Some things to consider. Was reading tomshardware forums and their quite much poking holes everywhere regarding these benchmark claims by Apple. First of all, the Zen2 3990x scores north of 35.000 in Apple's own geekbench 5 app, the M1 Ultra aint even nowhere close. As many on the toms forums wonder is where Apple has these numbers from.

How do random forums goers know what Apple's M1Ultra benchmarks will be? The fastest accessible M1 chip is the M1Max in the MacBook Pro 14" and 16" but the M1 Ultra is designed for non-laptop form-factors, is not limited to what power can be sustainably delivered by a battery and with vastly increased thermal ceilings. :???:

What considerations have been made for these rather massive differences?

3:18 So what they seem to be saying is that if you set both the M1 Ultra and a 12900K to 60w, it is 90% faster. That's not very difficult given how far the 12900k is away from it's maximum power consumption and thus clockspeed. It would be interesting to see a comparison of the 12900K and 5950X in the usual benchmark suite though.''
210
Efficiency is always about performance per watt. Unless that radically changes at different wattages, what difference does it make?
 
First this is not an Apple app.

Many claim that Geekbench is created by a Mac-review site. I have no idea if that is true or not, just mentioning it since i hear it everywhere. It is not a usefull benchmark to gauge pc performance as many others have already said, in special not between different architectures and platforms.

Second those 35k scores obviously are overcloked results; a result closer to reality is ~25k as shown on Geekbench processor summary page.

So what if these are OC'ed results? Also, why was Apple comparing to old Zen2 products? Theres Zen3, and as of yesterday Zen3 TR up to 64 cores. The comparison to Intel was very weird aswell, they went with a 60watts alder lake to compare to the M1 Ultra and made performance claims. its good to show efficiency in Apple optimized apps that make use of the media accelerators, but what consumer is going to limit their much cheaper alder lake cpu to 60watts?


Is it true or just hearsay?

That guy is full of shit.

Ofcourse he is, it wasn't too pro Apple so yeah theres that.

And it's not that hard to prove it

Apple never really went into what kind of things where exactly tested for most of their claims in their ads.

Frankly get over it, M1 is a great CPU

It is. But whenever some start claiming its so much better than anything else, your going to get counter-claims.
 
Just when I thought that PSman1700 was turning a new leaf, he's gone back to his old ways...

Many claim that Geekbench is created by a Mac-review site. I have no idea if that is true or not, just mentioning it since i hear it everywhere. It is not a usefull benchmark to gauge pc performance as many others have already said, in special not between different architectures and platforms.

Geekbench's developer used to run "Geek Patrol" which focused on Mac hardware. But the site shut down well before Geekbench and well before Apple Silicon. So it's a complete non-point, and anyone who thinks it favours Apple hardware is either stupid (not understanding how hardware works) or deceitful.

So what if these are OC'ed results? Also, why was Apple comparing to old Zen2 products? Theres Zen3, and as of yesterday Zen3 TR up to 64
cores.

Comparisons were made between M1 Ultra and Intel's Comet Lake, Cascade Lake, and Alder Lake.

The comparison to Intel was very weird aswell, they went with a 60watts alder lake to compare to the M1 Ultra and made performance claims.

It's called performance-per-watt.

its good to show efficiency in Apple optimized apps that make use of the media accelerators, but what consumer is going to limit their much cheaper alder lake cpu to 60watts?

This is fake news by this point. Not even going to bother.

Apple never really went into what kind of things where exactly tested for most of their claims in their ads.

It's a one-hour product announcement, not a Gamer's Nexus video. JFC, get over it.

It is. But whenever some start claiming its so much better than anything else, your going to get counter-claims.

It is so much better. Even HUB and JarrodsTech have documented Apple's vastly superior performance-per-watt, and that's despite the performance reduction due to Rosetta2 in some apps.
 
