Interesting video for us Mac users

[lclyman1]

This is a very well done comparison of most of the major DAWs and their usage of the new M chips...

Luna was not included, but perhaps fortunately so as I'm pretty sure Luna would not fair very well in this test..most likely sandwiched somewhere between ProTools and Studio One's results as Luna's poor CPU optimization is most likely similar to the other poorer performers:

Perhaps tests like this will stoke some of Luna's developers competitive juices.

Video link:

 

[EoSNJ]

I did a quick informal test when I saw this vid, and Luna does appear to be using the efficiency cores.

 

[lclyman1]

I did a quick informal test when I saw this vid, and Luna does appear to be using the efficiency cores.

How did you test?

 

[Alexxon]

I heard James Zhan is from our university, he is a tutor here in our music falculty. He make decent Mac related videos and I was wondering if one day I could meet him in person lol.

For what his video demonstrates, if you are not using those DAWs that doesn't make use of efficiency cores, you'll be good buying M3Pro / M3Max. But what Apple doing recent years is really collecting oily money, keep blurring the categories between different Mac, now they even have base model with 8GBs of RAM to sell over 2000CAD.

For those who rely not that much on GPU but on CPU, you need to buy M3Max models now to get better performance. In the past, there is no much difference between the CPU performance of Pro & Max Apple Silicons. I grabbed an used MacBook M2Pro as soon as I saw their M3 MacBook releases and I think I made a wise move, saving 1000CAD compared to M3Pro models.

 

[Alexxon]

What's more ridiculous about is that Logic Pro, an Apple's own DAW, doesn't make use of efficiency cores. It's hard to not convience myself that Apple is not intentionally to do so. Apple is always at the top of the race and I've been a user of Apple for more than 10 years, but still, they sucks at some point.

 

[taliosanchez]

So far, on my own tests, LUNA isn't using high-efficiency cores.

 

[lclyman1]

So far, on my own tests, LUNA isn't using high-efficiency cores.

According to Drew on another forum, Luna uses the same approach as Ableton in it's use of multiple cores...and according to the tests in the video, Ableton does not use the efficiency cores either.

And Ableton had the poorest performance of all the DAWs tested...

 

[SomeBassPlayer]

Why would UAD make their app able to run native plugins better than their own DSP hardware can?

At this point, I have seen enough videos of M2 Pro machines running 100+ tracks (usually Logic) with a Kontakt instrument and a set of 4-5 plug-ins on every track, eq's/reverbs included...the days of needing DSPs for anything are rapidly coming to an end (especially for mixing) especially if developers are not taking full advantage of apple silicon completely as of yet. When they really start getting in there...how many of us need 300 tracks with a convolution reverb on each one?

 

[kcatthedog2]

Not certain your parallel conclusion, luna would fare badly is valid, as luna can run plugs natively or on ua dsp?

 

[taliosanchez]

Not certain your parallel conclusion, luna would fare badly is valid, as luna can run plugs natively or on ua dsp?

LUNA can run plug-ins natively and on DSP just like any other DAW out there. No difference at all in that regard.

 

[EoSNJ]

How did you test?

MBP 16" M2 Max (8 Perf/ 4 Efficiency cores) 32G Ram
Informal test.... I didn't max it out, I was interested in seeing if the EC's were showing utilization.

Created a session with 3 instances of WaveState Native, each running complex wave sequences, and 3 SWAM instruments - Flugelhorn, Bari, and Trumpet. All instruments using A800 tape, 1073(x). 2 of the WS's also have 176's, the other a Waterfall Rotary Speaker. These three are going to a Neve Summing bus, ATR-102, and SCS in Drum Live mode. The 3 swam instruments are each going to a second Neve Summing bus, with ATR-102 and SCS in Ensemble Horns setup. Master bus is Neve Summing, ATR-102, Hitsville Mastering EQ, and CMC.
So only 6 tracks, but lots of heavy hitting CPU stuff.

Note the CPU graphs. Cores 5&6 (P-cores) are easiest to observe the following events:
Spike 1 is when I launched Activity Monitor.
Spike 2 is when I double clicked the Session file - Luna wasn't running before it, so that activity includes Luna booting up and loading the session.
after letting it sit for a bit, I hit Play. that's the big ramp up.

Clearly the Efficiency Cores are being used.

