Maybe that is why Apple did not release an updated Mac Mini with M1 Pro/Max chips. Then the desktop comparison could be legit (even though the mini is still not a workstation class machine).
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.
So M1 Max with 32 Core GPU matches RTX 3080 Mobile.
- Thread starter iBug2
- Start date
- Sort by reaction score
Sort of, nvm not being a workstation class machine, which is understating things a bit, … a desktop 3090 is bigger than the Mac mini
When one component is bigger than your entire computer and still potentially trades blows with it.Sort of, nvm not being a workstation class machine, which is understating things a bit, … a desktop 3090 is bigger than the Mac mini
I would hope Apple would use the desktop aspect (always on wall power, better cooling capability) to offer a Mac mini that was always in High Power Mode (and therefore have higher performance than the MacBook Pros)...?
This is from Geekbench too:
The Geekbench OpenCL I think, for a Mac Pro 2012 5,1, (tower) with a 5700 in it.
65754
I'm looking forward to some work being evaluated. yes, I know benchmarks is all we have at the moment ... but ... I'm sceptical about their relevance to real work flows, actual productivity, which are not about a single point of reference.
I wonder too - sorry way off topic - but will the new Mac Pros have replaceable M CPUs on daughter boards? That would be cool.
The Geekbench OpenCL I think, for a Mac Pro 2012 5,1, (tower) with a 5700 in it.
65754
MacPro5,1 - Geekbench
Benchmark results for a MacPro5,1 with an Intel Xeon X5690 processor.
browser.geekbench.com
I'm looking forward to some work being evaluated. yes, I know benchmarks is all we have at the moment ... but ... I'm sceptical about their relevance to real work flows, actual productivity, which are not about a single point of reference.
I wonder too - sorry way off topic - but will the new Mac Pros have replaceable M CPUs on daughter boards? That would be cool.
Attachments
Last edited:
They cater to a PC audience, and perhaps even more importantly, lives off advertising from the PC industry.
Tech reviewing on the Web (nevermind Youtube) is fundamentally corrupt. Where it isn’t, it’s only due to personal integrity.
By the way, I don't know if it was mentioned before, but I would like to point out that these M1 Max scores are very similar to AMD's 6700M — another 10TFLOPS GPU. So again what we have here is that Nvidia GPUs somehow perform better in Geekbench, either because Ampere's dual ALUs give it a big advantage, or because there is something in Geekbench itself that favors Nvidia.
My background is computational chemistry. Quantum Mechanics, Molecular Dynamics, that sort of thing. When GPGPU started up, I took an interest in it, and it soon became clear that (unsurprisingly) in many of the cases where ”GPU-accellerated” was the headline, the GPU itself had little to do with the speed-up, the code was actually memory bandwidth accellerated.
I was around back at the time of formation of SPEC, and (again unsurprisingly) the early SPEC-fp subtests were, quite appropriately, strongly influenced by memory bandwidth. But that makes for lousy benchmarketing, so over time the suite was reworked to be far less influenced by the memory subsystem. Commercial interests at work.
Enough intro - I was very happy to see the memory architecture of the M1PM. The bandwidth, especially in combination with the humongous SLC really lends itself to algorithmic experiments, and I’m sure we were several dozen ? chemists who raised eyebrows when the presentation hit those points. Well, maybe even more.
But I have no idea what the codes in other computationally intensive branches of science and technology looks like. Could you comment on the value of a flexible setup like the M1PM in your field?
Folks folks folks.
This is the first time in history.
First time.
Like man landing on the moon.
First time we have workstation performance in a laptop that doesn’t burn your lap.
The promise we have all been waiting for.
Geekbench 65,000 or Geekbench 70,000. Makes no difference.
The difference is peace of mind.
No loud fans.
No burning heat.
No throttling.
The only loud noise and heat is coming from Dell, Intel and Razer HQ. They are screaming and having a fever.
This is the first time in history.
First time.
Like man landing on the moon.
First time we have workstation performance in a laptop that doesn’t burn your lap.
The promise we have all been waiting for.
Geekbench 65,000 or Geekbench 70,000. Makes no difference.
The difference is peace of mind.
No loud fans.
No burning heat.
No throttling.
The only loud noise and heat is coming from Dell, Intel and Razer HQ. They are screaming and having a fever.
This Asian brother making me jealous! Why Asian brothers always getting the good stuff first hmpppph
That machine looks so solid.
Geekbench is recognizing the number of cores (CPU & GPU) just fine.
8-core CPU (6+2) - Downclocked?
14-core GPU
24-core GPU
8-core CPU (6+2) - Downclocked?
14-core GPU
24-core GPU
Right in what way? That M1 Max is faster than the RTX 3090?
Its definitely a very decent GPU and a very good offering in the professional market. I am sure it will excel at creative work, and I think that these new Macs would be ideal workstations for CAD, if only there was suitable software. But they are not magically faster than large desktop GPUs and also not faster than ultra-high-end laptop GPUs.
no, he is saying how Apple reach that their gpu is right between the 100W 3080 and the 160W 3080Right in what way? That M1 Max is faster than the RTX 3090?
Its definitely a very decent GPU and a very good offering in the professional market. I am sure it will excel at creative work, and I think that these new Macs would be ideal workstations for CAD, if only there was suitable software. But they are not magically faster than large desktop GPUs and also not faster than ultra-high-end laptop GPUs.
He never says anything about 3090 and how M1 max is faster..wtf ?!
Last edited:
Lol, I think you should wait two more days and get the real numbers when the new MBPs are out. All this extrapolation is truly tricky as software may cater more to specific hardware and we just don’t know. Having said that, I think you can take the M1 and extrapolate CPU performance but not GPU.
I was referring to the video thumbnail. As to the rest of it, it kind of depends on what one means. As I said, I am quite confident that in creative workflows (or even gaming) the M1 Max will be very close to the fastest mobile RTX 3080 variants. Not so much for pure straightforward number crunching, where it seems to sit somewhere around mobile RTX 3060/3070
I think the fact that apple is suddenly putting a lot of effort into getting blenders cycles renderer using metal API is a very good sign that these new GPUs will be great for 3d creative work
So you are confident that will be around 3070, and probably close to the 3080 mobile
Still a big jump no matter how you see it...the biggest jump from 1 generation to the new one in history i think
Did you order something?
Aren't they made in China? That guy is Vietnamese.
Anyways he's a half day ahead, and it sounds like he got his package early.
That happened to me way back with my iPad Air 2. I got my delivery the night before the official launch date. The shippers are supposed to hold the packages until launch day, but occasionally they slip through.
While CyclesM will be awesome, I would like to also have Metal acceleration in the viewports...!
yes hopefully thats in the works. I was under the misunderstanding that cycles metal would also be viewport acceleration
I am probably using the wrong terminology or whatever, but I just want the viewports in Blender to be as fast as possible, so one can see more of what they are really working towards without having to send off test renders...?