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

Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

wccftech.com

Post reply on HN
Print view View original

81–90 of 448 posts

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#81

Post by wtallis »

mrangle wrote: post #76
znpy wrote: post #49

Earlier quoted context omitted.

Just wondering, what is the cpu that you picked? But I agree, tdp should really be kept in consideration.

Look at the i3. Though it isn't much different than last gen, and that buying decision should be mostly determined by price (for the same price, go later). Unless there are significant enough non-speed related architecture differences that I am not aware of and that are important to you (possible). Try to catch a sale. Going upline will double the TDP for little gain, especially if dollars spent is important as well.…

> Look at the i3. Though it isn't much different than last gen

IIRC, the desktop i3 and i5 processors are in many cases literally last-gen, since Intel is using a mix of Alder Lake and Raptor Lake dies on most of the "13th gen" desktop products that do not have more cores or cache than Alder Lake did.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#82

Post by StewardMcOy »

seanmcdirmid wrote: post #67
adastra22 wrote: post #2

50% slower than the absolute best of class NVIDIA discrete GPU offering that consumes massively more wattage is hardly a damning statement.

Wait, a $7000 computer doesn’t come with a discrete GPU?

It's not just that it doesn't come with a discrete GPU. Apple Silicon doesn't support discrete GPUs to begin with.

As far as I understand it—and this is just from watching Apple's presentations on the architecture—the lack of a discrete GPU is a big part of how the Apple Silicon machines achieve good performance per watt.

Instead of having discrete RAM or a discrete GPU with its own VRAM, all of the RAM is accessible to the CPU and and the GPU in a unified memory architecture. On the M2 Ultra, this allows for 800 GB/s of memory bandwidth, and also eliminates a lot of the need to copy data from RAM to VRAM, as both the GPU and the CPU can access the same memory. In return, this allows the GPU to match the performance of discrete GPUs that have a lot more cores.

Of course, the big downside is that you can't expand the RAM or install a beefier GPU. It's all baked in to the logic board.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#83

Post by mrangle »

jacooper wrote: post #18
murmansk wrote: post #5

Ahhaha, Intel's TDP is 2-4 times bigger under full load (and I guess ~10-15x under moderate load) while having negligible perf advantage over M2. The author can not be serious.

Since when was TDP a deciding factor in desktop purchasing decisions? Sure M SOCs are great on laptops, but there are no reasons to get them on the desktop, slower, more expensive and with less compatibility.

TDP reduction = heat management = reliability management but mostly noise management.

I've never tried to water cool a 125 watt processor and so I can't speak to that. But especially if one uses air cooling or ideally passive cooling, one's ability to reduce noise is proportional to the chip's TDP. Noise reduction is important to many.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#84

Post by wtallis »

StewardMcOy wrote: post #82
seanmcdirmid wrote: post #67

Earlier quoted context omitted.

Wait, a $7000 computer doesn’t come with a discrete GPU?

It's not just that it doesn't come with a discrete GPU. Apple Silicon doesn't support discrete GPUs to begin with. As far as I understand it—and this is just from watching Apple's presentations on the architecture—the lack of a discrete GPU is a big part of how the Apple Silicon machines achieve good performance per watt. Instead of having discrete RAM or a discrete GPU with its own VRAM, all of the RAM is accessible…

> and also eliminates a lot of the need to copy data from RAM to VRAM, as both the GPU and the CPU can access the same memory. In return, this allows the GPU to match the performance of discrete GPUs that have a lot more cores.

Plenty of PC hardware reviewers have done sensitivity analysis experiments to see how discrete GPU performance is affected by running with a slower or narrower PCIe link. The consensus is usually that GPUs connected by PCIe have more than sufficient bandwidth, and cutting it in half only affects gaming framerates by a few percent. Tighter coupling between CPU and GPU can plausibly have a bigger impact for some GPU compute workloads, but for traditional 3D graphics it doesn't help performance much.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#85

Post by mrangle »

wtallis wrote: post #81
mrangle wrote: post #76

Earlier quoted context omitted.

Look at the i3. Though it isn't much different than last gen, and that buying decision should be mostly determined by price (for the same price, go later). Unless there are significant enough non-speed related architecture differences that I am not aware of and that are important to you (possible). Try to catch a sale. Going upline will double the TDP for little gain, especially if dollars spent is important as well.…

> Look at the i3. Though it isn't much different than last gen IIRC, the desktop i3 and i5 processors are in many cases literally last-gen, since Intel is using a mix of Alder Lake and Raptor Lake dies on most of the "13th gen" desktop products that do not have more cores or cache than Alder Lake did.

