Apple M1 SoC

M1 Ultra should be a monster at rendering because of available VRAM.

Redshift has their Moana Island Scene and it does crazy good on M1 Max.

2 x 1080 Ti = 77m
2 x 2080 Ti = 34m:17s
1 x 3090 (24GB) = 21m:45s
2 x 3090 = 12m:44s
M1 Max (64GB) = 28m:27s

M1 Ultra would probably halve that to 14 minutes.

Times taken from this thread.

A Redshift developer do mention some issues with the default benchmark here.

Please note that the benchmark is using 128x128 blocks which are not ideal to the M1 chips (due to their pretty high latencies). So the benchmark (as it stands today) might not show you particularly good scaling. The same is true for the M1 Pro/Max - but to a lesser extent.

As a stopgap solution I might look at adding a blocksize parameter to the benchmark. Or force it to 256 or larger if apple silicon is detected.
Click to expand...
 
Last edited:

Quite the difference compared to a 3080/3090 in the updated blender benchmark. In dave2d's other tests, the m1 Ultra 64c wasn't so far off from 3080/3090 and a 12900k combo (1132 for the ultra, 5552 for the 3090).
Would be a nice little pc i think if it could run windows natively, play games as intended etc. If it was at a lower-than 6000/7000 dollar price tag aswell then for the 64 core gpu ultra.

Dave2d's end-take is that, if you have a clear use for it, its a good one.
 
The GPU scaling is not that good between M1 Max and Ultra in most compute workloads. Especially in games, the famous multi GPU stutter is present in full force on the M1 Ultra.

so Tomb Raider was not a perfect or even particularly good experience: there was substantial, noticeable micro stutter at every resolution we tried.
Click to expand...
https://www.theverge.com/22981815/apple-mac-studio-m1-ultra-max-review

In compute, with Geekbench 5 using OpenCL, the RTX 3090 is 250% faster than M1 Ultra, keep in mind the OpenCL is not that best result for the 3090 as NVIDIA has very weak support for OpenCL. The scaling from M1 Max to Ultra is only 31% using OpenCL and 46% using Metal.

Blender without RT, the RTX 3090 is 4.9 times faster than M1 Ultra. The scaling from M1 Max to Ultra is only 60%.
 
Last edited:
For ultra-mobile, its not that weird with 2 performance cores i think. Iphones have been running with 2 performance cores/4 efficiency cores for a while now.
 
Rootax said:
What is the target audience for the M1 Max ?
Click to expand...

A lot of creative users fall in that segment, photographers, video editors, graphic designers, print and press etc. Basically anyone that runs Adobe CC or similar.

People seem to be preoccupied with only Blender and games here for whatever reason.

Zbrush performance in Photoshop are great on these SoCs.
 
It appears it doesn't max out power usage for neither the CPU or GPU. Blender only uses around 30-40W instead of up to 120W the GPU is capable of. CPU maxes out at 3064Mhz instead of 3.2Ghz.

I suppose we are waiting on some software updates, even from Apple.


I just ran 3DMark Wild Life Extreme and GPU power on my M1 Max (24-core GPU) topped out briefly at 28W, otherwise were in the 12-20W range. My FPS were 60FPS (max my screen does) and my score were 10020.
 
Last edited:


Power draw maxed, all media encoders running. Intel system is faster in Adobe aswell. Seems that dual chips being seen as one isnt all that smooth.
In cinebench the 12900k remains much faster in both single and multicore performance.
Its suprising to me that in Blender, just using the CPU, the ADL was faster than the m1 Ultra.
 
Last edited:
PSman1700 said:


Power draw maxed, all media encoders running. Intel system is faster in Adobe aswell. Seems that dual chips being seen as one isnt all that smooth.
In cinebench the 12900k remains much faster in both single and multicore performance.
Its suprising to me that in Blender, just using the CPU, the ADL was faster than the m1 Ultra.
Click to expand...

You need to use the built in power metrics tool in macOS to read CPU and GPU power draw. Wall power doesn't really tell you anything other than it doesn't max out clearly. The PSU is 370W in the Mac Studio.
 
milk said:
The more I look at this display, the worse it's design looks.
Click to expand...

Design is personal, could be nice to someone could be horrible to someone else. What mkbhd means is the price/performance, you can get very capable oled/high refresh rate displays for much less money. It is like its accompanying mac studio, it is for Apple-ecosystem users, and for them its a good deal.
 
PSman1700 said:

Quite unplayable, down to 23 fps or below, lags alot. Reviewer concludes its too high settings.
Click to expand...
That's Strange isn't even a15 already strong enough for sustained 60 fps genshin? Like on iPad mini

Hmm I wonder... Whether future iPad mini with m1 will be worth it or it'll be just a small jump from a15. Due to limited tdp setup
 
You must log in or register to reply here.

Similar threads

W
LLMs may force Apple to integrate more RAM into A-series SOCs
Replies
8
Views
2K
MfA
M
M
Batman Quest 3 is a graphical showpiece and a great game
Replies
2
Views
385
mavox01
M
L
Iphone SOC
Replies
17
Views
3K
Pressure
P
S
Deadlock (PC ... not sure what else)
Replies
3
Views
512
GhostofWar
G
G
Nvidia G-Assist
Replies
2
Views
1K
Davros
D
Back
Top