I provided you the reason why I responded to your post. If you feel that is being singled out then I would suggest you stop being so sensitive.
That said I am done responding to you about this. It's not constructive.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Would have bought an M4 Ultra Studio - decided not to get the M3 Ultra or maxed out Max - anyone else?
- Thread starter M4M
- Start date
- Sort by reaction score
That said I am done responding to you about this. It's not constructive.
chuckles back to form. you didnt provide a reason as much as an excuse, and now you are insulting, but since you refuse to actually discuss the topic I am cool with moving on. I suggest maybe you keep a more open mind to things.
The Mac Studio is just a tool.. Nothing more, nothing less. Apple released what they could design and build. We should all make such big mistakes.
It's your opinion the target market would rather have a Studio Ultra based on the last generation technology instead of the current generation?
It’s my opinion that a lot of people on these forums spend too much time chasing specs and numbers and then complain anytime Apple doesn’t meet their expectations instead of actually doing something with the tools that Apple provides, which are objectively quite incredible given where we were just a few years into the Apple silicon age. The slagging going on over the “M3” Ultra is a waste of time quibbling by people who only see the number, can’t get past it but don’t have any genuine use for that much horsepower. I certainly don’t and if I did, I’d just buy the thing and get my work done. Which is probably what is happening judging by the articles across the internet by those who are buying the M3 Ultra for their very specialized workloads and raving about the work they are getting done.
Ah, but are those people buying M3 Ultras because they're good for their specialised workloads making click-baity Youtube videos about it?
I’m sure there are those YouTube people and there are those who are simply writing blog posts about their specialized workloads and there are probably a whole bunch of others who bought one and you won’t hear a peep out of. The YouTubers on either side of the fence about the M3 Ultra are generally gearing their videos for eyeballs and little else.
EDIT: there are a few YouTubers who will put up honest videos of their experience with the M3 Ultra.
Last edited:
I see. So it's your position only those people who "have any genuine use" for something are the only ones who can offer their opinion on it? No other opinions are welcome?
Why do people keep saying this? Aren't the YouTubers there to help people make informed decisions? I watched the Luki Miani video and I felt he was right on the money. He was fair, balanced, and essentially said what many here have been saying: The Ultra offers the best performance for specific use cases despite it using the previous generation technology. However it is his opinion, which appears to be shared by others in this forum, that the lack of current technology puts a damper on it and is setting it up for failure. Why is that "click-baity"?
And if they do, so what?
It’s my opinion that most people would prefer to win the lottery given the choice. I also believe most people would prefer to buy Mercedes at Volkswagen prices.
So what? That’s not reality.
My opinion is if all things being equal Apple could have made an M4 ultra they would have. But for whatever reason, logistical, financial, technical, we can speculate all day, they couldn’t. So the machine they could release, the one we can judge on its own merits, is the M3 ultra. And the verdict from unbiased sources seems to be if you’re concerned more about the single core speed the Mac studio isn’t for you. Buy a m4 and be done with it. And if you are concerned about traditional graphics, the m4 max might be enough. But if you are manipulating huge data sets that 99.9 % can’t even contemplate and you need 512 gb ram, then the machine that actually exists and not just dreamed of, the m3 ultra is worth a look.
To call it a failure is just ignoring reality.
You forgot the sarcasm flag. You tubers exist to make money. They make money by getting clicks. It’s easier to get clicks by saying what people want to hear, which generally is complaints. If you don’t keep this in mind while watching a you tube video… you are not getting a complete picture. I look for you tubers that don’t tell me what to think, but just give me the facts as close as they can so I can make my own decision. Not what I want to hear, what I deserve to hear.
Yeah, but, unlike AMD, Apple designs their chips with phones in mind first, then scales that design up and out for everything else.
Yeah... no. The four video encoders of the M3 Ultra makes it the Mac of choice for those producing videos. The massive memory also means very high numbers of audio tracks, etc.
Tim is not victimizing you.
Of course they’d rather have m4 generation!
But it doesn’t exist!
You are seriously under estimating the risk and expense of building ultras and why Apple has always taken about 12 months after the pro to release them. They are built on field tested cores that the manufacturing bugs have been well and truly sorted out for.
They are likely accumulating top tier max dies for a year to build them with - and the process of joining them isn’t 100%. Risking a number of working max parts to get an ultra with a significant failure rate is expensive. They could have just been sold as max.
