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

381–390 of 448 posts

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

#381

Post by Dylan16807 »

nateb2022 wrote: post #376

CPUs are intentionally engineered for a target market. When engineering the M2, Apple had a certain market in mind, and design and performance trade-offs were made accordingly. As others have noted, power efficiency seems to be a big priority. The M2 Ultra has a TDP of 60W. The i9-13900K has a base consumption of 125W, and draws up to 253W under stress. So do the math. Intel achieves 32% better single-core performanc…

> If Apple wanted to, they could probably conjoin two M2 Ultra's and soundly beat the i9-13900K by a considerable margin and still use about half the same power to do so.

You can play that in the other way too. Use multiple 13700Ts at 35W each.

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

#382

Post by Dylan16807 »

kmeisthax wrote: post #275
aunty_helen wrote: post #265

Earlier quoted context omitted.

The ultra isn’t in laptops. The ultra in the Mac Pro starts at 7 grand, if you’re going to pay a 4 grand premium, is 200$ of extra power over its lifetime really that important? I have a m2 Max, the battery life is awe inspiring. M2 ultra doesn’t best the competition for its use case.

Don't waste your money on the Pro, you can get an M2 Ultra in a Mac Studio for $4000. The only reason to buy the Pro is if you have a bunch of really weird PCIe cards[0]. As for the desktop use case... sure, you aren't going to care about the $200 of saved power draw, but not having a very loud and hot machine at your desk has to count for something, right? [0] GPUs won't work, their memory access is locked out in ha…

A proper cooler won't get loud or hot.

And as far as heating the room, okay that's an extra $50 of air conditioning to remove the $200 of heat.

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

#383

Post by yunohn »

nateb2022 wrote: post #376

CPUs are intentionally engineered for a target market. When engineering the M2, Apple had a certain market in mind, and design and performance trade-offs were made accordingly. As others have noted, power efficiency seems to be a big priority. The M2 Ultra has a TDP of 60W. The i9-13900K has a base consumption of 125W, and draws up to 253W under stress. So do the math. Intel achieves 32% better single-core performanc…

> The real question is why any consumer would need that much compute, and the target market for such a niche is probably very small which is why Apple didn't do that

This is a lazy argument. Why did they make the M1 faster than other CPUs? I would argue that most people don’t need even that performance. Why do they put M chips in iPads - they can’t even use it optimally.

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

#384

Post by pocketarc »

nateb2022 wrote: post #376

CPUs are intentionally engineered for a target market. When engineering the M2, Apple had a certain market in mind, and design and performance trade-offs were made accordingly. As others have noted, power efficiency seems to be a big priority. The M2 Ultra has a TDP of 60W. The i9-13900K has a base consumption of 125W, and draws up to 253W under stress. So do the math. Intel achieves 32% better single-core performanc…

The M2 Ultra comes on a $7000 workstation; it's not exactly consumer-level hardware. What you're saying is absolutely true - they could just double it up (like they do from the normal to the Pro, and from the Pro to the Ultra). And I hope they do, for an M3 Ultra. It'd be great if they pushed it even harder and showed Intel/AMD that they can beat them at both the low end and the high end, with plenty of power consumption to spare.

As you said though, it's a very niche thing, so even without beating Intel at the top end they're not going to be losing much. It's just a bit of a shame. The i9-13900K is only $500, it's not like it's some wildly out-of-reach thing.

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

#385

Post by wtallis »

sudosysgen wrote: post #379
wtallis wrote: post #372

Earlier quoted context omitted.

> While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually look better. I think this is an instance of the coordinated omission problem: you cannot simply measure the latency of the frames that were completed, but instead have to consider the frames that should have been delivered during a stall. Looking at frame time percenti…

This isn't actually true. Looking at aveeages will in fact reflect correctly on average frame staleness over a sampling period. A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. Think of it like Little's law in queing theory , it's better to simply just look at different percentiles, like 99 or 99.9th. The average framerat…

> A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum.

It's not clear to me: are you saying those two scenarios should be quantified as equally bad? Because it seems pretty obvious to me that the 5x outlier is qualitatively much worse.

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

#386

Post by sudosysgen »

wtallis wrote: post #385
sudosysgen wrote: post #379

Earlier quoted context omitted.

