I've been building a patch that has several looping stations, sequencers, and effects all coinciding nicely. But, my drum and synth sequencers have been lagging quite regularly. Whilst trying different things to fix it I changed my sampling rate to 96k, and boom, no more lagging. BUT then my DSP cpu usage is doubled, making my patch run almost 80%... Afterwords, I tried switching back and forth between 96k and 44.1k to see if it was just needing to be reset or something, but it was still lagging just the same.

So, I could live at 96k, but would rather save on CPU and make my patch bigger.

I did some additional troubleshooting by making a few other small patches just to test this issue, and the same timing problems occurred.

-If anyone else has had an issue like this/has resolved an issue like this, I'd appreciate the input!

I've been building a patch that has several looping stations, sequencers, and effects all coinciding nicely.  But, my drum and synth sequencers have been lagging quite regularly.  Whilst trying different things to fix it I changed my sampling rate to 96k, and boom, no more lagging.   BUT then my DSP cpu usage is doubled, making my patch run almost 80%...  Afterwords, I tried switching back and forth between 96k and 44.1k to see if it was just needing to be reset or something, but it was still lagging just the same.

-If anyone else has had an issue like this/has resolved an issue like this, I'd appreciate the input!


lagging-at-44-1k

The obvious thing(s) to try are to try reduce the i/O vector size and signal vector size (rather than sample rate) which will also have a lpotentially significant effect on latency (and CPU usage as well).

Also, if t is possible, try to use poly~ where you have multiple identical voices going if you have not already done so. poly~ also enables patches to run at greater or lesser sample rates where that might be practical.

The obvious thing(s) to try are to try reduce the i/O vector size and signal vector size (rather than sample rate) which will also have a lpotentially significant effect on latency (and CPU usage as well). 

Also, if t is possible, try to use poly~ where you have multiple identical voices going if you have not already done so. poly~ also enables patches to run at greater or lesser sample rates where that might be practical.


Thanks for the reply! Reducing the size of the signal vector made my patch keep perfect time. So, thank you very very much!

Also, if you or anyone else knows of any good reading related to optimizing one's CPU usage in Max/making patches run more smoothly in general, I'd greatly appreciate it being sent my way.

Thanks for the reply!  Reducing the size of the signal vector made my patch keep perfect time.  So, thank you very very much!  

Lagging at 44.1k