First of all, the Zen2 3990x scores north of 35.000 in Apple's own geekbench 5 app, the M1 Ultra aint even nowhere close. As many on the toms forums wonder is where Apple has these numbers from.
https://browser.geekbench.com/v5/cpu/search?utf8=✓&q=AMD+Ryzen+Threadripper+3990X+

Are you actually comparing a 280w TDP 3990X to M1 Ultra, a SoC made to fit in a 3 litre case which consumes about 60w for its CPUs?

How can anyone take you or this comparison seriously?


The M1 Ultra is in the server/redering station class, not consumer gaming/creator cpu class

What?
It's made for people who work in studios (content creators); definitely not a server or a rendering farm product. Its TDP appears to be approximately 150w, less than the TDP of just a 3990X on its own.

EDIT: 3990X has a TDP of 280w not 180w.
 
Last edited:
Are you actually comparing a 280w TDP 3990X to M1 Ultra, a SoC made to fit in a 3 litre case which consumes about 60w for its CPUs?

How can anyone take you or this comparison seriously?




What?
It's made for people who work in studios (content creators); definitely not a server or a rendering farm product. Its TDP appears to be approximately 150w, less than the TDP of just a 3990X on its own.

EDIT: 3990X has a TDP of 280w not 180w.

Are we comparing 5a 2022 5nm chip to a 2020 7nm chip ? Not only that but this thread ripper is based on Zen 2. Wouldn't it be better to compare it to a more contempary chip like the upcoming Threadripper pro 5000 series based on 7nm and zen 3 ?

https://www.extremetech.com/computi...new-threadripper-pro-5000-cpus-based-on-zen-3

Even then its hard to tell what advantage the 5nm node gives over the 7nm node
 
Are we comparing 5a 2022 5nm chip to a 2020 7nm chip ? Not only that but this thread ripper is based on Zen 2. Wouldn't it be better to compare it to a more contempary chip like the upcoming Threadripper pro 5000 series based on 7nm and zen 3 ?

https://www.extremetech.com/computi...new-threadripper-pro-5000-cpus-based-on-zen-3

Even then its hard to tell what advantage the 5nm node gives over the 7nm node

He should be comparing to class-for-class CPU's. This kind of stuff is what Apple M1 Ultra (lol for naming) should be compared to:

https://www.tomshardware.com/news/i...nFX6S6O2IrVPwnN6UmCeNbpvTk10Z5qjZnuxTF7Uql8D8

What one can give Apple for their M1 lineup is performance per watt. Raw performance? Nah, their quite far off from say Intels or AMD's CPU's. Extremely far off from NV/AMD GPU's.
These M1 ultra/studio boxes that retail for over 7000 dollars are not intended for the 'wide market' just like these Intel Sapphire Rapids cpu's.

Even when considering an Alder Lake CPU, their very competitive against the m1 Ultra cpu prowess, remember that Apple was comparing per-watt achieved performance, but i strongly doubt anyone with a Alder lake intel cpu is going to limit it to 60watts. The alder lake is going to draw more power, but the M1 is going to cost alot more. If it takes 5 years to save that on electricity its probably already time to swap the machine anyway.
 
He should be comparing to class-for-class CPU's. This kind of stuff is what Apple M1 Ultra (lol for naming) should be compared to:

https://www.tomshardware.com/news/i...nFX6S6O2IrVPwnN6UmCeNbpvTk10Z5qjZnuxTF7Uql8D8

What one can give Apple for their M1 lineup is performance per watt. Raw performance? Nah, their quite far off from say Intels or AMD's CPU's. Extremely far off from NV/AMD GPU's.
These M1 ultra/studio boxes that retail for over 7000 dollars are not intended for the 'wide market' just like these Intel Sapphire Rapids cpu's.

Even when considering an Alder Lake CPU, their very competitive against the m1 Ultra cpu prowess, remember that Apple was comparing per-watt achieved performance, but i strongly doubt anyone with a Alder lake intel cpu is going to limit it to 60watts. The alder lake is going to draw more power, but the M1 is going to cost alot more. If it takes 5 years to save that on electricity its probably already time to swap the machine anyway.
It starts at $1999 and the model with M1 Ultra at $3999.

