Go to Post 2015 never happened - Chris is me [more]
Home
Go Back   Chief Delphi > Technical > Control System > FRC Control System
CD-Media   CD-Spy  
portal register members calendar search Today's Posts Mark Forums Read FAQ rules

 
Reply
Thread Tools Rate Thread Display Modes
  #1   Spotlight this post!  
Unread 11-01-2012, 01:51
wireties's Avatar
wireties wireties is offline
Principal Engineer
AKA: Keith Buchanan
FRC #1296 (Full Metal Jackets)
Team Role: Mentor
 
Join Date: Jan 2006
Rookie Year: 2004
Location: Rockwall, TX
Posts: 1,168
wireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond repute
Send a message via AIM to wireties
relative throughput of new/old cRIOs

I have heard that the 4-slot cRIOs are faster. Is this true? Anyone know how much faster?
Reply With Quote
  #2   Spotlight this post!  
Unread 11-01-2012, 01:55
EricVanWyk EricVanWyk is offline
Registered User
no team
 
Join Date: Jan 2007
Rookie Year: 2000
Location: Boston
Posts: 1,597
EricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond repute
Send a message via AIM to EricVanWyk
Re: relative throughput of new/old cRIOs

They have more memory and a better cache. It makes some use cases faster, but it isn't huge.
Reply With Quote
  #3   Spotlight this post!  
Unread 11-01-2012, 10:10
Racer26 Racer26 is offline
Registered User
no team
Team Role: Alumni
 
Join Date: Apr 2003
Rookie Year: 2003
Location: Beaverton, ON
Posts: 2,229
Racer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond repute
Re: relative throughput of new/old cRIOs

Not enough to matter, not unless you're doing seriously high end trigonometry, triangulation, vision, and a bunch of interrupting.

Until 2009, we ran on an 8-bit microcontroller that had 4kB of RAM and 128kB of program space, operating at 10MIPS, and we did vision with that.

The cRIO runs at something on the order of 400MIPS, and has far more RAM/Program Space. If any of us start saturating its capabilities, I'll be amazed.
Reply With Quote
  #4   Spotlight this post!  
Unread 11-01-2012, 11:25
wireties's Avatar
wireties wireties is offline
Principal Engineer
AKA: Keith Buchanan
FRC #1296 (Full Metal Jackets)
Team Role: Mentor
 
Join Date: Jan 2006
Rookie Year: 2004
Location: Rockwall, TX
Posts: 1,168
wireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond reputewireties has a reputation beyond repute
Send a message via AIM to wireties
Re: relative throughput of new/old cRIOs

Quote:
Originally Posted by 1075guy View Post
Not enough to matter, not unless you're doing seriously high end trigonometry, triangulation, vision, and a bunch of interrupting.
like continuously calculating trajectory? continuously tracking a target using vision algorithms? running 6-8 PID loops at 50Hz? get the point?

Quote:
Originally Posted by 1075guy View Post
Until 2009, we ran on an 8-bit microcontroller that had 4kB of RAM and 128kB of program space, operating at 10MIPS, and we did vision with that.
it was the 16-bit PICC I think and the vision code was so slow as to be almost unusable - and the camera did some of the work back then

Quote:
Originally Posted by 1075guy View Post
The cRIO runs at something on the order of 400MIPS, and has far more RAM/Program Space. If any of us start saturating its capabilities, I'll be amazed.
be amazed, have you seen all the other crap running on that processor? but I basically agree, it should be enough.
Reply With Quote
  #5   Spotlight this post!  
Unread 11-01-2012, 14:16
EricVanWyk EricVanWyk is offline
Registered User
no team
 
Join Date: Jan 2007
Rookie Year: 2000
Location: Boston
Posts: 1,597
EricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond reputeEricVanWyk has a reputation beyond repute
Send a message via AIM to EricVanWyk
Re: relative throughput of new/old cRIOs

If you want to get in to the gritty details, you can start here .

The FRC cRIO-II has twice the memory: twice the RAM, twice the FLASH, and twice the on-die cache in the processor. However, raw execution speed in the cores are (roughly) identical: both run at 400MHz and can issue up to 3 instructions per clock.

The extra memory is good for vision, but won't really affect things like trajectory calculations and PID loops. Simply put, anything that the PIC could do is trivially small on the new system. Bumping the number of PID loops from 8 to 80 would have no impact on performance, but bumping the image size from 160x120 to 640x480 certainly will.
Reply With Quote
  #6   Spotlight this post!  
Unread 11-01-2012, 14:35
Racer26 Racer26 is offline
Registered User
no team
Team Role: Alumni
 
Join Date: Apr 2003
Rookie Year: 2003
Location: Beaverton, ON
Posts: 2,229
Racer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond reputeRacer26 has a reputation beyond repute
Re: relative throughput of new/old cRIOs

Quote:
Originally Posted by wireties View Post
it was the 16-bit PICC I think and the vision code was so slow as to be almost unusable - and the camera did some of the work back then
It was a PIC18F8722. The specs I quoted came straight from Microchip.

You are correct that the camera itself did a bunch of the work, but the imaging code wasn't particularly slow. It kept up with the camera framerate.
Reply With Quote
Reply


Thread Tools
Display Modes Rate This Thread
Rate This Thread:

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

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 21:23.

The Chief Delphi Forums are sponsored by Innovation First International, Inc.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi