[PD] profiling patches and benchmarking platforms for Pd

Dan Wilcox danomatika at gmail.com
Fri May 15 12:26:40 CEST 2020


I think it will also be difficult to compare systems, largely because it also depends on what you are doing in your patches. Some systems may have a fast enough CPU but slow file access which will only bother you if you are doing a lot from disk, etc. OTOH maybe this is something solved by a library that we could leverage? IOhannes is right in that it depends on what you really measure and how. I don't think it's as simple as a focused metric like a GPU or file transfer metric.

This is not a real benchmark, but I can share what I do.

In general, I have done simple comparisons by running the same patch on my desktop and whatever mobile or embedded device I'm using for robotcowboy. I then have it spit out the basic CPU load percentage as calculated by the test patch included with Pd. That helps me have a general understanding that a song may take about 8-10x more CPU percentage on mobile as on the development machine and simply have another number with this applied. It's really simple & not refined but works well enough. Then again, what I am doing with Pd is in a lot of ways simple as compared to what many on this list do.



> On May 15, 2020, at 12:00 PM, pd-list-request at lists.iem.at wrote:
> 
> i think it's hard to impossible to find any meaningful benchmark.
> which kind of numbers do you expect in your textfile, and how would you
> compare them?

--------
Dan Wilcox
@danomatika <http://twitter.com/danomatika>
danomatika.com <http://danomatika.com/>
robotcowboy.com <http://robotcowboy.com/>



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20200515/00617399/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen Shot 2020-05-15 at 12.25.16 PM.png
Type: image/png
Size: 9606 bytes
Desc: not available
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20200515/00617399/attachment.png>


More information about the Pd-list mailing list