That’s a far cry from “over 7000 dollars”.
 
It starts at $1999 and the model with M1 Ultra at $3999.

That’s a far cry from “over 7000 dollars”.

Base M1 Ultra starts at 50.500kr here (almost 6000 dollars). For the 128gb/64core gpu we are looking much higher than that, over 7.500 dollars at the least (1tb ssd). The 2000 dollar (US) entry is the M1 Max.
These machines are certainly not aimed at the wider market for the average home user. These prices are way above what a Alder lake system will cost, even when coupled with a high-end RTX/RX gpu (msrp), not even offering more performance either, yes more performance per wattage given, but you aint getting more performance overall.
 
@PSman1700 Different localities will have price deviations, but he’s not wrong to say that $4k usd is the launch price of the studio with M1 ultra.

I’ll be curious to see how it stacks up against a 12900k or 5950x build with something like a 3070 or better.

Not sure why people get so worked up about boosting or downplaying these Apple socs. The node advantage is real and accelerators aren’t cheating. If the end user experience is good, then it’s a good product. There will be places where pcs and windows win, especially on the graphics side, but Apple will have its wins too.

Very easy to find game devs and programmers who are very happy with Apple silicon and the performance they get unplugged. Same with video editors etc.
 

That's with every add-on you can purchase. 64-core GPU, 128GB of unified memory and 8TB SSD.

The two standard models are $1999 (M1 Max / 10-core CPU / 24-core / 32GB unified memory / 512GB SSD) GPU and $3999 (M1 Ultra / 20-core CPU / 48-core GPU / 64GB unified memory / 1TB SSD).

I don't quite understand why some of you are so eager to shit on everything Apple here? Why even join the conversation if it's all just negative sentiment?

Base M1 Ultra starts at 50.500kr here (almost 6000 dollars). For the 128gb/64core gpu we are looking much higher than that, over 7.500 dollars at the least (1tb ssd). The 2000 dollar (US) entry is the M1 Max.
These machines are certainly not aimed at the wider market for the average home user. These prices are way above what a Alder lake system will cost, even when coupled with a high-end RTX/RX gpu (msrp), not even offering more performance either, yes more performance per wattage given, but you aint getting more performance overall.

Move to another country I guess? In Denmark the base M1 Ultra model is only $5280 with 25% VAT ($4225 with no VAT). But we both know it doesn't matter what the price is specifically for you. You weren't going to buy it either way.

The basic users don't need the Mac Studio or Alder Lake or a high-end RTX, jeez. The Mac mini at $699 and $1099 MacBook Air is great for general computing / office work. If you are a photographer the base Mac Studio is a straight up bargain at $1999.

When talking about overpriced and low value I will personally add the 2019 Intel based Mac Pro starting at $5999 for just an 8-core Xeon, 32GB RAM, Radeon Pro W5500 8GB RAM and 512GB SSD and the Intel based Mac mini starting at $1099.
 
Last edited:
Move to another country I guess?

Just for Apple products? Other electronics come in at better prices, even consoles (when msrp ofcourse).

In Denmark the base M1 Ultra model is only $5280 with 25% VAT ($4225 with no VAT). But we both know it doesn't matter what the price is specifically for you. You weren't going to buy it either way.

Only? The base model for well north of 5000 dollars still isn't anywhere what you'd pay for a Intel/NV/AMD etc combo. And thats the base model, which atleast here means no 64cores gpu and not the 800gb/s BW.
And no i wont be buying it because i do not do video/image export and creation, for that its the ultimate device, if you use the specific workloads. If your into CUDA then you'd want NV, for example.

The basic users don't need the Mac Studio or Alder Lake or a high-end RTX, jeez. The Mac mini at $699 and $1099 MacBook Air is great for general computing / office work. If you are a photographer the base Mac Studio is a straight up bargain at $1999.