(Quick follow up edit - out of curiosity I reloaded the session to check & try a couple things...

Render meter when playing is at 53%

Bypassing all processing dropped the Perf Cores way down, but Efficiency cores are consistent with prior test. Hmmm....

 

[EoSNJ]

Why would UAD make their app able to run native plugins better than their own DSP hardware can?

At this point, I have seen enough videos of M2 Pro machines running 100+ tracks (usually Logic) with a Kontakt instrument and a set of 4-5 plug-ins on every track, eq's/reverbs included...the days of needing DSPs for anything are rapidly coming to an end (especially for mixing) especially if developers are not taking full advantage of apple silicon completely as of yet. When they really start getting in there...how many of us need 300 tracks with a convolution reverb on each one?

UA developers have had a ceiling for 20 years. They've had to be incredibly efficient coders to produce such high quality plug-ins within the UAD-2 chip environment. It has kept them disciplined by necessity.

Going native hasn't removed that ceiling, but it's much, much higher now. Have no doubt though, we'll hit this new ceiling. And we'll need faster computers again.

 

[kcatthedog2]

LUNA can run plug-ins natively and on DSP just like any other DAW out there. No difference at all in that regard.

So, I don't understand what your concern is, do. you actually use an m2 or m3 and have you actually had a performance issue

 

[Steven A]

In my tests, Luna uses all E and P cores, but not at all in an optimized way.
(Not at all like Reaper and Cubase like in the video)
There is still a lot of CPU heardoom when LUNA render is a 100%. A shame, hope it will get optimized.

 

[kcatthedog2]

I understand the tests but was your session actually negatively affected?

i find the focus o the original video a little misleading for audio work: how often do we need to use 100% of the processing of a given computer?

How often do any of us get in our cars and drive at 100% of its power:like never.

I run typical logic sessions without ua dsp on an m1, never heard my fans come on or seen any cpu usage even at 50%.

I understand the theoretical concern sbout the different cores but the m2/m3 processing each is a roughly 25% increase over my m1, so I wonder how much of a real world issue the m3 differing cores would be for a typical audio user. I think, not much and if you are aldo using ua dsp even less.

Video editing/rendering: different story.

 

[Steven A]

My concern is that LUNA offers native extensions which are great but in a 50+ track scenario, which happens very often in my case, in 96k, render % goes up very quickly.
Some native plugins are super heavy too (SCS !)
Even working in 48k render % goes up quickly.

Freeze tracks is an answer, but I don't like because no editing possible. Commit would be best but doesn't exist yet. Bounce out then drag back is the only way to free up CPU cycles.

I can work without any hiccups, and have loads of sharc power, but it would be great that the code can use up more power since it's there

 

[taliosanchez]

So, I don't understand what your concern is, do. you actually use an m2 or m3 and have you actually had a performance issue

I have a MacBook Pro M1 base model which have served me perfectly fine during these 2 years, mixing/mastering primarily in REAPER in my commercial studio with clients. During this year I've had a few projects from friends that I've mixed in LUNA and the amount of power I can squeeze from it is like 40% less than what I can do in REAPER. Pretty much like ProTools which is the other DAW I have to use when clients send me their sessions. Both PT and LUNA are like 40% less efficient than REAPER and after I did my own tests I noticed the problem: they don't use the high-efficiency cores like it.

I just hope LUNA devs can find a way in the future to give the user the option to also take advantage of the high-efficiency cores.

 

[kcatthedog2]

Fair enuff !

 

[lclyman1]

I have a MacBook Pro M1 base model which have served me perfectly fine during these 2 years, mixing/mastering primarily in REAPER in my commercial studio with clients. During this year I've had a few projects from friends that I've mixed in LUNA and the amount of power I can squeeze from it is like 40% less than what I can do in REAPER. Pretty much like ProTools which is the other DAW I have to use when clients send me their sessions. Both PT and LUNA are like 40% less efficient than REAPER and after I did my own tests I noticed the problem: they don't use the high-efficiency cores like it.

I just hope LUNA devs can find a way in the future to give the user the option to also take advantage of the high-efficiency cores.

I brought this up with the same comparison to Reaper months ago...this was before the tests referenced in my original post which now point out where some of these not insignificant differences are caused.

However, these same differences were present before the M1 chips were as common, so at least in Reaper's case, it's superior coding all around while the current M series differences highlight that even more.