News for hackers. Stuff that matters.
Re:2022 Mac Studio (20-core M1 Ultra) Review by 2OEH8eoCRo0 on Thursday April 28, 2022 @05:03PM (#31198189)
Re:2022 Mac Studio (20-core M1 Ultra) Review by joakleaf on Thursday April 28, 2022 @05:58PM (#31198716)
That's kind of disingenuous.

Searching on Geekbench for Apple M1 ultra single core scores returns values mostly in the 1770-1780 range. E.g. https://browser.geekbench.com/v5/cpu/14597244

Most 12900K score are between 1900 and 2200 but then there is this outlier with single core score of 1252: https://browser.geekbench.com/v5/cpu/14572307

Intel certainly wins on single core, but the m1 Ultra multicore scores are still impressive in comparison being generally 23-24000, while the 12900k are around 15-20000.

Re:2022 Mac Studio (20-core M1 Ultra) Review by hu3 on Thursday April 28, 2022 @06:02PM (#31198753)
So 1900-2200 for Intel and 1770-1780 for M1?

Disingenuous would be to focus on the outlier.

Re:2022 Mac Studio (20-core M1 Ultra) Review by joakleaf on Friday April 29, 2022 @02:03AM (#31201922)
An outlier for m1 ultra was what was reported by parent.
Re:2022 Mac Studio (20-core M1 Ultra) Review by sliken on Thursday April 28, 2022 @06:19PM (#31198921)
Sure, Intel focuses on single thread perf, high power (241 watt max tdp), and automatically overclocks to 5.1 GHz, only if you have enough power, cooling, and a bunch of idle cores. Thus the 15% variation in submitted scores. It's also rather memory bandwidth constrained, and shows impressive numbers with a single core running.

Apple on the otherhand doesn't overclock, focuses on multi-core performance, has great memory bandwidth, and all the submitted scores are within 1%.

The M1 ultra is also 1.32x faster in the multiprocessing benchmark. Looks pretty impressive to me, even ignoring the much less power the M1 ultra uses.