Right. Price should determine what one buys. There is a marginal performance bump between them for the same advertised TDP, and so suppose if price is the same then next gen it is. Otherwise buy what is cheapest.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#86

Post by locustous »

daneel_w wrote: post #57
peoplefromibiza wrote: post #52

Earlier quoted context omitted.

> Economically and environmentally it's absolutely a problem. [citation needed] Economically the Intel CPU of the 90s that had a very bad perf/watt compared to today's standards have been awesomely worth it Environmentally, CPU have been getting better and better, the difference of a few 10s of watts doesn't really make any difference, unless you have numbers to back up your very strong claim.

Citation needed? Do you live in some alternate reality? > "... the difference of a few 10s of watts doesn't really make any difference, unless you have numbers to back up your very strong claim." We have a billion power-hungry PCs running on the planet. Power-efficiency matters for economy and for environment, because power isn't free and only a tiny speck of the world runs on clean energy. It always mattered.

Most of those systems sit at idle, when they are on at all. Idle power is not the same as TDP. The idle difference is not so great that it's a big deal.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#87

Post by gumby »

KerrAvon wrote: post #31
gumby wrote: post #19

Poorly written, resulting in what could look like a bias. The article mentions that apple has focused on single core performance while the x86 processors in question are designed for multi core use cases. This reflects two different markets being addressed and the sad state (small amount) of multicore code today. Also it’s silly to claim that the M2 Ultra is so expensive — you can get the same performance from a 3K “…

haven’t read the article, but there are no single threaded use cases on Apple platforms that Apple cares about, so I’d take any claim that Apple is optimizing for single threaded performance with an enormous bag of salt

I'm pretty sure I remember discussion from Srouji and others on this specific topic at the "apple silicon" WWDC introduction. But I can't find anything useful from Apple in a web search, so it seems my statement is just a "some guy on the Internet" assertion.

It seems pretty clear and unsurprising that Apple optimises their design for their use case (e.g. major consideration of bandwidth to screen in handheld devices, reminiscent of one of the Alto's design criteria) but how that plays out doesn't support my claim either. But Apple's intended use cases aren't the same as the threadripper's.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#88

Post by brucethemoose2 »

adastra22 wrote: post #2

50% slower than the absolute best of class NVIDIA discrete GPU offering that consumes massively more wattage is hardly a damning statement.

Take anything from wccf with a grain of salt.

And yeah, as others point out, this is Apples to oranges. x86 desktops are great at some things, M2 Ultras are great at others, and the overlap that really matters is pretty small... Like, you have to be crazy to buy an M SoC for gaming, or buy a Nvidia GPU for workloads that won't fit in VRAM.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#89

Post by Retric »

wtallis wrote: post #84
StewardMcOy wrote: post #82

Earlier quoted context omitted.

It's not just that it doesn't come with a discrete GPU. Apple Silicon doesn't support discrete GPUs to begin with. As far as I understand it—and this is just from watching Apple's presentations on the architecture—the lack of a discrete GPU is a big part of how the Apple Silicon machines achieve good performance per watt. Instead of having discrete RAM or a discrete GPU with its own VRAM, all of the RAM is accessible…

> and also eliminates a lot of the need to copy data from RAM to VRAM, as both the GPU and the CPU can access the same memory. In return, this allows the GPU to match the performance of discrete GPUs that have a lot more cores. Plenty of PC hardware reviewers have done sensitivity analysis experiments to see how discrete GPU performance is affected by running with a slower or narrower PCIe link. The consensus is usua…

You can generally ignore FPS numbers as irrelevant outside of extremes, what people notice is stutter and that’s where PCIe link bandwidth matters.

Re: Apple M2 Ultra SoC isn’t faster than AMD and Intel last year desktop CPUs

#90

Post by jitl »

lxgr wrote: post #40
jitl wrote: post #26

Earlier quoted context omitted.

3rd party GPUs don’t mesh with the Metal strategy. Apple dropped and deprecated OpenGL a long time ago - and they’re the only one building metal GPUs

Doesn’t Metal support multiple GPU architectures just like OpenGL?

You’re right, there is some support - https://support.apple.com/en-us/HT202239
Post reply on HN