Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Mac Studio M2 Ultra vs M2 Max Geekbench single-core score discrepancy

K

kgapp

macrumors regular
Original poster
Jan 6, 2004
107
7
Chicago, IL
Following last week's product announcements I purchased one Mac Studio M2 Ultra with 60 core GPU and 128GB Ram (Mac 14,14) as well as one Mac Studio M2 Max with 38 core GPU and 96GB Ram (Mac14,13).

I set both machines up identically (via migration assistant from my M1 MacBook Pro). Once both machines were up and running and had a day or two to spotlight index, iCloud sync, etc. I started doing some benchmarking on the two machines against each other.

I am having trouble understanding the results of the single core Geekbench 6.1 that I am seeing from the two machines I have. Admittedly, I am not a benchmarking expert but I know enough to work from a clean boot, quit all apps, etc. prior to running the benchmarks.

My Mac Studio M2 Max is consistently outperforming my Mac Studio M2 Ultra for single core CPU Geekbench 6.1 scores. I have run the tests a few times now on different days and I get consistent results in the Mid 2600's on the M2 Ultra while I get scores in the low 2800's on the M2 Max.

Does this make sense? My understanding is that I should be seeing nearly identical single core performance on both machines. Obviously there may be small fluctuations from run to run but in my case the M2 Ultra is consistently 10% slower than the M2 Max in single core scores. A summary of my scores from both machines are below.

As of today I am not yet seeing any Mac Studio M2 scores in the Geekbench 6 Mac Benchmarks repository so I can not yet compare my personal scores to other similar machines. I can only compare to what I have seen others post in various threads.

Either way these machines absolutely SMOKE the performance of anything I have used before and am I extremely happy with both of them. Just trying to understand if what I am seeing would be considered normal or not.

Any thoughts or comments on my score discrepancies are welcome.

1686936566485.png
 
  • Like
Reactions: greenteapanda
It’s interesting the CPU frequency fluctuates so much, but you’re right that these are very small differences and I wouldn’t agonize over them.

Your M2 Max is scoring ~3% higher than similar MacBook Pros so perhaps it’s enjoying a little bit more thermal headroom
 
  • Like
Reactions: greenteapanda, macguru212 and Icelus
I can confirm the base m2 ultra spec scores that I had tested below.

Single core 2691
Multicore 20959

There is someone else that posted here showed the high spec'ed version to be inline i think with m2 max for single core.
 
cpnotebook80 said:
I can confirm the base m2 ultra spec scores that I had tested below.

Single core 2691
Multicore 20959

There is someone else that posted here showed the high spec'ed version to be inline i think with m2 max for single core.
Click to expand...

I did see that score posted and a few others with the 76 core GPU that were all in the low 2800's. That had me scratching my head a bit as to why my 60 core GPU version and yours appear to be running a bit lower on the CPU benchmarks.

I guess we will have to wait for more reports from other M2 Ultra users but from initial glance it appears that for whatever reason the M2 Ultra 60 core GPU and 76 core GPU variants are in fact producing different CPU benchmarks. I expected different Metal and OpenGL results and nearly identical CPU benchmarks but that does not appear to be the case for whatever reason.
 
It could be the xtra RAM - try taking out 32 GB from the Ultra... oh whoops, I guess you can't do that. My old MacPro ran faster with 96 GB vs 128GB -- apparently due to the memory lanes in that machine.
...so RAM could be a factor in the Single core test, but I have no idea why that would be. So I'm curious as to how many apps are running in single core mode ?
 
  • Like
Reactions: macguru212
There aren't many data points, but, if you look at Geekbench and results with usernames attached, there's a pretty solid pattern of 38/76 GPU core CPUs just over 2800 and 30/60 GPU core CPUs in the mid 2600s. It also looks like there's a clock speed difference across the two (3.68 vs 3.49) that's roughly in line percentage-wise with the ~5.5 percent faster single core speeds.

The benchmark charts make it look like the 16-inch MBP with M2 Max is also clocked a little higher than M2 Pro chips, so perhaps that's just a change this generation that hasn't gotten a lot of attention.

It makes the small M2 Max 30-to-38 GPU core upgrade cost look even better, and it makes the M2 Ultra 60-to-76 GPU core upgrade look slightly less bad (but still not great). If you're on the fence between the Max 38 and Ultra 60, it makes the Ultra look a bit less appealing.
 
  • Like
Reactions: Mac Hammer Fan
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back