Live 11 offers less than 50% of Live 10

[properLofi]

I was noticing high CPU usage so I decided to carry out a performance test comparison using the Live 10 performance test set.

Without going into all the details as I’ll post those if a relevant thread starts, on Live 10 I could duplicate the set 5 times without issue and the CPU crapped out on the 6th. That’s at least 80 tracks without maxing out the CPU.

On Live 11 I got CPU overloads as soon as I pressed play and then it settled down to an above standard 37% CPU. It crapped out on the 3rd duplication of the set!

That’s a drop from 80 stable tracks to 32 tracks! Where the hell have those other 46 tracks gone?

MacBook Pro 15” Retina (mid 2012). 2.6GHz Quad-Core i7. 16GB 1600 DDR3. Intel HD Graphics 4000. MacOS 10.15.7 Catalina.

[Calagan]

Are you using an external monitor or egpu ?
It seems there is some issues with this setup...

[wildcon]

Hi,

Yes I’ve noticed something similar.

Not had that many tracks going but huge cpu usage from just a few tracks. Interestingly it doesn’t seem to increase related to the amount of tracks added. For instance I had 35% with 3 tracks!! But was able to add a lot more without too much of an increase. Adding MFL MPE device totally freaked it and it instantly hit 100%

I was using an external monitor and have noticed a lot of glitching when in the piano roll view!

[Calagan]

Another thing to watch is Live indexing.

If you just installed Live 11, I think it's indexing a whole lot of stuff, and until this is finished it can be pretty unstable and CPU-heavy...
Are you sure the indexing is finished ?

[perrycomo]

Same here, dismayed at how little 11 can handle before glitching out everywhere

[properLofi]

I’m not using any external displays or GPU. My system is exactly as stated in the OP and to reduce confusion that’s with no controllers or external sound cards attached.

Live 11 has been installed since release and I’ve completed a whole project so I’d hope initial indexing has finished.

The main reason for carrying out the test is that without a change of workflow or utilising the more system intensive additions to 11 I was seeing abnormal CPU loads compared to 10.

That’s why I went back to a known benchmark process. It seems like either Ableton have chosen to provide less optimised performance on older processors by some large degree or they’ve released a product that hasn’t been thoroughly tested. I just don’t get why the underlying engine would demand so much more CPU load if your not using the newer functionality of 11.

It’s also odd that others aren’t more vocal about this. I can’t be the only person to be seeing issue as not everyone has new hardware given 10s reliability on older machines.

[kitekrazy]

This seems to be a trend in any software development. We have more powerful systems and developers want to rape as much resources as possible. That indexing is annoying and the more you add to it so does it take up drive space.

[[jur]]

It's been tested thoroughly and shouldn't be the case.
Might be some issue with a device or something like that.
Which one are you using?

Someone mentionned MPE, if you'e using the MPE Control device can you check something please:
open the builtin Max, not from inside Live but by going into Live's "package content" in the FInder/Explorer > navigating to the Max app
Once Max is open can you check its Debug menu and see if Event/SIgnal/Matrix Probe is active or not please? If anything is active, turn it off, quit Max, relaunch Live and re-do your test.

[Jed]

Same problem here with a 10th gen Intel processor

Cheers
Jed

[properLofi]

It's been tested thoroughly and shouldn't be the case.
Might be some issue with a device or something like that.
Which one are you using?

Someone mentionned MPE, if you'e using the MPE Control device can you check something please:
open the builtin Max, not from inside Live but by going into Live's "package content" in the FInder/Explorer > navigating to the Max app
Once Max is open can you check its Debug menu and see if Event/SIgnal/Matrix Probe is active or not please? If anything is active, turn it off, quit Max, relaunch Live and re-do your test.

As I said, I’ve got nothing attached to the laptop, no controllers, no interface, no screen, no external GPU. It’s a vanilla MacBook Pro that was rock solid and churned through tracks in Live 10 and just feels crippled by 11.

