Hart
1 sibling, 2 replies; 10+ messages in thread
From: Jaswinder Singh @ 2007-11-13 13:30 UTC (permalink / raw)
To: Gregory Haskins; +Cc: Darren Hart, RT, Steven Rostedt, Thomas Gleixner
Hello Greg and Darren,
If possible, can you please also add processor ID/number in your
cyclictest and run more tasks (say 40, 60 or 80 tasks) then we have
good picture like how many processes are running in which processor
and how they switching between processors.
And can we force processes to run on specific processor or core.
Thank you,
Jaswinder Singh.
On Nov 8, 2007 7:26 PM, Gregory Haskins <ghaskins@novell.com> wrote:
> >>> On Thu, Nov 8, 2007 at 2:34 AM, in message
> <200711072334.40216.dvhltc@us.ibm.com>, Darren Hart <dvhltc@us.ibm.com> wrote:
>
> > Greg,
> >
> > Here are the results I promised you.
>
> Darren,
> First off, thanks a bunch for running through those tests!
>
> > I don't think they are particularly interesting
>
> They *are* interesting in the sense that peaks of 40ish for an 8-way are excellent IMO ;) But yeah, i see what you mean; theres no interesting difference between them.
>
> Do you have any comparative data from runs in -rt1 or earlier kernels?
>
> > perhaps more iterations, say 1000000?
>
> We generally load the system up a lot more (e.g. "make -j 128") or let it run for a long time with the make in a loop (or both). For instance, here is the last data I gathered (with plot attached). This was for a single pass of "make mrproper; make allmodconfig; time make -j 128"
>
> 23.1-rt7
> -----------------
> real 10m49.312s
> user 49m59.674s
> sys 16m15.404s
>
> 26.38 76.02 56.11 1/304 21301
>
> T: 0 ( 5959) P:90 I:100 C:7500381 Min: 2 Act: 2 Avg: 4 Max: 78
> T: 1 ( 5960) P:89 I:200 C:3750191 Min: 2 Act: 3 Avg: 5 Max: 91
> T: 2 ( 5961) P:88 I:300 C:2500127 Min: 2 Act: 3 Avg: 4 Max: 66
> T: 3 ( 5962) P:87 I:400 C:1875096 Min: 2 Act: 5 Avg: 6 Max: 106
> T: 4 ( 5963) P:86 I:500 C:1500077 Min: 2 Act: 6 Avg: 5 Max: 92
> T: 5 ( 5964) P:85 I:600 C:1250064 Min: 2 Act: 5 Avg: 6 Max: 93
> T: 6 ( 5965) P:84 I:700 C:1071483 Min: 2 Act: 3 Avg: 6 Max: 78
> T: 7 ( 5966) P:83 I:800 C: 937548 Min: 2 Act: 5 Avg: 6 Max: 97
>
> 23.1-rt10
> -----------------
>
> real 10m39.046s
> user 56m31.662s
> sys 12m51.158s
>
> 7.64 63.88 62.70 1/291 30774
>
> T: 0 (15455) P:90 I:100 C:8001856 Min: 2 Act: 3 Avg: 4 Max: 65
> T: 1 (15456) P:89 I:200 C:4000928 Min: 2 Act: 3 Avg: 4 Max: 46
> T: 2 (15457) P:88 I:300 C:2667286 Min: 2 Act: 3 Avg: 5 Max: 54
> T: 3 (15458) P:87 I:400 C:2000464 Min: 2 Act: 3 Avg: 5 Max: 46
> T: 4 (15459) P:86 I:500 C:1600372 Min: 2 Act: 3 Avg: 5 Max: 46
> T: 5 (15460) P:85 I:600 C:1333643 Min: 2 Act: 5 Avg: 5 Max: 75
> T: 6 (15461) P:84 I:700 C:1143123 Min: 2 Act: 5 Avg: 6 Max: 52
> T: 7 (15462) P:83 I:800 C:1000232 Min: 2 Act: 4 Avg: 5 Max: 66
>
> 23.1-rt11
> -----------------
>
> real 10m41.610s
> user 55m37.065s
> sys 13m16.394s
>
> 101.92 101.56 61.65 1/308 21349
>
> T: 0 ( 6028) P:90 I:100 C:6584414 Min: 2 Act: 3 Avg: 4 Max: 55
> T: 1 ( 6029) P:89 I:200 C:3292207 Min: 2 Act: 6 Avg: 5 Max: 58
> T: 2 ( 6030) P:88 I:300 C:2194805 Min: 2 Act: 5 Avg: 5 Max: 54
> T: 3 ( 6031) P:87 I:400 C:1646104 Min: 2 Act: 5 Avg: 5 Max: 79
> T: 4 ( 6032) P:86 I:500 C:1316883 Min: 2 Act: 5 Avg: 5 Max: 60
> T: 5 ( 6033) P:85 I:600 C:1097403 Min: 2 Act: 5 Avg: 6 Max: 45
> T: 6 ( 6034) P:84 I:700 C: 940631 Min: 2 Act: 3 Avg: 6 Max: 45
> T: 7 ( 6035) P:83 I:800 C: 823052 Min: 2 Act: 5 Avg: 6 Max: 47
>
> As Steven Rostedt pointed out on IRC, numbers coming from me are suspect ;). But we can at least use them to see if you can repro similar results.
