It is. Here's what Intel thinks a desktop with a 13K should look like:
That thing is 318mm x 129mm x 337mm. A Mac mini is 35.8mm x 197mm x 197 mm. IOW, the comparison is silly.
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.
First Geekbench Scores for New Mac Mini With M2 Pro Surface, Beats M1 Max
- Thread starter MacRumors
- Start date
- Sort by reaction score
It is. Here's what Intel thinks a desktop with a 13K should look like:
That thing is 318mm x 129mm x 337mm. A Mac mini is 35.8mm x 197mm x 197 mm. IOW, the comparison is silly.
First of all, there’s the intel nuc and other small formfactor PCs from minisforum and AsRock which have a much smaller form factor and are very powerful and potent for their size.It is. Here's what Intel thinks a desktop with a 13K should look like:
That thing is 318mm x 129mm x 337mm. A Mac mini is 35.8mm x 197mm x 197 mm. IOW, the comparison is silly.
Second of all, apple said it had to leave intel in the dust because performance was lagging. But what you must ask yourself is, given that intel has faster raptor lake chips, and more efficient and powerful chips coming like meteor lake and arrow lake, and amd also has server chips that could’ve gone in the Mac Pro (Genoa zen 4 96 core monster), is apple really better off (on its desktop lines) without x86?
In laptop, apple silicon is very impressive. But on desktop, nvidia intel and amd are leaving it in the dust. M2 max GPUs are much slower than nvidia Lovelace. Amd 7900 and intel 13600K are as fast if not faster than m2 pro/max cpu in multithreaded tasks. And both have faster CPUs available.
Amd zen 5 and intel arrow lake are coming which will be much faster than existing lines. So from the desktop perspective, m2 pro/max aren’t as impressive vs the competition. But in laptop, they’re very impressive.
In my humble opinion of course.
Last edited:
None of them can fit the 13600K.
I like to compare products that actually exist.
Apple’s desktop SoCs run at the same clock and at relatively low wattage. Apple could do a 5 GHz 200W SoC that would smoke Raptor Lake-S. They just aren’t interested in that market, and comparing the Mac mini with it is silly.
I was just going to ask if any of them had a several hundred watt power supply
They aren’t interested in that market yet they offer a Mac Pro with a Xeon. Internal inconsistency.
What you pay for hydro or electric bill month to month, year to year over the lifespan of your computer.
comparing the same usage of the same products in your home and ONLY changing your computer - not your computing and other electronic device habits you could see a significant savings and more $ to spend on other items.
I can assure you those that had the Power Mac G5 towers and switching to Intel Mac Pro's had a significant cost savings in their yearly electric bill close to $300+/yr if not more. x5yrs that's $1500, if not spent elsewhere over that time ... helps for the next upgrade.
If it don't make dollars, (or save ya dollars), it don't make sense.
Intel is also very impressive on the energy bill, too.
When you look at it you'll be shocked at how much those chips consume.
Definitely.
Last edited:
For something like a Mac mini, absolutely. We know that Mac Mini can handle around 65W of system power, maybe a few watts more for short bursts. With Intel, the fastest CPU that would get you is the i9-12900T (desktop variant, very bad GPU) or a mobile i9-12900HX. Both of them have multicore performance similar to M2 Pro at the cost of 2x power consumption (in fact, the Mini chassis probably won't be able to handle these Intel CPUs at turbo, so you can safely remove another 10-20% from their multi GB5 scores). And of course, you'd have to live with much slower Intel GPU. AMD is not that much better either. Most that could be expected in the Mini would be a 12-core 7900 (which would need to be underclocked slightly to fit a true 65W limit), which again performs very similar to M2 Pro but needs much more power. Or one of the new Dragon Ridge mobile chips with yet unknown performance.
For large machines the merit of Apple Silicon still remains to be seen. On the so-called "enthusiast desktop" x86 currently reign supreme, mostly because the marketing is willing to sacrifice any remaining pretence of energy efficiency and the customers don't seem to mind that the current-get cutting edge hardware uses 3x more power under load than just a few years ago. Then again Apple never competed in this market segment. If they make the right choices they should have decent chance in doing well in a personal workstation market where things are a bit different. At any rate, if we look at the performance per watt, which is a crucial factor for this type of workstations, Apple currently holds the best tech in the world by a wide margin. The question is how much that market is worth to them.
