Wiki page : Plugins CPU Usage

[jon]

@rcf

For those, wouldn’t it be advisable to state the length of reverb IR used? Surely a loaded 15 secs IR is going to use more resources than a 3.5 secs one? Please ignore me if I’m wrong about that.

You are right about this.
Maybe it would be worth having a column with “extras” or “comments” where info like this would be dropped if needed(?)

[semmo]

that’s strange that most plugins cost the same CPU when turned on and off

[Rom]

It seams like the plugins don’t use the LV2 bypass described by @falkTX, am I right ?

[funkypou]

CPU usage with Convolution reverb

this is not so much related to cpu usage, but jump in them. cpu load jumping from 75% to 96% is certainly not normal on a full realtime system. the load is meant to be consistent, assuming the PB is also not changing too (bypassing some plugins can reduce load if they implement a custom bypass, like it is the case of chow centaur). this is definitely a bug that needs to be dug into and resolved.

[eggsperde]

@funkypou

That would be awesome. I took this initiative because there seemed to be demands from users and we had no better alternative. But I would love to see a more accurate and automatic process to collect these data though I don’t have the skills to create this tool.

If @falkTX lends me a hand wrt to the plugin-specific interactions, I would give it a go. For each plugin a JSON file with the pedalboard description would have to be generated and then loaded. After this happened, the plugin parameters would have to be changed and CPU consumption measured. If said parameters are exposed to Python I foresee no problems (although there certainly are some).

[CarloDossi]

In section simulator i’ve added the estimation for Amps.
I’ve reported the values of the same plugin in 2 row if it changes cpu usage in diffrent modality (clean/lead).
Titan had too much combination so i gived up
For Cab plugins I was unsure since some has different level of fidelty… maybe different raw for different level?

[funkypou]

Thank you very much for your contribution, @CarloDossi. You filled one of the most important categories. I suspected some plugins would use different amounts of CPU depending on their settings. That’s a good idea to create several lines with the general configurations that have an effect on CPU usage (clean, lead, drive on).

And thanks to you, I am also pleased that I chose the Onyx as my main amp as it seems to be the less consuming among the versatile amps and I am satisfied with the quality of the varied tones (clean, crunch, heavy rock) I get from it.

@Rom

I made measurements with the delays and Utilities I have installed

Hi @Rom,
I’m taking advantage of this post to thank you as well for the important number of lines you filled in these two categories.

You guys are awesome.

[oeSmash]

Running effects in parallel when possible saves CPU, and you can see on that board above that only one of the chains out of the toggle-switch actually goes through the amp/cab, the other ones are so heavily doped up in reverb/delay etc. that they sound fine without running through an amp/cabsim. I’m also using the Alembic amp, which is lower CPU (on a new board I tried the Roamer, sounds great but I think just uses too much so going back to Alembic). There are also some plugins that really give you a lot of bang for your buck… for example the Floaty is on most of my boards (hidden by the sequencer) and really can do a lot with some tweaking… Some of the pedals are worth the high CPU (brummer’s distortion, for one…) but it definitely takes trial and error to find your own system of pedalboard construction-- don’t give up on getting everything you want in there though, there’s likely a way to do it…

[CarloDossi]

Making an example using “drop” and “vintage cabinets”:

Fiedelty normal (both):

• Drop: 25%
• Vintage cabinet: 19%
• Parallel: 24% cpu
• Series: 32% cpu
• Absolute difference: 8%
• Relative difference (to series value): 25%

Fiedelty High (both):

• Drop: 35%
• Vintage cabinet: 32%
• Parallel: 37% cpu
• Series: 58% cpu
• Absolute difference: 21%
• Relative difference (to series value): 36%

As you can see the diffference is noticeable and increase with level of fiedelty (or cpu comsumption), and as said before the wiki page cpu level are indicative and are really valid only in that specific experimental context.

[Lyford_Rome]

Had a chance to update the chart with some MOD DUO values. Don’t own some of the paid plugins - would the trial versions be comparable?

Interesting to see how the older CPU compares with the Dwarf, especially noticeable with the amp sims!

[Rom]

Nice ! Thanks !!!

@Lyford_Rome

Don’t own some of the paid plugins - would the trial versions be comparable?

I think they are exactly the same, and I don’t think the mute generator done by the OS is much CPU consuming !