Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

2020 Mac Mini – Putting Apple Silicon M1 To The Test

anandtech.com

Post reply on HN
Print view View original

761–767 of 767 posts

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#761

Post by sliken »

m4rtink wrote: post #303
Aperocky wrote: post #184

Earlier quoted context omitted.

No reason why servers will always stick with Intel. Amazon already has their own Graviton ARM chips - And that's EC2, cloud native workflows might have already migrated.

Not to mention HPC, with the top three supercomputers being Power & ARM.

Intel isn't doing to well in HPC either. The top intel system is now at #6, and the vast majority of the performance in inside the matrix-2000 accelerator. The highest pure Intel system is at #9.

Even with high end servers Intel seems to be losing to AMD Epyc.

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#762

Post by ch_sm »

tpetry wrote: post #348
em500 wrote: post #228

Earlier quoted context omitted.

Here's one data point: a WebKit compile took 25min on the Air vs 20min on the Mini/Pro. That 25min is still a bit faster than the Intel 16-inch Pro, which took 27min and waaay faster than Intel 13-inch Pro at 46min. The crazy thing is that both M1 MacBooks still had 91% battery left after the compile, vs 61% on the 16-inch Pro and 24% on the 13-inch Intel Pro. [1] https://techcrunch.com/2020/11/17/yeah-apples-m1-macb…

The interesting information after which time the air was throttled and how much performance is lost when throttling.

Several tests seem to show it throttling after the 8.5-9 minute mark.

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#763

Post by mthoms »

nomel wrote: post #759
mthoms wrote: post #758

Earlier quoted context omitted.

If Apple weren't constantly trying to increase their market share, I imagine their shareholders would like to have a word with them :-) I get what you're saying though. I don't think they should go after the budget PC market. There's still lots of room for growth at the mid to high end. There's also servers.

The goal for the investors isn't to increase market share though, it's to increase profit. I'm questioning the assumption that market share and profit are directly, and linearly, related. The average smartphone price worldwide is around $300 [1]. I don't have access to the full report at that link, but with the graphs shown, some significant portion must be below that price. My naive assumption is that market share i…

>The goal for the investors isn't to increase market share though, it's to increase profit.

Right, but since the wholesale cost of their phones isn't likely to change much at this point, increasing marketshare is the most obvious path to increased profit. Also, with Apple increasingly focused on services revenue, getting more customers into its ecosystem makes perfect sense.

I agree they likely won't go after the low end. This site [0] claims Apple only owns 52% of the high end market so there's lots of room for growth at the mid and high end.

Anyways, since we were comparing Apple to Intel/AMD I assumed we were talking about PC's.

[0] https://www.gizchina.com/2019/12/08/apple-still-holds-first-...

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#764

Post by nomel »

mthoms wrote: post #763
nomel wrote: post #759

Earlier quoted context omitted.

The goal for the investors isn't to increase market share though, it's to increase profit. I'm questioning the assumption that market share and profit are directly, and linearly, related. The average smartphone price worldwide is around $300 [1]. I don't have access to the full report at that link, but with the graphs shown, some significant portion must be below that price. My naive assumption is that market share i…

>The goal for the investors isn't to increase market share though, it's to increase profit. Right, but since the wholesale cost of their phones isn't likely to change much at this point, increasing marketshare is the most obvious path to increased profit. Also, with Apple increasingly focused on services revenue, getting more customers into its ecosystem makes perfect sense. I agree they likely won't go after the low…

> Anyways, since we were comparing Apple to Intel/AMD I assumed we were talking about PC's.

Oh jeez, that's embarrassing. I went off on a tangent without realizing it.

For the PC market, I absolutely agree!

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#765

Post by sliken »

mcintyre1994 wrote: post #429
arvinsim wrote: post #241

Earlier quoted context omitted.

As someone who owns both Mac and PC, I am excited on what Ryzen can offer on 5nm. If Apple has these gains, I am sure Ryzen will have great performance leaps too.

I’m not massively familiar with CPU architectures, would you expect to see similar performance gains going to 5nm on x86 as you do on ARM?

Not really. AMD could make a 7nm version of the Apple core, but they instead build more cores. Much how AMD outmaneuvered intel with smaller chiplets (more flexible in design, higher volume, more tolerant of process errors, higher yield, etc). Apple has done similar with their design. It's better in obvious ways, larger caches, larger number of rename registers, more outstanding transactions, more memory channels, etc. Apple could make a core just as fast, maybe slightly less power efficient if it spent less on the neural cache, image processing, or GPU.

Another big win is that apple runs the memory at 4500 MHz, standard, without overclocking. Even the Zen 3 often runs the ram at 3200 MHz, and standard support goes up to 3800 MHz or so. You can run it higher, but then you have to decouple the memory clock from the CPU clock, which reduces performance. The DDR4x also supports 2 channels, instead of 1. So you get as many memory channels as the AMD threadripper, which is an expensive, hot, and low volume chip.

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#766

Post by thirdsun »

The_rationalist wrote: post #652

while the M1 is impressive, everybody avoid the Elephant in the room which is: they dropped Windows bootcamp support which make it DOA for me. (and no a VM is not a replacement) Zero media report it but there is a BIG performance issue with the M1: Compilers and JIT have had decades of optimization both for ARM and x86. But actually no, code that targeted ARM historically only really was either C, C++, swift, js and…

> While the M1 is impressive, everybody avoid the Elephant in the room which is: they dropped Windows bootcamp support which make it DOA for me. (and no a VM is not a replacement).

I'm not quite sure if it's even a mouse in the room. While it certainly is a problem for you and a niche group of like-minded users, the vast majority of Mac users, including me, simply don't care about Windows support.

Plus, there is good hardware to be found on the Windows side as well, so it isn't the end of the world in my opinion.

Re: 2020 Mac Mini – Putting Apple Silicon M1 To The Test

#767

Post by moz42 »

yxhuvud wrote: post #126
Thaxll wrote: post #10

I don't think there is enough test yet, why is it only Cinebench and Geekbench? Show us real test with ffmpeg, gcc etc ...

Does ffmpeg even compile on M1 yet?

Yes, seems so. Looking forward to check out how this will perform.

http://www.ffmpeg-archive.org/FFmpeg-on-Apple-Silicon-Succes...

Post reply on HN