Re:Apple fans returning new MacBook Pros with Core i9
by valarauca1 on Friday July 20, 2018 @03:32PM
(#17578170)
Geekbenchmark runs it tests in bursts with short delays between them to allow for the SoC to dissipate heat. They implemented that so laptops and cellphones wouldn't overheat, or thermally throttle while running the benchmark.
https://www.xda-developers.com/geekbench-ceo-fireside-chat-p...
I get that Geekbenchmark is extremely accessible, but it is not rigorous.
Citing a benchmark that's designed to avoid thermal issues, as proof a chip doesn't have thermal issues is just bad.
Re:Apple fans returning new MacBook Pros with Core i9
by dullin on Friday July 20, 2018 @03:45PM
(#17578290)
The article in question above is from the geekbench team compiling their geekbench suite over and over to force the overheating issue.
The blog post is not about benchmarking with geekbench at all.
John Poole found that using only CPU bounds tasks (compiling) wasn't that bad for the thermals.
Re:Apple fans returning new MacBook Pros with Core i9
by karanlyons on Friday July 20, 2018 @03:43PM
(#17578272)
My tests do not show this, but show that putting the CPU under load alone is enough to get rather disheartening results: https://imgur.com/a/ehtleu4
https://www.geekbench.com/blog/2018/07/macbook-pro-mid-2018-...