M2 Ultra in the first benchmarks: the numbers are with Apple, and Mac Pro goes double


The first results in the benchmarks for M2 Ultra of Apple carry it above any Apple Silicon chip. The product announced during WWDC 2023 passed on Geekbench 6 showing what it is capable of. Taking the results for granted, we see how Apple’s latest and most powerful Arm chip arrives at overcome there threshold of 2,800 points in single core and 21,000 in multi core.

The Apple Silicon chip that before the event of recent days was the most powerful in the range, M1 Ultra, was just over 2,300 points in single core and 17,000 in multi core, so if the Geekbench numbers on M2 Ultra were reliable, it would be of an increase in single core of about 17% and of almost 20% in multi core, percentages in line with those announced during the WWDC 2023 keynote where it was also said that the 76-core GPU was up to 30% faster and the 32-core Neural Engine up to 40%.

The numbers on the M2 Ultra are doubly significant, because the chip “signs” the first generation Mac Pro on Arm architecture. Apple’s most powerful tower had remained the only representative in the range of the old Intel course, but now he too has abandoned the x86 architecture for the new Apple Silicon branded Arm course. And comparing the previous Mac Pro to the latest, one emerges notable difference especially if the savings at the cash register are taken into accountwhich reaches levels that are almost embarrassing for the previous generation Mac Pro if you look at the maximum configurations.

The old Mac Pro with Intel Xeon W processor stops at about 1,370 points in single core and just over 10,000 in multi core, which means, using the scores that would have been measured by Geekbench, that the new Mac Pro “goes” twice as fast costing less even in the basic configuration.

Previous articleThe RTX 4060 would be released earlier than expected, see you on June 29, 2023
Next articleTwitter will pay verified creators for ads shown in their responses
Expert tech and gaming writer, blending computer science expertise