How M1 Pro and M1 Max differ from M1 are shown in the core images below:
The number of large nuclei in both (new) cases is eight, small nuclei are present two. Apple’s philosophy is somewhat different from Intel’s. While Intel compensates primarily for multi-core performance with a high number of small cores, Apple uses small cores primarily to save power (when high performance is not required); they are rather marginally signed on the overall performance.
However, the main reason for the difference in the area of both chips is the integrated graphics. It increased from (up to eight) blocks for M1 to (up to) 16 blocks for M1 Pro and to (up to) 32 blocks for M1 Max. Apple boasts that it outperforms the Playstation 5 in overall TFLOPS (10.4 TFLOPS M1 Max vs. 10.3 TFLOPS Playstation 5), but that doesn’t say anything about real graphics performance. But due to the graphics part, we are not here today, let’s look at the performance of the processor:
GeekBench 5
graph-13
The multi-core performance of the Apple M1 Max catches up with the Core i7-12700K samples from the Intel generation Alder Lake. I dare to estimate that this is a similar level of performance as the 65W Ryzen 9 5900 (unfortunately we do not have its results in GeekBench 5 – do you not run someone Ryzen 9 5900X in ECO mode? The score should not be much lower than the native 65W Ryzen 9 5900 – you can mention it in the discussion for comparison).
graph-14
Single-core performance is between Zen 3 a Alder Lake. The advantage of the Apple M1 Max will probably be lower consumption related to 5nm production at TSMC. Zen 3 a Alder Lake are 7nm / Intel 7. The disadvantage, of course, will be the absence of a socket and the link to Apple products, which will be traditionally somewhat differently priced elsewhere.
–
