Up next in 10
Mac Studio w/M1 Ultra - What to know BEFORE you buy…
Apr 30, 2024
We had high hopes that the M1 Ultra Mac Studio performance would be off the charts, but in most tests it hardly outperforms the M1 Max. Watch full episodes of the CultCast at http://watch.thecultcast.com and catch the live show Thursdays around 11:45am PST. Search "The CultCast" in your favorite podcast player for full audio episodes.
View Video Transcript
0:00
All right, let's send it back over to our special guests here
0:03
We got Max and Vadim here from Max Tech. Guys, I have been watching a lot of your benchmarks on the M1 Ultra
0:12
And I guess my question is, is there any point? Is there any point for anyone buying the M1 Ultra
0:17
I mean, you all did. I did. I spent the extra money, the $2,000 upgrade
0:22
But watching all your benchmarks, it really seems like it just doesn't benefit me or even
0:29
99% of people at all to actually spend the extra money, even though I feel compelled like a
0:34
moth to a flame to spend the extra money and do it. Is there any purpose in buying the M1 Ultra
0:41
Do you want to feel better about your purchase? So let me know what angle should we take on
0:46
You're going to tell your answer. I want the cold, hard truth here. I want the cold hard truth
0:50
I mean, I'm doing what you're doing, right? So we're content creators and we're making YouTube
0:54
videos, at least you are. So does it benefit people like us? Not really, no
1:00
So actually what we're doing right now before getting on this podcast, thank you guys for inviting us
1:06
And then afterwards, we're benchmarking video editing. Both DaVincior resolve, Final Cut and Premier Pro
1:13
Final Cut was just updated. Premier Pro is just updated. So we're running new tests, retesting things
1:18
And for a lot of content creators, most YouTubers out there, there is no reason to go for the M1 Ultra
1:26
Now, obviously, you can get up to 128 gigs of RAM, which you could with previous Intel Max
1:31
but with unified memory, it's way more efficient. It's way faster. The SSDs are faster
1:35
And you don't really need 128 gig if you needed it before. 64 is plenty
1:40
So for us, unless you're high-end filmmaker working with raw footage and crazy color
1:45
corrections, multi-layered, denoising. So even for people like yourself, because you're putting out content, your content's 4K, right
1:54
Yeah. And even for you, it seems. like, did you go with the M1 Ultra Mac Studio
1:59
Because I know you bought like every single one they make. But which one do you keep? We have three of them
2:03
And I guess the top N1 with a 64 core 128 gig, definitely not worth it
2:10
The $4,000 one is great for some people. We also have the M1 Mac
2:15
So we're going to make the final decision for Angelica, our main editor. But for people like us, it is not really worth it
2:22
Now, on the flip side, if you're somebody that's working, with applications that are CPU-based
2:28
which we talked about, I think a couple days ago, we put out a video, that 20-core CPU
2:34
can make a world of a difference. For example, if you're somebody that's
2:38
it's your profession, you're making money with the system, you're using Lightroom, you're using Xcode
2:43
you're using those applications that can actually, logic that can scale well, it can make you a lot of money
2:49
So I gave one example in the recent video, if you're somebody that's doing a bunch of photo editing
2:54
for example, for real estate, you're sitting there half the time waiting for photo merge to run, and now it's twice as fast
3:00
You can make a couple grand more a month every month by taking on more work because of the time it will save you
3:07
And then if that's a single person, if you have a team, there's a bunch of companies out there that edit real estate stuff, and they have 20, 30 editors sitting there editing
3:16
Imagine how much more productivity that will bring you how much more money you can make. That's 20 grand a month if you have a time of that
3:24
So for CPU, it could be a massive difference if you can actually benefit from it
3:30
So, but when you say for people who are constrained by a CPU performance, right
3:35
What applications is that? Because it seems to me like they really the only two that might take advantage of all the extra cores is lightroom oddly enough and and Blender But other than that it seems to me this is just my sense of it It seems like for regular folks the M1 Ultra
3:52
might be 20 to 30 percent faster than an M1 max. And then for applications like Lightroom
3:58
like Blender, it might be closer to like 80 percent faster than an M1 max. But for normal everyday usage
4:05
it doesn't seem like it's that much faster, maybe 20 to 30 percent. Is that, is that sense
4:10
seem right to you? Well, off the top of my head, I think in some cases, lightroom was, yeah, like
4:15
80% faster, 90% faster. So for some people, that can make a world of the difference. If you can
4:22
sit on two, three more edit jobs a day, you can make a couple hundred bucks more a day. Now
4:28
if you're somebody just sitting there as a hobby, doing it for fun and you work on one project
4:32
a day, who cares if you can export it 80% faster if most of the time you're just, you know
4:37
moving sliders around? But like Vadim said, logic. X code. So if you're somebody that's coding and you're working on code most of the day and you're only
4:45
