Welcome, Unregistered.

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

Reply
Old 26-Sep-2012, 10:53   #1
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default Android memory bandwidth benchmark

My second benchmark for Android is now also live on Google Play here. It is a simple memory bandwidth test based upon the classic STREAM benchmark. I made some modifications, such as using pthreads instead of OpenMP.

For best performance, try using more threads than number of cores.

On a Snapdragon S3 dual-core, I get about 1.5 GB/s of copy bandwidth.

Link to app: https://play.google.com/store/apps/d...vine.rgbenchbw

Please give your feedback. This app is not 100% consistent in different runs, but it is still good for getting a good ballpark estimate

Last edited by codedivine; 26-Sep-2012 at 11:14.
codedivine is offline   Reply With Quote
Old 26-Sep-2012, 17:49   #2
ToTTenTranz
Senior Member
 
Join Date: Jul 2008
Posts: 2,146
Default

AT&T One X

4 threads:
Copy 2274 MB/s
Scale 2375 MB/s
Add 2440 MB/s
Triad 2476 MB/s

Less than 4 threads gives very inconsistent results between runs, usually lower than the above, and for 8 threads the result is about the same as 4 threads.
ToTTenTranz is offline   Reply With Quote
Old 26-Sep-2012, 17:57   #3
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Quote:
Originally Posted by ToTTenTranz View Post
AT&T One X

4 threads:
Copy 2274 MB/s
Scale 2375 MB/s
Add 2440 MB/s
Triad 2476 MB/s

Less than 4 threads gives very inconsistent results between runs, usually lower than the above, and for 8 threads the result is about the same as 4 threads.
Thanks! I guess that's the Snapdragon S4?
Oh btw I have just pushed a v1.1 of the benchmark to the market. It is the same benchmark code, but just adds an "Auto mode" to the UI that automatically tests all thread configs and just displays results for the one with best copy bandwidth
codedivine is offline   Reply With Quote
Old 27-Sep-2012, 00:52   #4
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Pushing an update to the market that does slight changes to the kernel, which I *think* should reduce variability a little bit. The update should show up as v1.2 in an hour or so.

Also, for reference purposes, I took my C code, added a bit of code to convert it into a Linux command line app and measured performance. On my Phenom II X4 desktop, which has 17GB/s of bandwidth theoretically, I got 9 GB/s of copy bandwidth peak using my benchmark.

This was almost the same as that reported by the standard STREAM benchmark using default values so I suppose my benchmark does do its job.
codedivine is offline   Reply With Quote
Old 27-Sep-2012, 09:19   #5
Rys
Tiled
 
Join Date: Oct 2003
Location: Kings Langley, UK
Posts: 2,675
Default

Are you planning to add more bandwidth tests to try and probe what's capable under a number of different memory access scenarios? STREAM is fairly narrow in its assessment.
__________________
A major redesign of the core ALU pineapple boomerang fortress.
Rys is offline   Reply With Quote
Old 27-Sep-2012, 15:06   #6
Nigma
Registered
 
Join Date: Mar 2012
Posts: 1
Default

How does this factor in cache sizes ?... STREAM allows you to configure array sizes.
Nigma is offline   Reply With Quote
Old 27-Sep-2012, 18:16   #7
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Quote:
Originally Posted by Rys View Post
Are you planning to add more bandwidth tests to try and probe what's capable under a number of different memory access scenarios? STREAM is fairly narrow in its assessment.
Not sure Rys. What would you recommend? People also like using LMBench for testing latency, but that seems like a lot of work to port

Quote:
Originally Posted by Nigma View Post
How does this factor in cache sizes ?... STREAM allows you to configure array sizes.
Good question. Current version is set to use 8MB arrays (3 of them, so total 24MB usage). I think I can provide that as a user-configurable setting in the application.
codedivine is offline   Reply With Quote
Old 27-Sep-2012, 18:28   #8
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Some short results collected from the webs, take with grain of salt:

1. Nexus 7: About 1.3 GB/s of peak copy bandwidth reported. Much lower than what I was expecting.

2. Atrix (Tegra 2): About 900-1GB/s of peak copy bandwidth.

3. Totally unconfirmed: Exynos 4 Quad about 2.3 GB/s of peak copy bandwidth.
codedivine is offline   Reply With Quote
Old 28-Sep-2012, 03:50   #9
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Posted major update. Now I have added a simple "memcpy" benchmark in addition to STREAM based ones. The app by default now runs memcpy and you should just "Misc" if you want to test STREAM.

On my phone, on memcpy I am getting about 1.8GB/s on memcpy compared to the lower 1.5GB/s I was getting on the STREAM based benches.

If you have 10-15 seconds, please do try it out and post your results. Thanks

Last edited by codedivine; 28-Sep-2012 at 04:25.
codedivine is offline   Reply With Quote
Old 28-Sep-2012, 21:34   #10
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

A user with Snapdragon S4 reported almost 5GB/s on the memcpy benchmark ("simple" under the new app).