I ordered the 60-core and 80-core models but after doing some tests decided to return them today and am waiting to see if something is announced at WWDC. My main reason for it over the M2 Ultra I have is for video AI upscaling. The 80-core was 10% faster than the 76-core M2 Ultra in the model I use. I was expecting a bit more uplift based on Apple's own performance charts. If nothing new is announced at WWDC, I may still pick one up but go with a higher RAM config this time.
I think for the most part they update the devices they sell in the highest volume first. iPhones above all, and on the Mac side they generally update laptops much more often than desktops. I'd bet money the high-end Macs everyone is talking about in this thread amount to a small fraction of the revenue they make from MacBook Airs in a given quarter.
Oh, I understand why. It’s a question of logistics and development approach rather than a halo/prestige market play. Not at all surprising with Tim Cook in charge. It just means I’ll rarely do an impulse purchase (on timing) since weird anomalies like the M3 Ultra will appear unexpected and I won’t know if/when an Ultra variant of a Max chip will appear.
Everyone understands it doesn't exist. The discussion is about why and does it make sense to have released an M3 Ultra when the M4 Max offers highly competitive performance in the majority of benchmarks? There are those who appear to have an issue with such discussions and even more so with YouTubers who raise that question.
True. I should have worded that better.
If you need single core performance, you should buy an M4 Max. Basically the only reason to buy an M3 Ultra is if you do work where the narrow areas and situations where the M3 Ultra is the superior option are hit. For example, the large amount of RAM with the M3 Ultra has a great deal of use for people who want to run current LLMs on device.
I'd also note that for most of my adult life server/high-end workstation CPUs often lagged behind the bleeding edge of consumer technology in terms of core quality. What Apple was doing with the M1 and M2 generations - getting the workstation/server grade SOC out before the next generation consumer grade - was more of a welcomed exception to the norms in this space.
Apple needs to create more market differentiation with their workstation or higher grade SOCs. Especially if they finally figure out how to cost-effectively make the 4-Max/2-Ultra Frankenchip rumored in the M1 days. If the M4/M5 Ultra and Extreme do happen, I think it would be better to harken back to the xServe days and call them X1 and X1 Extreme. They'll always be a generation behind on individual core advancement but they're the options for people who need more or MUCH more cores than a non-binned M[n] Max can provide.
Last edited:
And you've gotten the answer. Multiple times.
So now, why don't you accept the answer?
My thoughts exactly. The current linear single prefix doesn't accurately reflect the range. This will affect the potential resale value down the road. People wishing to flip their M3 Ultra will have to be prepared to go into a deep dive to explain why M3 Ultra is, or was, the bee's knees at the time. How tedious.
That simple fact is enough for me to not consider shelling for a new M3 Ultra from the M4 era. Whether or not some people think we "spend too much time chasing specs and numbers", it's still a marketing disaster IMO, for both Apple and customers wishing to upgrade later.
As the old saying goes: "Never make it hard for people to give you money".
Last edited:
Yup, you and @rgwebb are right on the money( no pun intended ), the biggest problem here is Apple is literally being *too* clear about arch generation in their naming and spec hounds are grumpy.
The M3 Ultra is, as y’all have said, the “server” iteration (quotes because, at least for now, Apple doesnt make an actual server - though I do kinda expect the xserve to return with the same AI focused custom design apple is reportedly starting to use internally to power the next gen of Apple Intelligence/Siri since selling them to consumers would drop per chip cost internally), which traditionally lags for most CPU designers. Higher core count so better multicore perf and lower single core perf is common. If Apple named the Ultra (and possible future rumored “extreme”) line something else you wouldn't have this problem
The Ultra SoC is not fatally flawed in any way shape or form. It’s a use case scenario. Just because he doesn’t see the value doesn’t mean other don’t. I have the maxed out M3 Ultra and for me it’s an absolute bargain. 512GB of unified memory beats anything I could do for 5x the costs of just Nvidia GPUs. Nvidia would be faster with the same RAM but therein lies the problem. Have to cluster together many Nvidia systems to get the same amount of RAM. So for LLMs, advanced GPU needs, or video editing at scale, the M3 Ultra is very fairly priced. The only thing that isn’t fair is the storage price. But for $10k can get the M3 Ultra with 512GB of unified memory and 4TB of storage. Can’t beat that. Can even come close. Run LLMs locally. Nearly 2x the GPU performance. The video encoders and decoders are doubled. Forget the whole M3 vs M4 argument. It’s all just not relevant. If you don’t need an Ultra, you know it. If you do need an Ultra, you know it.