>
> Thanks again!
>
> Regards,
> -Greg
>
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-08 7:34 cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11) Darren Hart
2007-11-08 13:56 ` Gregory Haskins
@ 2007-11-13 13:40 ` Jaswinder Singh
2007-11-13 14:15 ` Steven Rostedt
1 sibling, 1 reply; 10+ messages in thread
From: Jaswinder Singh @ 2007-11-13 13:40 UTC (permalink / raw)
To: Darren Hart; +Cc: RT, Thomas Gleixner, Steven Rostedt, Gregory Haskins
hello Darren,
On Nov 8, 2007 1:04 PM, Darren Hart <dvhltc@us.ibm.com> wrote:
> # ./cyclictest -n -i 10000 -l 10000 -p 95
10000 (10 milliseconds) interval seems to be quite big for current
machine. 10 milliseconds is good for 10 to 15 years old machine but
not for latest machines.
I think we should try -i 1000 or -i 4000 .
Thank you,
Jaswinder Singh.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-13 13:40 ` Jaswinder Singh
@ 2007-11-13 14:15 ` Steven Rostedt
2007-11-13 16:15 ` Darren Hart
0 siblings, 1 reply; 10+ messages in thread
From: Steven Rostedt @ 2007-11-13 14:15 UTC (permalink / raw)
To: Jaswinder Singh; +Cc: Darren Hart, RT, Thomas Gleixner, Gregory Haskins
On Tue, 13 Nov 2007, Jaswinder Singh wrote:
> On Nov 8, 2007 1:04 PM, Darren Hart <dvhltc@us.ibm.com> wrote:
> > # ./cyclictest -n -i 10000 -l 10000 -p 95
>
> 10000 (10 milliseconds) interval seems to be quite big for current
> machine. 10 milliseconds is good for 10 to 15 years old machine but
> not for latest machines.
>
> I think we should try -i 1000 or -i 4000 .
>
heh, I test with -i 250.
-- Steve
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-13 13:30 ` Jaswinder Singh
@ 2007-11-13 14:20 ` Steven Rostedt
2007-11-13 16:19 ` Darren Hart
1 sibling, 0 replies; 10+ messages in thread
From: Steven Rostedt @ 2007-11-13 14:20 UTC (permalink / raw)
To: Jaswinder Singh; +Cc: Gregory Haskins, Darren Hart, RT, Thomas Gleixner
On Tue, 13 Nov 2007, Jaswinder Singh wrote:
> Hello Greg and Darren,
>
> If possible, can you please also add processor ID/number in your
> cyclictest and run more tasks (say 40, 60 or 80 tasks) then we have
> good picture like how many processes are running in which processor
> and how they switching between processors.
>
> And can we force processes to run on specific processor or core.
>
Cyclic test isn't really about testing RT balancing. Even though it does
have an effect. For that type of test you want my rt-migrate-test.c
http://rostedt.homelinux.com/rt/rt-migrate-test.c
or Gregory's (when he fixes it ;-)
http://rt.wiki.kernel.org/index.php/Preemption_Test
-- Steve
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-13 14:15 ` Steven Rostedt
@ 2007-11-13 16:15 ` Darren Hart
2007-11-13 17:13 ` Steven Rostedt
0 siblings, 1 reply; 10+ messages in thread
From: Darren Hart @ 2007-11-13 16:15 UTC (permalink / raw)
To: Steven Rostedt; +Cc: Jaswinder Singh, RT, Thomas Gleixner, Gregory Haskins
On Tuesday 13 November 2007 06:15:03 Steven Rostedt wrote:
> On Tue, 13 Nov 2007, Jaswinder Singh wrote:
> > On Nov 8, 2007 1:04 PM, Darren Hart <dvhltc@us.ibm.com> wrote:
> > > # ./cyclictest -n -i 10000 -l 10000 -p 95
> >
> > 10000 (10 milliseconds) interval seems to be quite big for current
> > machine. 10 milliseconds is good for 10 to 15 years old machine but
> > not for latest machines.