compiling once at the end of the day and it is 30% faster, for example, which I think that's what we saw
4:51
Does it really matter? No. But if you're compiling 20 times a day and each time you're saving a time and you're sitting there waiting
4:58
that's where you have to look at your workflow. What are you doing? What are you limited by? You can go into activity
5:03
monitor and see is it the CPU that's maxed out when you're sitting there waiting for something to happen or the graphics
5:08
and then are you making money from it? And does that extra time matter
5:12
if it could be 80% faster or 30% faster? And you said logic is one of the applications
5:18
because I use logic and I use ISAT menus. I think do you guys use ISTat menus as well
5:24
Whatever you use to track, you do. Okay, so it sits in the menu bar and you can see what's happening with your CPU
5:28
and your GPU and everything. It's actually really cool. And with logic, every time I export
5:33
it seems like only one CPU core is being utilized. all the rest are just sitting there
5:40
So you think that editing and logic will actually benefit from the M1 Ultra
5:46
So there are definitely improvements. It is CPU-based. There's also memory if you're doing a bunch of plugins and stuff
5:52
So this is where it comes down to how heavy are you hitting the system
5:56
If you don't have a bunch of tracks, a bunch of plugins, things that really take resources, you don't need all that power
6:02
Now getting back to the video editing side, for a stream like this, where you're going to cut off the beginning, maybe the end, maybe add some music
6:08
a couple titles, you can buy an M1 Ultra, but you're not going to benefit because you can even
6:14
buy an M1 MacBook Air, and for the work you do, if you're not pushing it that hard, you're not
6:19
going to see a difference. So that's where you also put into perspective. Are you maximizing the current system
6:24
Because even the M1 MacBook Air is so much faster than a lot of previous Intel higher NMAX
6:29
So that's why we always get into it, like with the Mac Studio video compared to the Mac Mini
6:33
I said, for a lot of people, you can buy the $700 Mac Mini, and it will do the job and do it
6:38
better than your previous Mac, Intel-based one that you might have bought for $2,000.3,000
6:43
Yeah, it's an interesting observation. I feel like the M1 is just so much more powerful than Intel
6:47
that, and we even said this when it came out, it was beating my MacBook Pro. I have a 2019 MacBook
6:54
Pro, and I felt like it was blowing away my MacBook Pro and a lot of the video editing tests that
6:58
were happening at the time. And I was like, do this thing's insane. Like, finally, not only can
7:02
you actually edit on a MacBook Air, but you can edit 4K content on a MacBook Air. I don't think
7:07
that was possible pre-M-1. And so they came out swinging so hard with the M-1
7:13
that I feel like they actually shot themselves in the foot in a way
7:17
because now they're struggling to find a way to actually give pros a better option
7:22
And the M-1 Ultra is their way of saying, well, you can get double the performance, which theoretically is true But when you do the actual benchmarks I was shocked to see that that not really what you get because most applications can even take advantage of it
7:38
Or if they do, like they updated After Effects, Adobe After Effects this week to be M1 optimized
7:45
And they said that with the, was it with the M1 Max
7:49
it would run up to two times faster in some of their render tests. And then with the M1 Ultra
7:55
it would run up to 3. three times faster. And I was like, wait a second. With the M1 max, you're getting double the performance
8:00
but the ultra, which is two max, two maxes, you're only getting an extra, you know, 100% or 50
8:06
performance increases. The returns there, the diminishing returns there, again, make it hard to justify the extra, you know
8:16
$2,000 with the M1 max. Some information that we got recently based on what's limiting the M1
8:24
Ultra. And I made a tweet about this, and we put it into our recent review, but maybe it's not
8:30
explained that well, but basically there's like a bottleneck that's impacting the M1 Ultra more than
8:35
the M1 Max, and it has to do with something within the hardware, and because of that, it's not able
8:41
to utilize the GPU all the way. And then on top of that, there's another issue with the encoders
8:47
You know, it has double the video encoders and decoders for like H.264 and H.D.65. However, because
8:54
of how the compression works and because of how the encoding and decoding works, it
9:00
seems like right now it's only using one set. So it's basically
9:04
like the encoders and decoders are still like the M1 Max. And we
9:08
actually ran tests where we included one set out of how many
9:14
Well, it comes with two sets, right? So two encodes, two decodes. So it's supposed to... It sounds
9:18
like a lot of these are optimization problems that you're talking about with the software, right
9:22
Is that what you're seeing? that the developers haven't optimized their stuff properly
9:26
for this new architecture. Yeah. Kind of, right? Because you're saying that part of it is software
9:32
but part of it is actual like a hardware limitation. Yeah, there was basically something that was overlooked
9:38
and I guess the Apple engineers did not think it would be a limitation
9:43
but it ended up being a limitation as they started adding more cores with the highest in N1 Max and the M1 Ultra