Can anyone with a S4 confirm the result?
codedivine is offline   Reply With Quote
Old 01-Oct-2012, 00:37   #11
cal_guy
Member
 
Join Date: Jun 2008
Posts: 155
Default

HTC Desire: 582MB/s Simple, 648MB/s Misc.
cal_guy is offline   Reply With Quote
Old 01-Oct-2012, 21:27   #12
swaaye
Entirely Suboptimal
 
Join Date: Mar 2003
Location: WI, USA
Posts: 6,845
Default

Nook Color running Cyanogenmod 7.2
1100 MHz conservative gov

719MB/s simple

Misc
498.9MB/s copy
511.4MB/s scale
465.3MB/s add
438.8MB/s triad

I did 3 runs of both test modes and results were pretty close each time.
swaaye is offline   Reply With Quote
Old 01-Oct-2012, 21:40   #13
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Thanks cal_guy and swaaye
codedivine is offline   Reply With Quote
Old 09-Oct-2012, 16:50   #14
Rys
Tiled
 
Join Date: Oct 2003
Location: Kings Langley, UK
Posts: 2,675
Default

Can you provide an APK version? I've got some devices here that I can't sign in to the Play store with....
__________________
A major redesign of the core ALU pineapple boomerang fortress.
Rys is offline   Reply With Quote
Old 09-Oct-2012, 17:56   #15
ToTTenTranz
Senior Member
 
Join Date: Jul 2008
Posts: 2,146
Default

Quote:
Originally Posted by codedivine View Post
A user with Snapdragon S4 reported almost 5GB/s on the memcpy benchmark ("simple" under the new app).

Can anyone with a S4 confirm the result?
My result is from a S4. Maybe that was a S4 Pro? Oppo Finder 5, Optimus G and some other top-end smartphones are coming with with the quad-core Krait + Adreno 320, and maybe that chip has bandwidth improvements.
ToTTenTranz is offline   Reply With Quote
Old 09-Oct-2012, 19:23   #16
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Quote:
Originally Posted by Rys View Post
Can you provide an APK version? I've got some devices here that I can't sign in to the Play store with....
PMed.

Quote:
Originally Posted by ToTTenTranz View Post
My result is from a S4. Maybe that was a S4 Pro? Oppo Finder 5, Optimus G and some other top-end smartphones are coming with with the quad-core Krait + Adreno 320, and maybe that chip has bandwidth improvements.
No, I had updated the test after you ran them thus the results are different. Essentially, I have added a simple "memcpy" test which simply benches the "memcpy" call in C. The results you had are from my STREAM variant. Apparently there is a big difference on S4, no idea why.
codedivine is offline   Reply With Quote
Old 10-Oct-2012, 01:18   #17
ToTTenTranz
Senior Member
 
Join Date: Jul 2008
Posts: 2,146
Default

Yep, I get ~5,1GB/s in the updated version.
ToTTenTranz is offline   Reply With Quote
Old 11-Oct-2012, 16:48   #18
Rys
Tiled
 
Join Date: Oct 2003
Location: Kings Langley, UK
Posts: 2,675
Default

Simple 5376.8 (this varies quite a bit, run to run).
Copy 3496.5
Scale 3674.6
Add 4031.2
Triad 4055.5

Qualcomm APQ8064, no power saving.
__________________
A major redesign of the core ALU pineapple boomerang fortress.
Rys is offline   Reply With Quote
Old 11-Oct-2012, 22:25   #19
Zaphod
Remember
 
Join Date: Aug 2003
Posts: 2,031
Default

OMAP 4460 (Huawei Ascend P1): Results are not consistent with simple/auto. On average it yields ~1250 MB/s, but on a couple of runs I've seen as high as 1370 MB/s and as low as 1120 MB/s. Power saving on or off (which seems to impact max CPU speed) makes no discernible difference.
Zaphod is offline   Reply With Quote
Old 12-Oct-2012, 13:50   #20
Gubbi
Senior Member
 
Join Date: Feb 2002
Posts: 2,544
Default

LG Optimus 2X (Tegra 2):
Simple, auto: 1198.0 MB/s

Misc, auto:
Copy: 945.2 - 1054 MB/s
Scale: 891.5 - 1025 MB/s
Add: 725.3 - 834 MB/s
Triad: 711.1- 828 MB/s


Cheers
__________________
I'm pink, therefore I'm spam

Last edited by Gubbi; 12-Oct-2012 at 13:58.
Gubbi is offline   Reply With Quote
Old 12-Oct-2012, 16:32   #21
codedivine
Member
 
Join Date: Jan 2009
Posts: 215
Default

Thanks everyone!
codedivine is offline   Reply With Quote
Old 12-Oct-2012, 21:50   #22
Shifty Geezer
Grumpy Mod
 
Join Date: Dec 2004
Location: In a pretty pink padded cell
Posts: 25,990
Default

ASUS Transformer TF101 (Tegra 2)
Auto threads :
simple - ~1060 MB/s
misc -
copy ~660
scale ~660
add ~ 710
triad ~710

Incompatible with Samsung Galaxy Ace
__________________
Shifty Geezer
...

Tolerance for internet moronism is exhausted. Anyone talking about people's attitudes in the Console fora, rather than games and technology, will feel my wrath. Read the FAQ to remind yourself how to behave and avoid unsightly incidents.
Shifty Geezer is online now   Reply With Quote
Old 13-Oct-2012, 05:18   #23
Rurouni
Member
 
Join Date: Sep 2008
Posts: 131
Default

Huawei Mediapad (oc'd S3@1.51Ghz)
Simple: ~1500
Misc: 1400-1600

The S4(pro?) benches are mouth watering.
Rurouni is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 15:08.


Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.