When I do attach my interface, keyboard, push etc nothing changes with regards to the performance difference. 11 is a massive resource hungry beast and I’m worried about upgrading hardware if I’m going to see the same issue. Even in this one thread people have identified I’m not alone so something is going on if 10th Gen CPUs are being hit as well.

[Mark Williams]

No issue on my Ryzen 3600x

[[jur]]

It's been tested thoroughly and shouldn't be the case.
Might be some issue with a device or something like that.
Which one are you using?

Someone mentionned MPE, if you'e using the MPE Control device can you check something please:
open the builtin Max, not from inside Live but by going into Live's "package content" in the FInder/Explorer > navigating to the Max app
Once Max is open can you check its Debug menu and see if Event/SIgnal/Matrix Probe is active or not please? If anything is active, turn it off, quit Max, relaunch Live and re-do your test.

As I said, I’ve got nothing attached to the laptop, no controllers, no interface, no screen, no external GPU. It’s a vanilla MacBook Pro that was rock solid and churned through tracks in Live 10 and just feels crippled by 11.

When I do attach my interface, keyboard, push etc nothing changes with regards to the performance difference. 11 is a massive resource hungry beast and I’m worried about upgrading hardware if I’m going to see the same issue. Even in this one thread people have identified I’m not alone so something is going on if 10th Gen CPUs are being hit as well.

Still, that doesn't offer us any view on what's in those sets where you notice a cpu increase compare to L10, so it's hard to help...
e.g on a 11 years old macpro I don't see any relevant CPU difference between L10 and L11... and I did test this quite extensively for quite some time.
So, there's something wrong because of "something" that doesn't work correctly with L11 only on your side, and we have to find what it is. It can't be just L11 as a whole.

[properLofi]

As I said I’m using the set used for the standard Live 10 Performance Test. The one from lesson 16 of Live 10’s help.

I standardised as much as I could to carry out a direct comparison that would provide useful information.

Nothing connected to the laptop, a set built by Ableton with known requirements that could translate to anyone else who wants to try the same.

I get that it’s frustrating when new software gets bashed but I’ve been pretty clear about what I’ve done to test this. I’m not just arbitrarily saying 11 is worse on my CPU than 10. I’ve been methodical and thought I’d explained myself pretty well. If I need to provide any more info then I’m happy to but I think it’s pretty clear that it’s to do with Live 11 rather than any other esoteric combinations of controllers, devices or sets.

[[jur]]

As I said I’m using the set used for the standard Live 10 Performance Test. The one from lesson 16 of Live 10’s help.

I standardised as much as I could to carry out a direct comparison that would provide useful information.

Nothing connected to the laptop, a set built by Ableton with known requirements that could translate to anyone else who wants to try the same.

I get that it’s frustrating when new software gets bashed but I’ve been pretty clear about what I’ve done to test this. I’m not just arbitrarily saying 11 is worse on my CPU than 10. I’ve been methodical and thought I’d explained myself pretty well. If I need to provide any more info then I’m happy to but I think it’s pretty clear that it’s to do with Live 11 rather than any other esoteric combinations of controllers, devices or sets.

Sorry I didn't notice you were using this set, my bad. That's indeed not good sign then.
Just to be sure: you have the same buffer and sample rate settings in L10 and L11?
Is the indexing finished (it takes longer when you install a new version)?

Btw, I have no problems with people bashing out Live, I'm just trying to help actually.

[properLofi]

Hi Jur, I appreciate the help. It’s just frustrating to see such a dramatic change. I can shift my workflow and rely on freeze tracks more but it feels like something untoward is happening.

I’ve been running 11 everyday since release for at least a couple of hours. How long are we talking for indexing and how do you tell if it’s completely finished? Would it show up in activity monitor?

At this stage the only difference I can see is that I had a MaxFpsMac set in options that is no longer working in 11. Single thread use seems really high but overall system CPU use is below max.

I double checked buffer and sample rate settings before running the benchmarks so it does feel like either 11 is the culprit or it’s interacting with something else on my system that 10 didn’t. To be honest I can’t see what that would be unless it doesn’t like certain software like Adobe or Office being installed. It’s a pretty empty machine to be honest.