12 ms·
The GPU claims on the M1 Pro & Max were, let's say, cherry picked to put it nicely. The M1 Ultra claims already look suspicious since the GPU graph tops out at
by kllrnohj 5y ago
The GPU claims on the M1 Pro & Max were, let's say, cherry picked to put it nicely. The M1 Ultra claims already look suspicious since the GPU graph tops out at ~120W & the CPU graph tops out at ~60W yet the M1 Studio is rated for 370W continuous power draw.
Since you mention ML specifically, looking at some benchmarks out there (like https://tlkh.dev/benchmarking-the-apple-m1-max#heading-gpu https://tlkh.dev/benchmarking-the-apple-m1-max#heading-gpu & https://wandb.ai/tcapelle/apple_m1_pro/reports/Deep-Learning-on-the-M1-Pro-with-Apple-Silicon---VmlldzoxMjQ0NjY3 https://wandb.ai/tcapelle/apple_m1_pro/reports/Deep-Learning... ), even if the M1 Ultra is 2x the performance of the M1 Max (so perfect scaling), it would still be far behind the 3090. Like completely different ballpark behind. But of course there is that price & power gap, but the primary strength of the M1 GPUs seems to really be from the essentially very large VRAM amount. So if your working set doesn't fit in an RTX GPU of your desired budget, then the M1 is a good option. If, however, you're not VRAM limited, then Nvidia still offers far more performance.
Well, assuming you can actually buy any of these, anyway. The M1 Ultra might win "by default" by simply being purchasable at all unlike pretty much every other GPU :/
- brigade 5y ago100W of that is probably for the USB ports; afaik TB4 ports are required to support 15W, and I don’t think there’s been a Mac didn’t support full power simultaneously across all ports. (if that’s even allowed?)
- dingle_thunk 5y agoI suppose given that this is two M1 Max glued together, assuming cooling is a solved problem, the max SOC power consumption just twice as high as usual, plus interconnect overhead. Right? Based on the thermal and power consumption characteristics of previous chips I would not be surprised if say ~120W is the max power draw of this thing. edit: Of course the M1 max only shipped in laptops, so... who knows.
- masklinn 5y ago> assuming cooling is a solved problem I’d assume that’s what most of the chonk is about, no? > Based on the thermal and power consumption characteristics of previous chips I would not be surprised if say ~120W is the max power draw of this thing. The Max could be brought up to 90W or so.
- fnord123 5y agoI wouldn't assume heat is solved as it's been Apples weak point in the past. The cube would crack, g5 iMac's would melt capacitors, MacBooks would burn users' laps.
- scoopertrooper 5y agoTwo thirds of the volume seems to be dedicated to cooling. Assuming they’re not complete idiots, they must be doing something!
- fnord123 5y ago> Assuming they’re not complete idiots https://www.theverge.com/22967776/apple-magic-mouse-charging-port-bottom-upside-down-its-2022 https://www.theverge.com/22967776/apple-magic-mouse-charging...
- Bud 5y agoThose issues were all almost 20 years ago. If you think this is still a problem, you haven't used any recent Macs. The current MB Air and MB Pro both run very cool even under prolonged heavy loads. Apple's management of any and all heat issues has been far better than any competitors for a while now.
- fnord123 5y agoThe last intel macs run way too hot still. They get up over 80dC. The 2014 ones would hard reboot when building large Java projects.
- apohn 5y agoThe 3090 also can do fp16 and the M1 series only supports fp32, so the M1 series of chips basically needs more RAM for the same batch sizes. So it isn't an Oranges to Oranges comparison. Back when that M1 MAX vs 3090 blog post was released, I ran those same tests on the M1 Pro (16GB), Google Colab Pro, and free GPUs (RTX4000, RTX5000) on the Paperspace Pro plan. To make a long story short, I don't think buying any M1 chip make senses if your primary purpose is Deep Learning. If you are just learning or playing around with DL, Colab Pro and the M1 Max provide similar performance. But Colab Pro is ~$10/month, and upgrading any laptop to M1-Max is at least $600. The "free" RTX5000 on Paperspace Pro (~$8 month) is much faster (especially with fp16 and XLA) than M1 Max and Colab Pro, albeit the RTX5000 isn't always available. The free RTX4000 is also a faster than M1 Max, albeit you need to use smaller batch sizes due to 8GB of VRAM. If you assume that M1-Ultra doubles the performance of M1-Max in similar fashion to how the M1-Max seems to double the gpu performance of the M1-Pro, it still doesn't make sense from a cost perspective. If you are a serious DL practitioner, putting that money towards cloud resources or a 3090 makes a lot more sense than buying the M1-Ultra.
- mehmetoguzderin 5y ago> The 3090 also can do fp16 and the M1 series only supports fp32 Apple Silicon (including base M1) actually has great FP16 support at the hardware level, including conversions. So it is wrong to say it only supports FP32.
