View unanswered posts | View active topics It is currently Thu Oct 23, 2014 6:12 am






Reply to topic  [ 8 posts ] 
nOpcodesPerTimeslice Safety Issues 
Author Message
Rookie

Joined: Mon Jan 10, 2011 11:39 pm
Posts: 16
Post nOpcodesPerTimeslice Safety Issues
How safe is it to adjust the Opcodes per Timeslice?
Because I have run some tests and it seems like the firmware becomes unstable when its adjusted to another value from its original 255.
I have attached a zip file with the raw data and the results.
NOTE: These tests were done using the PC-Emulator and not on a physical microcontroller.

Attachment:
File comment: Opcode raw data and graphs made with Excel.
Opcode Zip.zip [31.27 KiB]
Downloaded 210 times

_________________
Regards
Stefan Andres Charsley
charsleysa@gmail.com


Wed Sep 28, 2011 1:29 am
Profile
Moderator
Moderator
User avatar

Joined: Wed Mar 05, 2008 8:14 am
Posts: 3225
Location: Rotterdam, The Netherlands
Post Re: nOpcodesPerTimeslice Safety Issues
It would seem that it isn't safe :)

- Xander

_________________
| Professional Conduit of Reasonableness
| (Title bestowed upon on the 8th day of November, 2013)
| My Blog: I'd Rather Be Building Robots
| ROBOTC 3rd Party Driver Suite: [Project Page]


Wed Sep 28, 2011 7:12 am
Profile WWW
Rookie

Joined: Mon Jan 10, 2011 11:39 pm
Posts: 16
Post Re: nOpcodesPerTimeslice Safety Issues
UPDATE:

I did the benchmarks on an actual Cortex Microcontroller and the results are so different that you can't compare them!
The PC-Emulator is highly flawed!
Anyway, here are the results for the Cortex, they are very promising, up to 40ms faster than default.

Attachment:
File comment: The Opcode tests done on an actual Cortex Microcontroller.
Opcode Tests.zip [29.49 KiB]
Downloaded 207 times

_________________
Regards
Stefan Andres Charsley
charsleysa@gmail.com


Fri Sep 30, 2011 7:57 pm
Profile
Moderator
Moderator
User avatar

Joined: Wed Mar 05, 2008 8:14 am
Posts: 3225
Location: Rotterdam, The Netherlands
Post Re: nOpcodesPerTimeslice Safety Issues
charsleysa,

Can we see the program you used to do these tests? Data without information on the actual tests is fairly meaningless :)

- Xander

_________________
| Professional Conduit of Reasonableness
| (Title bestowed upon on the 8th day of November, 2013)
| My Blog: I'd Rather Be Building Robots
| ROBOTC 3rd Party Driver Suite: [Project Page]


Sun Oct 02, 2011 1:29 am
Profile WWW
Rookie

Joined: Mon Jan 10, 2011 11:39 pm
Posts: 16
Post Re: nOpcodesPerTimeslice Safety Issues
DOH! Haha, that may be useful :D
The code will run each benchmark 30 times on the Opcode setting you choose in the source code.
The output is a CSV file output format which you just save into a CSV file using windows notepad.
On the PC-Emulator the benchmarks are quick but on the Cortex they may take longer, I estimated
that it takes roughly an hour to wait for the benchmarks to finish and push the data into a file.
Enjoy!

Attachment:
File comment: Source code for the benchmarks, comments in the code should be enough to guide you.
SourceCode2.c [2.36 KiB]
Downloaded 201 times

_________________
Regards
Stefan Andres Charsley
charsleysa@gmail.com


Mon Oct 03, 2011 4:15 am
Profile
Moderator
Moderator
User avatar

Joined: Wed Mar 05, 2008 8:14 am
Posts: 3225
Location: Rotterdam, The Netherlands
Post Re: nOpcodesPerTimeslice Safety Issues
Your program doesn't test reliability, only incrementing a variable. You need to test things like motor control and readings sensors. I'd be far more curious to find out how those aspects are impacted by tweaking that variable.

- Xander

_________________
| Professional Conduit of Reasonableness
| (Title bestowed upon on the 8th day of November, 2013)
| My Blog: I'd Rather Be Building Robots
| ROBOTC 3rd Party Driver Suite: [Project Page]


Mon Oct 03, 2011 4:37 am
Profile WWW
Rookie

Joined: Mon Jan 10, 2011 11:39 pm
Posts: 16
Post Re: nOpcodesPerTimeslice Safety Issues
As I do not have access to a physical cortex I cannot do those tests, especially since I learned that the PC-Emulator is highly flawed. I will try to put together some code for benchmark IO of the Cortex to its peripherals. But until I score a Cortex and the extra equipment I cant benchmark and report back. I will see what I can do about acquiring the equipment but until then I will later post the source code for the new benchmarks.

_________________
Regards
Stefan Andres Charsley
charsleysa@gmail.com


Mon Oct 03, 2011 4:47 am
Profile
Rookie

Joined: Mon Jan 10, 2011 11:39 pm
Posts: 16
Post Re: nOpcodesPerTimeslice Safety Issues
So I whipped up some code in a few minutes, but I think its a bit too runty to show any real results.
Anyway to keep consistency I quickly ran through the benchmarks on the PC-Emulator.
As usual, the results were mostly inconclusive because of the instability of the emulator.

Attachment:
File comment: The graphs for the PC-Emulator benchmarks which are inconclusive.
Opcode Tests.zip [29.04 KiB]
Downloaded 194 times


Attachment:
File comment: Source code for IO test (it was the best I could think of in a few minutes).
SourceCode1.c [2.79 KiB]
Downloaded 192 times

_________________
Regards
Stefan Andres Charsley
charsleysa@gmail.com


Mon Oct 03, 2011 5:37 am
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 8 posts ] 

Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  



Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.