9:49
and then they found that out. So hopefully it's going to be fixed in the future, but if everything was optimized perfectly it would run just all the way to the top you know 105 watts
9:59
but now it's being limited because a lot of developers are just getting it over to metal
10:04
they're like oh it works apple silicon sweet good to go we're done you know we're not going to put
10:09
any more money towards investing and recoding and redoing everything basically but getting back to
10:14
the encoders it seems like only the video encoders and decoders from the m1 max are running
10:21
while you're exporting a video instead of using both how the compression works
10:26
So we actually tested it. So we compare the M1 Max and the M1 Ultra encoding H.2664 to H.264
10:33
with no other GPU effects, and it was exactly the same time
10:37
Same thing for HVC. Exactly the same time. So it's only using basically one M1 Max media engines for that
10:45
However, for ProRes, we did see almost two times more, right? Yeah, close to L
10:51
linear 50 to 80% oh okay that's there's some good news so we just got a final cut 10.6.2
10:57
which I know you have been waiting for I have been too because I was like when we got the M1 Ultra
11:01
and we started seeing the benchmarks I was like wait a second there's something way off here
11:05
in fact it was so way off that I thought maybe I was not understanding the problem correctly
11:09
but it just the M1 Ultra just didn seem to be performing and much better than the M1 max in most tests And I know you all have been testing the new version of Final Kit Pro And can you give us some high insights here
11:22
Like, does it resolve any of the problems that we've been seeing in performance with the M1 Ultra? No
11:28
That is truly mind-boggling. That's disappointing. I mean, so any application company
11:35
they could say, oh, up to two times faster, or we made something better. But just because they say it could be up to two times faster
11:40
It means maybe in one task for one specific case, at the hardest workload possible, they achieve that
11:48
But then everything else, you know, hasn't changed much. So the new version has a couple of features added
11:54
I don't know why they waited a month since the announcement or more to release it when on stage the numbers they gave
12:02
That's the software they were using. So in general, all the stuff we looked at, we're not seeing any of the bottlenecks that were there before
12:10
that they've been fixed. Maybe somewhere in the timeline, something can get smoother
12:15
something very vague, but overall, it's not a big difference at all
12:20
So I don't know if you said this already. I know you were considering your decision here
12:24
but are you going to keep the M1 Ultra Max Studio? Because if you guys don't keep it
12:30
now I'm seriously thinking that I shouldn't keep mine. Because honestly, like if you guys don't need it
12:34
you're creating so much content that it makes me think, okay, I should just get rid of it
12:39
and buy the M1 Max version because that's probably, I mean, that value proposition, I think that was probably
12:45
my most, like, the most exciting thing that they announced at the keynote event. I was like
12:49
you can get an M1 Max now for $2,000 and you look at a spec-out MacBook Pro with the same specs and
12:55
it was like $3,500. I'm like, this is a crazy good value if you don't need the screen and the keyboard
13:00
So do you know what you're going to do yet? So basically, after we're done with this video that
13:05
we're going to finish testing and shooting today, we got to talk about it and see which one
13:09
So Angelica's been using an iMac, the 24-inch iMac, with the M1 and 16 gigs of RAM since that came out for the main editing machine
13:19
And we're making... And it works. It works great. The only limitation we have sometimes is the encoder speed, just if we need to get a video out as fast as possible
13:27
And sometimes 16 gigs of RAM can be a limitation. Now, logically, and we should be keeping the $2,000 Mac Studio because the $4,000 and $5,000,000
13:39
$600 ones that we ordered, we're not going to get any benefit for what we're doing. Literally
13:44
Even right now, our main thing we wanted is be able to export faster with the quad encoders that
13:49
they have in there compared to M1, but we're not seeing that improvement right now
13:53
Oh my gosh, that's so unbelievable. And you should keep the $2,000
13:58
Now for us, the other difference is we got to create content when new Macs come out, other PCs
14:03
We'd have a system here to compare it to. So that's where the other side comes in of maybe we should keep it, even though
14:09
won't make a difference just so that we can use it for future videos
14:13
You know, the funny thing is, is like, that's where mind is at too. I'm like, you know, I don't seem to be realizing any actual benefits from this
14:20
but maybe in the future things will change or maybe Apple will release some kind of update that
14:25
We'll release the hidden power. Actually, we have some information where there's a chance the next version of macOS
14:32
could utilize some of those encoders and decoders. Okay. We'll leave it at that right now
14:38
Is that some inside Intel direct from your sources? Yes. Very nice
14:45
Very nice. Juicy
#Computer Hardware
#Consumer Electronics
#Laptops & Notebooks
#Multimedia Software