- oneplane 5y agoI'm not sure if he was talking about the ML engine, the ARM cores, the microcode, the library or the OS. But it does indeed have FP16 in the Arm cores.
- apohn 5y agoI should have been more clear. I didn't mean the hardware, but the speedup you get from using mixed precision in something like Tensorflow with an NVIDIA GPU.
- 5y ago
- joshspankit 5y agoWatching the keynote I was almost thinking that Nvidia missed the boat when they chose not to sign whatever they had to to make OSX drivers. Thank you for recalibrating me to actual reality and not Apple Reality (tm)
- zdw 5y agonVidia missed the boat in releasing a bunch of "replace the whole laptop logic board" chips that died in the 2008-2012 timeframe and annoyed a whole host of OEMs: https://www.techpowerup.com/64683/nvidia-admits-to-selling-faulty-mobile-gpus-shares-plummet https://www.techpowerup.com/64683/nvidia-admits-to-selling-f... Apple specifically: https://support.apple.com/en-us/HT203254 https://support.apple.com/en-us/HT203254
- wil421 5y agoMy 2011 MPB 15” suffered from the Nvidia GPU issue. It was so bad my computer wouldn’t boot properly and the Apple Geniuses kept denying the claim because it couldn’t finish a test. Anyway, it was still my longest lived Laptop. My Sony VAIOs were great but I liked that Mac better.
- ja27 5y agoBy 2011 it was AMD(ATI) not Nvidia. I had one that failed but Apple did the replacement for free.
- JiNCMG 5y agoI am certain AMD like most responsible vendors (Seagate) also worked with Apple to correct the issue. Nvidia's issue was that it told all of the laptop vendors to just deal with it. It's why they are hated by other vendors and AMD/Intel worked hard to keep them from creating x86 cpus.
- fouc 5y agoI pulled the logic board out and baked it in the oven and it lasted another 7 months before needing another bake. By then I moved onto a newer macbook.
- Macha 5y agoAn M1 Ultra is $2000 incrementally over a M1 Max, so there is no price gap, even with the inflated prices 3090s actually go for today.
- kllrnohj 5y agoTo be fair that $2000 also gets you +32GB RAM, +512GB storage, and +10 CPU cores. It's not just the GPU. Although yeah you can definitely fit a 3090-equipped PC into a $4k budget even with ebay pricing if pure GPU performance is all you really want.
- fivea 5y ago> To be fair that $2000 also gets you +32GB RAM, +512GB storage, and +10 CPU cores. Even though it's not an apples to apples comparison, keep in mind that a 1x32GB DIMM sells for less than 150$, and you can buy 1TB SSDs for less than 100$.
- tyrfing 5y agoI'm confident the marketing oversells it, but it's likely very good in comparison. 3090 is about 28B transistors on Samsung 8nm, with some budgeted for raytracing. This is on TSMC 5nm, a process with 3-4x the density, and the 114B transistor count could potentially allow for similar GPU size - although I'd wait for Locuza or someone to analyze it. It should be very competitive in performance, and the winner by far in perf/watt, at least until RDNA3 and Lovelace GPUs release towards the end of the year.
- lmilcin 5y ago> The M1 Ultra claims already look suspicious since the GPU graph tops out at ~120W & the CPU graph tops out at ~60W yet the M1 Studio is rated for 370W continuous power draw. And that is expected, a lot has to be reserved for USB devices.
- jsjohnst 5y agoSSD and those two large fans also use a fair amount of power too.
- kllrnohj 5y agoThose fans are unlikely to break 5w combined.
- jsjohnst 5y agoGuess we will see once one is available to be tested, but I doubt that’s the case. I was guessing 5-8W each, but I could be wrong of course.
- kllrnohj 5y agoMost PC fans are 1-2w. Given the claims of "near silent", I think it's plenty safe to say these are not 5,000+ RPM rippers that are going to be in the 5-8W each range.
- jsjohnst 5y agoGiven the M1 Ultra has a 2lb heavier heat sink just to dissipate the extra heat, not sure that’s a safe bet. There are other ways to reduce noise than just using lower RPM fans.
- lmilcin 5y agoMassive (by weight) heatsink can be used as thermal mass. This delays the need to increase fan speed and allows spreading that increase over time. One thing that is more noticeable than fan speed is rapid fan speed changes -- avoiding those makes entire system seem quieter.
- oneplane 5y agoYeah, there are a lot more parameters here than just 'there are two chips and one is the bestestst', like the availability you pointed out. There is raw performance, but there is also performance per watt, availability and scalability (which is both good and bad - M1 is available, but there is no M1 Ultra cloud available). If you want a multi-use setup, an RTX makes more sense than most other options, if you can get one and at a reasonable price. If you want a Mac, the M1U is going to give you the best GPU. In pretty much all other setups there are so many other variables it's hard to recommend anything.
