[Logic Pro] share performance (CPU) data

cyberheater

I’m guessing that Logic Pro will come with default projects and performance meters

Can we compare the same project and cpu usage vs different iPads. Would be good to understand what iPad to get.

dendy

bear in mind that only meaningful way to measure performance is to use max up project with same type of plugin to get into steady 80-90% area and then compare number of plugin instances you can run on other device to get to 80-90%

comparing just projects which may take lets say 20-30% of cpu on different devices has not much information value because if you don’t push high load, CPU turns partially off some cores - as a result of that on two completely different iPads same project may eat same amount of CPU - cause on more powerful one unused cores are put to sleep - they are awaken only when cpu demand grows.

So - pick plugin, add instances to project until you get to 80-90%, count number of instances. Then do same on other iPad - with same plugin (and very inportantly also with same buffer size)

As a better reference you can try with same plugin in different DAWs to see what partnofnload is consumed by DAW itself .. it’s also good check if given DAW uses multicore - to my knowledge only Cubasis is capable to use multiple cores for audio processing which manifests in capabiloty to run MANY more instances of same plugin than in any other DAW - i am prettt curious how will be Logic, i am expecting it will use multicore but we will see ..

cyberheater

We good points Dendy. We can come up with our own benchmark and share the file so it’s an fair comparison

Samu

Just remember that if the 'test project' is created on the most powerful iPad it might choke and crash lesser capable iPads due to memory restrictions etc.

As an example my iPad Air 2 dives to a reboot when I try to load a Cubasis project playing 20+ Model D instances with a bunch of Eventide effects enabled. Another AUM session with 25+ instances of another sample heavy synth plays fine on the M1 but totally trashes the Air 2...

So personally I have like zero interest in benchmarks as long as things play without cracking up

cyberheater

We need a progressive benchmark where more load is processed as time moves on. When it grinds to a halt then that’s the benchmark time.

It would be interesting running the same benchmark on desktop.

dendy

@cyberheater said:
We good points Dendy. We can come up with our own benchmark and share the file so it’s an fair comparison

you don’t need to share project itself.. you just need agree on methodology.. will do some tests for sure, with both Model D as probably most CPU hungry AUv3 plugin and also with some logic stock plugin .. will share my methodology and result numbers for sure, ao anybody can easily replicate test on own device and share their result

cyberheater

@dendy said:

@cyberheater said:
We good points Dendy. We can come up with our own benchmark and share the file so it’s an fair comparison

you don’t need to share project itself.. you just need agree on methodology.. will do some tests for sure, with both Model D as probably most CPU hungry AUv3 plugin and also with some logic stock plugin .. will share my methodology and result numbers for sure !

Thanks Dendy. I have Model D on desktop so will figure out a benchmark and share.

Samu

On Intel Mac Moog Model D is a real pita, barely one instance runs reliably as a plug-in on a 3Ghz 6-Core i5...
...on an M1 Mac no issues what so ever, similar performance as I get on the M1 iPadPro...

cyberheater

@Samu said:
On Intel Mac Moog Model D is a real pita, barely one instance runs reliably as a plug-in on a 3Ghz 6-Core i5...
...on an M1 Mac no issues what so ever, similar performance as I get on the M1 iPadPro...

Thanks Samu. Might be better to use another plug in.

dendy

@cyberheater said:

@Samu said:
On Intel Mac Moog Model D is a real pita, barely one instance runs reliably as a plug-in on a 3Ghz 6-Core i5...
...on an M1 Mac no issues what so ever, similar performance as I get on the M1 iPadPro...

Thanks Samu. Might be better to use another plug in.

if incorrectly understand you are interested in comparing performance of logic on desktop vs on iPad ?

That would be interesting too.. i have MacBook Air M1 and iPad Air M1 so in theory both should perform more or less same .. looking forward for this test, will be interesting !

cyberheater

@dendy said:
if incorrectly understand you are interested in comparing performance of logic on desktop vs on iPad ?

Yes. That's right. We'll stick to Model D.

That would be interesting too.. i have MacBook Air M1 and iPad Air M1 so in theory both should perform more or less same .. looking forward for this test, will be interesting !

Same. I wonder what performance options we have in Logic Pro iPad.

cyberheater

Can folks try this Logic Pro Project on Mac M1 and let me know how they get on?

https://www.dropbox.com/s/w0ldsvkagsjlsdg/Moog Performance Test.logicx.zip?dl=0

Samu

@dendy said:

if incorrectly understand you are interested in comparing performance of logic on desktop vs on iPad ?

Well that could be fun but in practice only be applicable to Mx iPad vs Silicone Macs as there's not too many cross platform AUv3's that work on Intel Mac's and even those that do suffer from pretty miserable performance with the Moog ones being the worst case I've encountered so far.

That would be interesting too.. i have MacBook Air M1 and iPad Air M1 so in theory both should perform more or less same .. looking forward for this test, will be interesting !

That test would definitely make sense but I have a feeling that the MacBook Air M1 would totally crush the M1 iPadPro when it comes to disk-performance while stuff that is kept in RAM should be more or less on-par.

My suggestion here would be to do all bench-marks without any 3rd party plug-ins.
Alchemy in 'Ultra' mode can be real CPU hogger as one example.

Cheers!

TheGarageBandGuide

The Logic Pro benchmark project found here:

https://music-prod.com/logic-pro-benchmarks/

might be worth a try too.

cyberheater

Try this Alchemy test

https://www.dropbox.com/s/w0ldsvkagsjlsdg/Moog Performance Test.logicx.zip?dl=0

ghost_forests

@Samu said:
Just remember that if the 'test project' is created on the most powerful iPad it might choke and crash lesser capable iPads due to memory restrictions etc.

Yup have an M2 Pro and will not be doing this 😂

cazel

released
logic pro for ipad

tja

@cyberheater said:
Try this Alchemy test

https://www.dropbox.com/s/w0ldsvkagsjlsdg/Moog Performance Test.logicx.zip?dl=0

My poor Air 3 just broke into tears, nothing else.

cyberheater

@tja said:

@cyberheater said:
Try this Alchemy test

https://www.dropbox.com/s/w0ldsvkagsjlsdg/Moog Performance Test.logicx.zip?dl=0

My poor Air 3 just broke into tears, nothing else.

Thanks for trying it.

cyberheater

I got to 17 on the Moog test before it crapped out on me

Simon

Logic Pro for iOS - How Does It Run?....