This isn't actually true. Looking at aveeages will in fact reflect correctly on average frame staleness over a sampling period. A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. Think of it like Little's law in queing theory , it's better to simply just look at different percentiles, like 99 or 99.9th. The average framerat…

> A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. It's not clear to me: are you saying those two scenarios should be quantified as equally bad? Because it seems pretty obvious to me that the 5x outlier is qualitatively much worse.

The fact that the 5x outlier may be qualitatively worse is something that you pin down using percentiles, which is what everyone does, either 1% or 0.1% low. The argument about time-weighted frame staleness is not mathematically valid.

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

#387

Post by gautamcgoel »

cjbprime wrote: post #112

Reasons to still prefer an M2 for a HEDT situation, despite the article claims: * The latest Intel/AMD desktops only have two channels of DDR5, and if you put as little as 64GB RAM in them they can drop to DDR4 speeds (2DPC, dual-rank), less than 10% of M2 memory bandwidth. * You can (surprisingly cheaply) buy an M2 with 192GB GPU VRAM, but you can't buy a DDR5 PC with the same amount. If we all start wanting to run…

Why is it that using 64GB of RAM causes bandwidth to drop?

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

#388

Post by Retric »

sudosysgen wrote: post #378
Retric wrote: post #365

Earlier quoted context omitted.

Yes occasional 90ms stall reduces 1% frame rates, but less than how annoying it is. A 10 minute test at 120 FPS is 10 * 60 * 120 ~= 72,000 frames. Your 1% lows is an average of ~720 frames but so what if of 300 them are at 60 FPS you’re going to have trouble noticing. While if you’re almost rock solid at 120FPS a dozen 70ms stalls it’s really obvious that something is wrong but the calculated 1% low’s may actually lo…

This is a completely misguided analysis. First of all, there is nothing in the M series architecture that's going to avoid pipeline stalls, let aloe huge 70ms graphics pipeline stalls. As I've explained above, games keep multiple frame pipelines in flight at once, so you don't have these kind of issues unless there is a scheduling or I/O issue. There is no advantage here. The only advantage is that the transfer from…

> 100ms+ frames are not a problem on modern games and modern hardware, and when they are

There’s a multitude of such bugs which ship with modern titles. Benchmarking sites don’t use the release day build of 2077 for very good reasons.

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

#389

Post by Retric »

sudosysgen wrote: post #386
wtallis wrote: post #385

Earlier quoted context omitted.

> A single frame that's 5x too long or 4 frames that are 2x too long have the same impact on the average delta of frame staleness on the total sum. It's not clear to me: are you saying those two scenarios should be quantified as equally bad? Because it seems pretty obvious to me that the 5x outlier is qualitatively much worse.

The fact that the 5x outlier may be qualitatively worse is something that you pin down using percentiles, which is what everyone does, either 1% or 0.1% low. The argument about time-weighted frame staleness is not mathematically valid.

There’s nothing inherently mathematically special about 1% or 0.1%, you can also use 0.0042% or 0.00069% which highlights this kind of extreme outliers more.

Benchmarking sites use 1% or 0.1% because they better highlight the commodity PC hardware rather than the games bugs and architecture.

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

#390

Post by THENATHE »

menus wrote: post #269
IneffablePigeon wrote: post #107

Earlier quoted context omitted.

I bought an m series device to replace my windows desktop mostly for efficiency reasons. It saves me hundreds a year in electricity costs at the moment (quite literally, I calculated it beforehand and checked real usage afterwards). It also means my office is now silent and cool. It’s a big QoL improvement for me.

Was cost of the device also taken into consideration in your overall cost calculation? What was the cost of your Windows desktop and what was the cost of your M series device you replaced it with? Have you ever upgraded any components of your Windows desktop (RAM, GPU, CPU, motherboard) or did you discard the entire thing?

Realistically, assume you were going to upgrade anyway.

$1200 PC vs $1500-$2000 MacBook is really only 300-800 difference, and MacBooks hold their value on the second hand market EXPONENTIALLY better than windows.

So 3-8 years at 100$ a year, but also it’s worth 3x as much as the comparable PC in 5 years

Post reply on HN