-
lacuna
Maybe the answer is in this video of Millers classes at 1:04:00
https://msp.ucsd.edu/syllabi/206.20s/movies/5b.may1.mp4
sorting starts at 59:19:15
overview: https://msp.ucsd.edu/syllabi/206.20s/index.htm -
lacuna
A comma in a message means "now comes another message"
so [1, 2(
is the same as
[t b b]
| |
[2( [1(So you could separate your two messages to subpatch with a comma instead of writing ;subpatch again. (where the semicolon means "send to")
-
lacuna
@jameslo And when filtering zero, there are different timings at different samplerates and dsp on/off also differs.
At 48kHz samplerate and dsp off, lowest is 4 ms but very very rarely goes as low as 2.66667 ms. -
lacuna
@jameslo wtf I expected 64 samples. It can go down to zero:
slider-min-update-rate.pd
Is this related related? [Pd-list] pd-gui update rate (2011)
https://lists.iem.at/hyperkitty/list/pd-list@lists.iem.at/thread/NFL5EIGEJQII7G5X5FXGM2WXDM6W4263/#KJSQGTRBJVBZGMQPHFYRGYBQXN7GSOZL -
lacuna
Press control+shift+escape to open the task manager, in details you can right click on wish86.exe and or pd.exe and define the allowed cores for the application.
I never tryed this but just saw it there now.
As @whale-av said, maybe you have to copy the pd folder and maybe rename the second application.
I guess there are other ways to start and run applications on a specific thread in win, but you can search for yourself as this is not Pd specific.And as @whale-av said, these instances are completly independent and they
won't run in sync. They run asynchronous. Useful for special tasks, for example to do sth with fast-forward-message while the other patch stays in real-time.If you want to spread patches across different cpu cores but still in sync, use the [pd~] object.
Either way shmem lib is very useful to share memory / share arrays between pd-instances.
(Looking at the task-manager when using [pd~] I see the cores are constanly changing, Looks like Windows is dynamicly changing the cores. Might be good enough or even better to let Windows manage it instead of assigning fixed threads. Same might be true when running asynchronous instances. )
-
-
-
lacuna
Updated the patches:
Fixed 2 bugs:
Whole array size as default output.
Now also works if array-size changed.And cleaned up messy counter.
Added array-sort example to helpfile, changed thread title.
@ddw_music This was also my second try, giving up the first. And my thoughts where similar, especially if LUA would be handy here? I did not care too much about speed as I don't need this for realtime. Anyway an object written in C would be faster.
(This array-sort is much faster than [list-abs/list-sort], did not try [text]sort.)
Still I don't understand your idea of building a list without rescanning the array or list for each peak? But don't worry if you are done with it ... If I only had known how much time I spend with this .... -
lacuna
@ddw_music Oh! Thank you! Yes, a minimal change in loading the 4th argument caused the bug! Moses! I might not have tested before uploading... blushing!!! Now I reuploaded it. Thank you
And this is quite slow... it is actually very slow. Wondering about a different approach. -
lacuna
Vanilla abstraction, made with [array-max] and [array-min] by nulling found peak and run again.
array-maxx.pd
array-maxx-help.pd
array-minn.pd