Which Mac Studio for music production? Max M2 vs. Ultra M2

[Eastie]

Saw a video on YouTube yesterday of a guy who does orchestral/cinematic stuff running almost 2300 tracks of Kontakt instruments @ 48khz, 24 bit, 256 sample buffer on a new m2 ultra studio.

Ordering mine asap, Imho Intel doesn't hold a candle when it comes to performance vs power consumption, form factor, hassle setting up/building system etc and longevity.

Long time user of pc's btw, always die, and age badly. using macs for just over 2 years now and its just been a breath of fresh air.

I'll be sure to maybe uploads some extreme actual real world tests when my system arrives as those are pretty hard to find in a world of graphs, offline testing and specification comparisons.

So I'm sure a M2 max with the aforementioned ssd upgrade will be pretty much beast compared to most setups available today

 

[ratspg]

Saw a video on YouTube yesterday of a guy who does orchestral/cinematic stuff running almost 2300 tracks of Kontakt instruments @ 48khz, 24 bit, 256 sample buffer on a new m2 ultra studio.

Ordering mine asap, Imho Intel doesn't hold a candle when it comes to performance vs power consumption, form factor, hassle setting up/building system etc and longevity.

Long time user of pc's btw, always die, and age badly. using macs for just over 2 years now and its just been a breath of fresh air.

I'll be sure to maybe uploads some extreme actual real world tests when my system arrives as those are pretty hard to find in a world of graphs, offline testing and specification comparisons.

So I'm sure a M2 max with the aforementioned ssd upgrade will be pretty much beast compared to most setups available today

That’s amazing! Can you please share that YouTube video? Thanks!!

 

[Eastie]

 

[Eastie]

Here we are

 

[Think77]

Saw a video on YouTube yesterday of a guy who does orchestral/cinematic stuff running almost 2300 tracks of Kontakt instruments @ 48khz, 24 bit, 256 sample buffer on a new m2 ultra studio.

Ordering mine asap, Imho Intel doesn't hold a candle when it comes to performance vs power consumption, form factor, hassle setting up/building system etc and longevity.

Long time user of pc's btw, always die, and age badly. using macs for just over 2 years now and its just been a breath of fresh air.

I'll be sure to maybe uploads some extreme actual real world tests when my system arrives as those are pretty hard to find in a world of graphs, offline testing and specification comparisons.

So I'm sure a M2 max with the aforementioned ssd upgrade will be pretty much beast compared to most setups available today

While I certainly don’t run 2300 tracks in Logic in my projects, I will say that my M2 Max Mac Studio is crazy good! 😊 It sits at around 20-30% cpu usage even with more complex arrangements with lots of cpu hungry plugins/instruments that would bring my old M1 Mac Mini to its knees. Lots of headroom going into the future! And it’s absolutely silent too. Couldn’t be happier.

 

[ratspg]

Here we are

That's such a wild test! I use Cubase for years and it almost hits 1k tracks , logic blowing it away with an extra 1300+ tracks lol. WILD TIMES!

 

[savvn001]

I chose the i9 13900K.

View attachment 2222889

View attachment 2222894


It is much more efficient than the M2 Ultra and I made myself a RAID storage on Sample.

View attachment 2222893

Just stumbled across this through a google, 500 tracks is insane! From my research I saw the i9 13900 actually performs slightly worse than the M2 Ultra in synthetic bencharks due to macOS not being able to use the P/E cores properly, but your system is smoking it, odd. Can you run Geekbench 6 on your rig? Very curious...

 

[ori69]

Just stumbled across this through a google, 500 tracks is insane! From my research I saw the i9 13900 actually performs slightly worse than the M2 Ultra in synthetic bencharks due to macOS not being able to use the P/E cores properly, but your system is smoking it, odd. Can you run Geekbench 6 on your rig? Very curious...

No one knows what Geekbench is really measuring. Did you notice that with this test not all cores are used?
But despite this Intel still manages to be better.

13900kf macos - Geekbench 6 CPU Search - Geekbench

Logic Benchmark and Cinebench evenly and consistently use all cores and in these tests Intel comes out much better.

 

[Eastie]

Meh, who cares.
New ultra studio can run 2300 tracks of Kontakt and takes up the same desk space as a cheese sandwich.
Ez win!
Keep your windows headaches 😂😅

 

[ori69]

Meh, who cares.
New ultra studio can run 2300 tracks of Kontakt and takes up the same desk space as a cheese sandwich.
Ez win!
Keep your windows headaches 😂😅

Since M2 Ultra run 2300 tracks of Kontakt, and can't pull more than 400 tracks in Logic Benchmark, i9 13900K can much more. Unfortunately the youtube user did not post the project he tested.
And the small computer box = clutter on the desk. I can have it all inside.

 

[Eastie]

Not my fault you work in such a mess…
In a m2 max vs ultra thread, you derail by posting, “my hippopotamus is better than either of the two horses you specify”.

I came here with the same question as the op, found a more useful answer elsewhere, so I thought I’d reply here to help op and not distract with apple butthurt.

No offence, just it’s a better machine, I’m a long time windows user also.

It’s just preference, I have a nice neat situation 😊

The studio will be a nice upgrade for me!

 

[PeterKnockaert]

I went with the M2 Ultra (hasn’t shipped yet) with 192 gigs of ram since I currently use two computers with 64 gigs of Ram each and easily fill both up. I would think you would want at least 128 to give yourself a little headroom. I do music production with large orchestral templates so if you’re doing anything like that you’re going to want that extra ram, since you won’t be able to upgrade in the future. My two cents.

