I'm using [list-inter] and controlling it with a slider sent through [line 0 20]. Dragging the slider around is giving me CPU thread spikes at like 80% on an i5 quad core at 4.0Ghz and the lists only have 5 floats. Ideally, I'd like to be able to smoothly interpolate these lists with many more numbers (possibly even at audio rate with [line~]).

Just wondering if anyone has any tips for managing this or if I basically need to buy a new CPU/scale back my vision for this application of [list-inter]

@Matt-Stalcup Is this any better on your machine? I used arrays instead of lists to do the computation, precomputed the delta, and didn't use expr. list-inter.pd

Edit: Hmm, isn't it true that sending a float to [line 0 20] does nothing? I think a single float just makes line jump to that value.

Edit 2: yeah, I think you mean this instead:

I think this should be the fastest update rate you can achieve in control domain.

Hey thanks for replying, I'm gonna try out your method here and report back.

I messed up when writing the initial post, I was actually using [line $1 $2] to test things out and wanted to write it just with default arguments to show that I have the issue with defaults. Misremembered the first argument as the initial value instead of ramp length in ms.

Hey so I tried it out and I gotta say, this is a really pretty way to carve out some CPU.

I'm not sure what was going on with my system, but I think when I first made this post pd must have had some persistent hiccup (maybe an issue with connecting/disconnecting from Jack) that was causing the CPU spikes.

I have your solution saved and may return to it later, but for now it's just not much cheaper than [list-inter] so it'd be more of a headache to implement it into what I've already got than to push forward as is.

Sorry to waste your time (and anyone reading this) by making a post before just restarting my computer, but it really did seem like [list-inter] was causing massive spikes. Thanks a bunch, Jamesio!

Dragging the slider around is giving me CPU thread spikes at
like 80% on an i5 quad core at 4.0Ghz and the lists only have 5 floats.

this is something else what casuses this, most likely in the graphics thread.

your computer can easily move some 50,000 messages per scheduler tick (1 ms?) between 2 pd externals.

try doing the same with 200 numbers instead of 5 and you will see what is going on.

@obineg, to your point: Adding a ton of numbers to my lists doesn't seem to increase load that much. Load increases a lot when I am running multiple instances of [line-inter] simultaneously. My goal is to interpolate between 4 lists plotted at corners of an xy field, which I can do but it turns out actually does spike my CPU to like 98 percent on Purr-Data's thread. Currently trying to hook up @jameslo's solution and see if it takes the load off.

My issue there is that I'd like to be able to dynamically set the number of floats in the array, so that's holding me back presently. Is it as simple as putting a $1 at the end of that array object? I don't know! Here I go to find out.

@Matt-Stalcup said:

Is it as simple as putting a $1 at the end of that array object?

Yes. You can also use [array size <arrayName>]--see the array object help file for more details.

Edit: Here, this should be a drop-in replacement if I haven't misunderstood how list-inter works. myList-inter.zip

Oh my god it was because I was printing the lists. The moment I cut the cord to [print], i went from 97% on a single thread to >3% for all threads combined.

Thanks everybody for helping in this thread. I appreciate that the pd forum is a place where I can come to learn without worrying too much about having to know everything already.

@Matt-Stalcup I remember learning that lesson, still have to relearn it on occasion. Gave me a chuckle.