checking CPU load of parts of the patch

k9's icon

My patch is growing toward 50% CPU and I am working with a bunch of objects that I am quite unfamiliar with.

what is the best way to find out where the CPU load in the patch is happening? simply deleting objets? or is there a more elegant way?

I came across the "probing" menu entry in the Debug menu, but I couldn't find it in the documentation and it didn't make much sense poking around with it. Is this the way to go?

what a luxury it would be if maxobjects.com had an indication of how much CPU a given object uses...

any advice greatly appreciated!

Kai

Roman Thilenius's icon

the problem is that it is not so simple that you could say object
[cycle~] uses 0.0057% in order to have some kind of comparison.

some objects start using the more CPU the more intances you run.

try making copies of cycle~ and see how much they use and then try
the same with zerox~ and you will see what i mean.

of course it can also depend on CPU type, OS, background apps
and last but not leat about the current CPU requirement for
message stuff in the max runtime.

are you using specific objects maybe? that is where i would start
optimising beore i count my line~s ...

-110

.

Matthew Aidekman's icon

david posted a tester a long time ago I wish I knew the right search terms. something about

Max Patch
Copy patch and select New From Clipboard in Max.

try this

Emmanuel Jourdan's icon

I often use Apple's Shark to look at my patches. It usually tells you how much percentage each (mainly DSP) object takes. Of course if you're just wondering if one instance of trigger takes more CPU than one of bangbang that's another story

tep's icon

While this thread is about "parts of a patch", is there a way to optimize cpu by making some parts of the patch "inactive" when not used ?

I'm using rather old computers with Tuio objects and the Reactivision software (which is a big cpu-eater) so it requires the Maxmsp patches to be as light as possible ...

First i put some gates everywhere but it doesn't seem to make a big difference. I wonder if there's a way to on/off in a more effective way ...