No. Don't get the Max unless GPU matters a lot.
Barely at all.
In the most ideal cases, saturating all cores, Anandtech was able to pull 243 GiB/s out of the M1 Max CPU:
So, even then, it doesn't actually come close to using the 400. It's barely above the Pro's 200.
Not worth it. The 400 GiB/s (instead of 200) are almost entirely for when you need the GPU, not for the CPU.
Now, it's possible the M2 Pro/Max are a bit different in this regard than the M1 Pro/Max, but I don't expect it.
Well https://www.gpu.audio FX utilise Apple Silicon GPU's.
Tried it on my M1 and it tasked the gpu like 99% before dropouts, CPU not even moving.
So GPU might not be completely wasted for audio in near future. However, i don't expect all devs to jump on utilising GPU.
I would think GPU for AS Macs will be used more as even the base M1 has a good GPU where as previously the best selling Intel Macs have weak GPUs. Development effort spent targeting the M1 GPU will benefit all AS Macs.Well https://www.gpu.audio FX utilise Apple Silicon GPU's.
Tried it on my M1 and it tasked the gpu like 99% before dropouts, CPU not even moving.
So GPU might not be completely wasted for audio in near future. However, i don't expect all devs to jump on utilising GPU.
You are comparing 16-core i7 and 24-core i9 with 12-core M2 Pro. 24-core M2 Ultra should go over i9 based on these benchmarks.Better than expected but still behind Intel 13th gen. Apple needs to bring their A Game for M3 chips.
Edit: Adding some benchmarks. The i9 chips are not out yet but leaked and the i7 are already on par with M2 Pro 12 core.
I’ll say. It looks like the M2 Pro does with 12 cores what the i7 needs 16 cores for. And 78% of what the (not out yet) i9 does with twice the cores. All the while using less power.
BTW, the M3 Pro mops the floor with the i9
Future competitor products beat current Apple products every day.I’ll say. It looks like the M2 Pro does with 12 cores what the i7 needs 16 cores for. And 78% of what the (not out yet) i9 does with twice the cores. All the while using less power.
BTW, the M3 Pro mops the floor with the i9
Including future Apple products.
Your argument makes no sense.
ETA: I just now got your joke! Me, slow of brain!
Last edited:
Anandtech just posted a review of the 13900KS: https://www.anandtech.com/show/1872...ks-review-taking-intel-s-raptor-lake-to-6-ghz
Great scores at high power and heat. What I don't understand, is why Anandtech doesn't use the stockfish benchmarks... Won't somebody please think of the stickfish users?
It seems to me that people are generally under-estimating the effect on battery life in every day use with the new chips, probably because of 4 vs 2 efficiency cores. The Verge is reporting 14-18 hours of actual use with M2 Max, vs 10 hours with M1 Max. That is a huge difference, which will be meaningful for a very large group of normal users, as opposed to theoretical battery drain under max load. The fact that M1 always had four efficiency cores is quite interesting in that regard. This is the part of Apple Silicon that I'm most excited about, Apple clearly understands that it's not just about optimizing for best performance under ideal conditions. Optimizing for everyday use is clearly a priority.
Not only that; each of those efficiency cores also got a lot faster (unlike the p-cores). Something like 30-40%.
Ran my M2 Pro through Geekbench. Results - Single core 2685, but multicore is only 12281. Whilst the single core score is great, not sure why Multicore so slow. Anyone else with similar results?
Attachments
Last edited:
Because you ran Geekbench 6 while previous scores are Geekbench 5 scores.
Geekbench 6 uses a new baseline.
See: https://forums.macrumors.com/threads/primate-labs-launches-geekbench-6-benchmarking-suite.2380767/
You can't compare Geekbench scores across major versions. You're on 6; this article is about 5.
Your result is very typical of an M2 Pro on Geekbench 6. Nothing to worry about.