But its the Ultra 64core gpu variant that is being compared to Alder Lake and other x86 cpu's, mainly. And these are very far from eachother in price. Same for the GPU.
Macbook air at 1100 dollars is a good deal for sure (and again here we pay 1500 for the 8gb model), but then again it depends on use-case and what your going to do with it. If its office-work on the go its a great choice in special due to battery. On the other hand for that price you can get a Zen3 5800H/RTX3070m/32gb combo, 120hz ips screen and generally acceptable quality (aluminium case, quite good cooling and build quality) with a asus G15 laptop at around the same price. I wouldnt know why in the world anyone would choose for the m1 air other than ecosystem or battery-life. Performance they are very, very far apart. And then you can game on the asus aswell. With W11, the android integration is really good, its not as solid as Apple's tight integration, but then again your free to use whatever other manufacturer's device you'd like to pair with windows.

I don't quite understand why some of you are so eager to shit on everything Apple here? Why even join the conversation if it's all just negative sentiment?

In the same vein, some of you are shitting on everything that isn't Apple. Look at this one, this post is very much against MS, non-Apple products: https://forum.beyond3d.com/posts/2245352/
He has already doomed non-Apple before that has happened.

@PSman1700 Different localities will have price deviations, but he’s not wrong to say that $4k usd is the launch price of the studio with M1 ultra.

I’ll be curious to see how it stacks up against a 12900k or 5950x build with something like a 3070 or better.

Not sure why people get so worked up about boosting or downplaying these Apple socs. The node advantage is real and accelerators aren’t cheating. If the end user experience is good, then it’s a good product. There will be places where pcs and windows win, especially on the graphics side, but Apple will have its wins too.

Very easy to find game devs and programmers who are very happy with Apple silicon and the performance they get unplugged. Same with video editors etc.

The launch price of the base Ultra, which is not the product that is generally being used to 'demolish' the high-end Intel Alder Lake gaming cpu in the graphs. What about comparing in the same price-class, like we usually get reminded of in the console sections? You'd be looking at the base M1 Ultra at 4000 dollars (US pricing), with that money, what would you get going with an alder lake and RTX gpu system? (msrp ofcourse).

How it stacks up against a 5950x/3070dgpu etc, probably depends alot on what tasks are being used. When the m1's media accelerators are put to use, its going to be crazy fast. It will be the same as we see with the M1 max now, just in a higher performance-tier, were comparing desktop/stationary products this time.

Accelerators arent cheating no, they kinda exist on NV gpus too. It all depends on workloads. What is interesing is raw compute power, perhaps not geekbench but other real world scenerario tests, which will come soon enough.
And to cover the last bit, its that some are dooming Microsoft, intel, Amd, Sony, NV etc because of Apple Sillicon, which i dont find abit early to claim. In special with Alder Lake we see quite nice improvements even on x86, its not all that bad as some imagine it is. And indeed, Apple has its advantages, where they always have had them, content creation/media and its ecosystem, if your into those its probably better then a windows/linux etc machine.
 
My M2 Air arrived Friday and it's fabulous. My only other experience with Apple Silicon is the office's 14" MacBook Pro with a M1Max.

My current/to-be-traded-in laptop is a March 2020 13" MacBook Air with a quad-i7 and I like the new machine better in every respect. I use this machine for some light development (Xcode) and other general tasks like web browsing and this feels as fast as the M1Max MBP to me.

On-chip H.265 1080p encode for most content is between 150-200fps which more than fats enough for my light encoding needs. The M1Max was typically around 240-260fps.
 
Isn't that pretty good for a laptop running on battery?

When I play Cult of the Lamb on my M1 Max (24-core GPU) it only uses around 36W peak at 3840x4320.
Yeah it's an excellent result. Most Windows gaming laptops are much slower on battery and also drain the battery super fast (1.5 hours or less)
 
DF tested RE8 on the desktop M1 Ultra and M1 Max, at max setrings, 4K, no RT, the M1 Ultra is providing performance similar to a 3080m (140w) laptop, M1 Max provides half the fps of the M1 Ultra.

I remember Apple claiming M1 Ultra is similar to a desktop 3090! So obviosuly, this statement is incorrect.

 
Last edited:
Back
Top