EDIT: I have an identical Mac Mini to yours. Without question you will see a major increase in power with the M2. No comparison at all. I was hoping to make the mini my work horse machine but it just didn’t have the power - even compared to my 12 core 5.1 Mac Pro.

Hi @dpcon, I have a question regarding the orchestral template: what do you use for (external) devices to run your samples from? External hard disks? What's the best system for that?
I will order a Mac Studio with 4TB and 128 RAM. I was thinking about the OWC ThunderBay 8 to run my samples from (I now have 3 SSD's packed with samples, about 5 TB in total).

 

[ratspg]

Hi dpcon, I have a question regarding the orchestral template: what do you use for (external) devices to run your samples from? External hard disks? What's the best system for that?
I will order a Mac Studio with 4TB and 128 RAM. I was thinking about the OWC ThunderBay 8 to run my samples from (I now have 3 SSD's packed with samples, about 5 TB in total).

Not @dpcon , but I can chime in and share what I’m using with my M2 Ultra Mac Studio.
Had tons of SSDs prior with the 2013 Mac Pro, now I’m using this nvme enclosure with 32tb of storage. Super fast, has tonssssssss of sample and orchestral sample libraries on them. No issues, smoothest sailing I’ve had in a long time!!
NOTE: The stock fan on that enclosure is SUPER loud. I purchased a Noctua fan that makes barely any noise and replaced it and now it’s not noticeable.

OWC Express 4M2

Four M.2 SSD bays with pure adrenaline inducing Thunderbolt 3 performance.

eshop.macsales.com

 

[savvn001]

No one knows what Geekbench is really measuring. Did you notice that with this test not all cores are used?
But despite this Intel still manages to be better.

13900kf macos - Geekbench 6 CPU Search - Geekbench

Logic Benchmark and Cinebench evenly and consistently use all cores and in these tests Intel comes out much better.

View attachment 2243009

I saw you have processing threads set to 'automatic'. Thing is that setting is odd and nobody knows what it really does on automatic mode. If you set it to the max (I think it will be 32 in your case) what results do you get? In theory it should be even higher.

I know why the i9 still smokes it after doing extensive research, because of hyperthreading you're essentially getting 32 procesing threads for logic, vs an M2 ultra with 24 processing threads (no SMT on apple silicon). On top of that some tests are showing that logic for some reason doesn't want to use up the efficiency cores on apple silicon... so on m2 ultra it's like having 16 effective processing threads!

 

[Soundflut]

Finally I bought the M2 Max.
And it is a real game changer! Awesome
fast and not to compare with my intel-powered Mac Mini (64GB).

It was the right decision. Even video rendering is very fast, so - if anybody has the same question in mind:
I guess you won‘t need an ultra. But the M2 in general is the money worth!

 

[ori69]

I saw you have processing threads set to 'automatic'. Thing is that setting is odd and nobody knows what it really does on automatic mode. If you set it to the max (I think it will be 32 in your case) what results do you get? In theory it should be even higher.

I know why the i9 still smokes it after doing extensive research, because of hyperthreading you're essentially getting 32 procesing threads for logic, vs an M2 ultra with 24 processing threads (no SMT on apple silicon). On top of that some tests are showing that logic for some reason doesn't want to use up the efficiency cores on apple silicon... so on m2 ultra it's like having 16 effective processing threads!

Here you have an explanation of why M2 processors lose in tests in Logic Pro.

 

[Chancha]

Here you have an explanation of why M2 processors lose in tests in Logic Pro.

View attachment 2266821

View attachment 2266820

The M1 Pro and M2 Pro are not supposed to have the same 10 cores total, this looks like with his test the M2 Pro is binned where the M1 Pro is not. In this case the M2 Pro has only 6 p-cores while the M1 Pro has 8, if unbinned the M2 Pro was supposed to have the same number of 8, with extra 2 e-cores.

 

[savvn001]

The M1 Pro and M2 Pro are not supposed to have the same 10 cores total, this looks like with his test the M2 Pro is binned where the M1 Pro is not. In this case the M2 Pro has only 6 p-cores while the M1 Pro has 8, if unbinned the M2 Pro was supposed to have the same number of 8, with extra 2 e-cores.

Yeah in his test he's compared the 8P/2E M1 Pro/Max to the 6P/4E M2 Pro. There is an 8P/4E M2 Pro which would of definately beat the M1 gen at least for the sake of testing.

But still it's odd that logic, apple's own app isn't using all available cores! Strange, wonder if it's on purpose to make sure the efficiency cores are left to background tasks.

 

[savvn001]

Here you have an explanation of why M2 processors lose in tests in Logic Pro.

View attachment 2266821

View attachment 2266820

Yeah this is the vid I watched also. Hackintosh is still offering much better performance per $ if you don't need mobility. Considering doing the same i9 13900 build as you.

 

[ori69]

Yeah this is the vid I watched also. Hackintosh is still offering much better performance per $ if you don't need mobility. Considering doing the same i9 13900 build as you.

The Intel 13900K has 16 e-cores. Logic you use all cores.

 

[Ethosik]

Assuming those two Mac Studios are both 512GB base models, then there is one downgrade happening where the sequential SSD speed went from 2022's 7000/5000 to 2023's 3500/3000. May be the slightly slower read speed of the 2023 M2 Max model becomes the bottleneck since the test reads loads of data from disk?

Anyway for any M2 Max config, just spec the SSD to 1TB or more then it should perform better than M1 Max Studio.

My 4TB M1Ultra Mac Studio only gets 3,500 MB/s FYI.