> >
> > I think we should try -i 1000 or -i 4000 .
>
> heh, I test with -i 250.
Someone, I'm sorry I can't recall who atm, suggested that using a larger
interval would allow for more variance to be introduced - not keeping the
caches so hot for this particular test by not spending so much time on the
cpu. Is this a valid approach? Perhaps running multiple runs with both very
tight intervals (like Steve's case) and some longer intervals to ensure we
can handle both cases - since both are common in practice.
Thanks,
--
Darren Hart
IBM Linux Technology Center
Realtime Linux Team
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-13 13:30 ` Jaswinder Singh
2007-11-13 14:20 ` Steven Rostedt
@ 2007-11-13 16:19 ` Darren Hart
1 sibling, 0 replies; 10+ messages in thread
From: Darren Hart @ 2007-11-13 16:19 UTC (permalink / raw)
To: Jaswinder Singh; +Cc: Gregory Haskins, RT, Steven Rostedt, Thomas Gleixner
On Tuesday 13 November 2007 05:30:52 Jaswinder Singh wrote:
> Hello Greg and Darren,
>
> If possible, can you please also add processor ID/number in your
> cyclictest and run more tasks (say 40, 60 or 80 tasks) then we have
> good picture like how many processes are running in which processor
> and how they switching between processors.
Sorry I'm not quite sure what you are requesting.
What do you mean by "add processor ID/number in your cyclictest"?
Are you suggesting running 40-80 copies of cyclictest in parallel?
>
> And can we force processes to run on specific processor or core.
We certainly can... it seems to me that pinning the tasks to CPUs just
reducing scheduling interference and makes the test easier... what are you
trying to accomplish by doing this?
--Darren
>
> Thank you,
>
> Jaswinder Singh.
>
> On Nov 8, 2007 7:26 PM, Gregory Haskins <ghaskins@novell.com> wrote:
> > >>> On Thu, Nov 8, 2007 at 2:34 AM, in message
> >
> > <200711072334.40216.dvhltc@us.ibm.com>, Darren Hart <dvhltc@us.ibm.com>
wrote:
> > > Greg,
> > >
> > > Here are the results I promised you.
> >
> > Darren,
> > First off, thanks a bunch for running through those tests!
> >
> > > I don't think they are particularly interesting
> >
> > They *are* interesting in the sense that peaks of 40ish for an 8-way are
> > excellent IMO ;) But yeah, i see what you mean; theres no interesting
> > difference between them.
> >
> > Do you have any comparative data from runs in -rt1 or earlier kernels?
> >
> > > perhaps more iterations, say 1000000?
> >
> > We generally load the system up a lot more (e.g. "make -j 128") or let
> > it run for a long time with the make in a loop (or both). For instance,
> > here is the last data I gathered (with plot attached). This was for a
> > single pass of "make mrproper; make allmodconfig; time make -j 128"
> >
> > 23.1-rt7
> > -----------------
> > real 10m49.312s
> > user 49m59.674s
> > sys 16m15.404s
> >
> > 26.38 76.02 56.11 1/304 21301
> >
> > T: 0 ( 5959) P:90 I:100 C:7500381 Min: 2 Act: 2 Avg: 4 Max:
> > 78 T: 1 ( 5960) P:89 I:200 C:3750191 Min: 2 Act: 3 Avg: 5
> > Max: 91 T: 2 ( 5961) P:88 I:300 C:2500127 Min: 2 Act: 3 Avg:
> > 4 Max: 66 T: 3 ( 5962) P:87 I:400 C:1875096 Min: 2 Act: 5
> > Avg: 6 Max: 106 T: 4 ( 5963) P:86 I:500 C:1500077 Min: 2 Act:
> > 6 Avg: 5 Max: 92 T: 5 ( 5964) P:85 I:600 C:1250064 Min: 2
> > Act: 5 Avg: 6 Max: 93 T: 6 ( 5965) P:84 I:700 C:1071483 Min:
> > 2 Act: 3 Avg: 6 Max: 78 T: 7 ( 5966) P:83 I:800 C: 937548
> > Min: 2 Act: 5 Avg: 6 Max: 97
> >
> > 23.1-rt10
> > -----------------
> >
> > real 10m39.046s
> > user 56m31.662s
> > sys 12m51.158s
> >
> > 7.64 63.88 62.70 1/291 30774
