Given the source of the article, the author might just have a vested interest in making their own systems look good compared to the Macs?
I don't think they succeeded. The Apple LAPTOPS held up really well against the optimized relatively high end desktops. Every advantage (price, size, cooling, software, architecture maturity) was on the side of the desktop and they are only twice as fast. That's a pretty slim gap.
They need to update the benchmarks to show the performance when they're unplugged from the wall.
"These units only cost around $1,500 as configured, however, so the fact that they are half the performance shouldn't be unexpected since they are also half the cost. What will be very interesting to see is how performance will shake out as Adobe improves support for native Apple Silicon and when (or if) Apple launches a higher-end laptop - or even better, a more powerful desktop variant than the Mac Mini."
"Another way to look at it is that these new MacBooks (and presumably the Mac Mini) are roughly equivalent to a desktop that is around 4-5 years old. So if you have a PC from 2015 or 2016, be aware that these MacBooks likely won't be any faster."
Lots of hand waving at the end. Did this fictional PC also cost $4k 4-5 years ago?
I'm a fan of both PCs and Macs. Just way too much marketing and not enough substance here. Someone wanted their SEO juice from a blog article and they got it.
And you're right, it's not even native binaries. They really should've showed how fast image processing in Photoshop on the iPad was compared to Wintel.
It's just all marketing BS.
Also they compare the apple silicon performances to older systems as well, to have a better comparison point. But again, they could have made a (much) better job of displaying the data.
And still, indeed, Rosettta2 is in the way, so that's no real result here.
It's just not even close to an apples-to-oranges comparison. Not only is it Rosetta2 but it's also Mac vs Windows binaries. They couldn't add in another variable to make it more different if they tried.
But if you were to build/buy a "AMD Ryzen Threadripper 3960X", I would check their numbers on geekbench.com. It seems like there's a lot of builds with the same specs and better numbers than even this. For the amount of money they're asking of a desktop, I'll check these numbers elsewhere.
https://www.anandtech.com/show/16252/mac-mini-apple-m1-teste...
Assuming 75% on average that would increase the M1 benchmarks about 33%, bringing them much closer to the Ryzen Systems tested.
That said, it doesn't seem hard to do and I wouldn't be surprised if most instructions have near 1:1 equivalents. The only question is if Rosetta 2 authors got around to translating those.
https://arstechnica.com/gadgets/2020/11/google-chrome-is-ava...
For photoshopped I have no idea of which of the micro benchmarks have the most predictive ability for Photoshop performance. If it’s mostly memcpy, Rosetta2 is nearly as fast at that as native. But if it’s a lot of math we could see some substantial speed ups from the native Photoshop.
I’m not sold on these yet, but at least read the entire analysis. These 10-25w machines are matching 4-5 year old desktops, or running the same $/performance as current machines in their metric.