Re: [swarm-hackers] Help on Heatbugs (Git)

From:

Nima Talebi

Subject:

Date:

Thu, 19 Nov 2009 16:02:54 +1100

On Thu, Nov 19, 2009 at 3:41 PM, Scott Christley <address@hidden> wrote:

On Nov 18, 2009, at 8:27 PM, Nima Talebi wrote:

Hi Scott, I've placed an "assert(1==0)" inside the threaded block to make sure it never gets hit - and it doesn't, so it's strange that removing it made things better for you.

I put a breakpoint at the if statement and the debugger stopped there. Then I looked at the detailed command line in the Build Results window it has -DTHREADED=1 in it. So I just put in "#if 0" in place.

Okay I can confirm that it's not running with -DTHREADED=0 ...

Here is the result of Instruments(ObjectAlloc) though; Again it seems to be deep in Swarm... how did you figure out it was the HeatCells (I'll look into this right now)?

...Did you have a look at this screenshot? Notice the largest chunk of memory is in [Value2dDisplay updateDisplay]? I can't see HeatCell anywhere in that deep tree where the most memory is used - The only section from Heatbugs in that tree is the call to [HeatbugsGUI updateGraphicalDisplays].

Nima

Fix the threading problem before you do the instruments otherwise it may give you confusing results because you are creating threads like they are going out of style.

Essentially I just ran the simulation for a short time, then drilled down into the section with the highest MB count, yeah it was pretty deep.

Scott

_______________________________________________
swarm-hackers mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/swarm-hackers

http://ai.autonomy.net.au/People/Nima