> >
> > T: 0 (15455) P:90 I:100 C:8001856 Min: 2 Act: 3 Avg: 4 Max:
> > 65 T: 1 (15456) P:89 I:200 C:4000928 Min: 2 Act: 3 Avg: 4
> > Max: 46 T: 2 (15457) P:88 I:300 C:2667286 Min: 2 Act: 3 Avg:
> > 5 Max: 54 T: 3 (15458) P:87 I:400 C:2000464 Min: 2 Act: 3
> > Avg: 5 Max: 46 T: 4 (15459) P:86 I:500 C:1600372 Min: 2 Act:
> > 3 Avg: 5 Max: 46 T: 5 (15460) P:85 I:600 C:1333643 Min: 2
> > Act: 5 Avg: 5 Max: 75 T: 6 (15461) P:84 I:700 C:1143123 Min:
> > 2 Act: 5 Avg: 6 Max: 52 T: 7 (15462) P:83 I:800 C:1000232
> > Min: 2 Act: 4 Avg: 5 Max: 66
> >
> > 23.1-rt11
> > -----------------
> >
> > real 10m41.610s
> > user 55m37.065s
> > sys 13m16.394s
> >
> > 101.92 101.56 61.65 1/308 21349
> >
> > T: 0 ( 6028) P:90 I:100 C:6584414 Min: 2 Act: 3 Avg: 4 Max:
> > 55 T: 1 ( 6029) P:89 I:200 C:3292207 Min: 2 Act: 6 Avg: 5
> > Max: 58 T: 2 ( 6030) P:88 I:300 C:2194805 Min: 2 Act: 5 Avg:
> > 5 Max: 54 T: 3 ( 6031) P:87 I:400 C:1646104 Min: 2 Act: 5
> > Avg: 5 Max: 79 T: 4 ( 6032) P:86 I:500 C:1316883 Min: 2 Act:
> > 5 Avg: 5 Max: 60 T: 5 ( 6033) P:85 I:600 C:1097403 Min: 2
> > Act: 5 Avg: 6 Max: 45 T: 6 ( 6034) P:84 I:700 C: 940631 Min:
> > 2 Act: 3 Avg: 6 Max: 45 T: 7 ( 6035) P:83 I:800 C: 823052
> > Min: 2 Act: 5 Avg: 6 Max: 47
> >
> > As Steven Rostedt pointed out on IRC, numbers coming from me are suspect
> > ;). But we can at least use them to see if you can repro similar
> > results.
> >
> > Thanks again!
> >
> > Regards,
> > -Greg
--
Darren Hart
IBM Linux Technology Center
Realtime Linux Team
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11)
2007-11-13 16:15 ` Darren Hart
@ 2007-11-13 17:13 ` Steven Rostedt
0 siblings, 0 replies; 10+ messages in thread
From: Steven Rostedt @ 2007-11-13 17:13 UTC (permalink / raw)
To: Darren Hart; +Cc: Jaswinder Singh, RT, Thomas Gleixner, Gregory Haskins
On Tue, 13 Nov 2007, Darren Hart wrote:
> On Tuesday 13 November 2007 06:15:03 Steven Rostedt wrote:
> > On Tue, 13 Nov 2007, Jaswinder Singh wrote:
> > > On Nov 8, 2007 1:04 PM, Darren Hart <dvhltc@us.ibm.com> wrote:
> > > > # ./cyclictest -n -i 10000 -l 10000 -p 95
> > >
> > > 10000 (10 milliseconds) interval seems to be quite big for current
> > > machine. 10 milliseconds is good for 10 to 15 years old machine but
> > > not for latest machines.
> > >
> > > I think we should try -i 1000 or -i 4000 .
> >
> > heh, I test with -i 250.
>
> Someone, I'm sorry I can't recall who atm, suggested that using a larger
> interval would allow for more variance to be introduced - not keeping the
> caches so hot for this particular test by not spending so much time on the
> cpu. Is this a valid approach? Perhaps running multiple runs with both very
> tight intervals (like Steve's case) and some longer intervals to ensure we
> can handle both cases - since both are common in practice.
>
I don't think it would hurt to test both cases. Perhaps three kinds.
-i 250
-i 1000
-i 10000
do them separately, and that should give us a good idea of running fast as
well as cache cold.
-- Steve
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2007-11-13 17:13 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-11-08 7:34 cyclic test results on 8 way (2.6.23.1-rt7 through 2.6.23.1-rt11) Darren Hart
2007-11-08 13:56 ` Gregory Haskins
2007-11-08 15:56 ` Darren Hart
2007-11-13 13:30 ` Jaswinder Singh
2007-11-13 14:20 ` Steven Rostedt
2007-11-13 16:19 ` Darren Hart
2007-11-13 13:40 ` Jaswinder Singh
2007-11-13 14:15 ` Steven Rostedt
2007-11-13 16:15 ` Darren Hart
2007-11-13 17:13 ` Steven Rostedt
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.