- kllrnohj 5y agoFor the market this is aimed at, performance per watt is really irrelevant. Performance per dollar or just outright performance are far more important the vast majority of the time. That's how we ended up with 125w+ CPUs and 300w GPUs in the first place.
- oneplane 5y agoThere are dedicated ML cards from Nvidia for that, far most powerful than a 3090, so that is indeed true. But PPW is never irrelevant when someone is doing things at scale, so the question becomes: who is doing this for money but somehow not at scale?
- blitzar 5y agoThe cost of powering my 3090 for a year is now more than the cost (RRP) of a 3090.
- GekkePrutser 5y agoBut do you really run it all year round?
- kllrnohj 5y agoWhere do you live that power is anywhere close to that expensive? And are you overclocking your 3090? Even assuming literal 24hr/day usage at a higher, "factory overclocked" 450w sustained, at a fairly high $0.30/kWh that's $1200/yr. Less than half the retail price of a 3090. And you can easily drop the power limit slider on a 3090 to take it down to 300-350w, likely without significantly impacting your workload performance. Not to mention in most countries the power cost is much less than $0.30/kWh. At a more "realistic" 8 hours a day with local power pricing I'd have to run my 3090 for nearly 10 years just to reach the $2000 upgrade price an M1 Ultra costs over the base model M1 Max.
- hajile 5y agoA 16" macbook with an M1 only uses around 100w and that's when maxing all the things. It runs at about 40w for CPU and 60w for GPU. Based on those numbers, 120w seems totally expected for two chips at the same frequency.
- ChrisMarshallNY 5y ago> 370W continuous power draw. Don't know if it's the same, these days, but when I was designing electronic stuff, we were always told to spec the power supply at twice the maximum draw.
- Melatonic 5y agoI am sure it has great GPU performance for what it is but comparing it to a top Nvidia chip just seems ridiculous on Apples part. Apple I think is going to have trouble conquering back the semi pro workstation market they abandoned not that many years ago if they do not start offering M1 chips along with Nvidia GPUs. Once again we get a Mac Semi-Pro Mini (seems like the Studio is more like a replacement for the Trashcan) that their marketing implies is maybe as good as a Mac Pro but is obviously not. It does look a lot better this time around - at least it has more ports :-D
- scarface74 5y agoJust a note: they explicitly said the Mac Pro is coming.
- Melatonic 5y agoI'm interested to see if they shoot themselves in the foot and try to make it all M1 architecture or partner with AMD or Nvidia
- Tagbert 5y agoYou are comparing the power source (370W) to the CPU/SOC (120W). The power supply provides power for USB-C/Thunderbolt ports and it’s never a good idea to spec a power supply too low and run it too close to capacity.
- 64bittechie 5y ago
- Sosh101 5y ago> look suspicious since the GPU graph tops out at ~120W & the CPU graph tops out at ~60W yet the M1 Studio is rated for 370W continuous power draw. Interested to know what you think a reasonable PSU would be for A machine that was consuming close to 200W for processing...
- anaisbetts 5y agoAt some point can we finally admit that Apple's GPU claims just aren't....true? Like, every Apple keynote they put up incredible performance claims, and every time people actually get their hands on the product, it doesn't even come close to holding water in any domain where GPU performance matters (Game performance, ML training perf)
- jaimex2 5y agoI thought that would be a given by now. Won't even get into that it can't run most things you'd want that kind of hardware for.
- hirako2000 5y agoeven their graphs don't look legit. they look like what you see in boxes of vitamins found on ransom websites. Got to wait for independent benchmarks with well explained scenarios. anything can run twice faster than anything with some specific tweaks to the program that runs.
- threeseed 5y ago> GPU performance matters (Game performance, ML training perf) No one plays games on a Mac. And it has nothing to do with GPU performance but rather the fact that the audience simply isn't interested in gaming on it and so there is no economic incentive to target them. So the GPU performance that matters to Mac users and is relevant to Apple is not games but rather content creation, production etc.
- kllrnohj 5y agoM1's blender rendering performance also wasn't very good, though. Assuming the scene was small enough to fit in the smaller VRAM of competitive consumer GPUs anyway. Video content creation specifically is where it mostly achieves what the graphs indicate, and that's mostly a "well yeah, video decoder ASICs are really efficient. I'll take things we already knew for $100, Alex"
- tambourine_man 5y ago
- trodrigues 5y ago> Well, assuming you can actually buy any of these, anyway. The M1 Ultra might win "by default" by simply being purchasable at all unlike pretty much every other GPU :/ Can we stop it with the meme that these GPUs are unobtainable? Yes, they are still overpriced compared to their supposed original prices and they'll likely never return to that price given that the base prices of manufacturing, materials and such have increased for multiple reasons. But stock has been generally available for many months now and it's possible to get them as long as you can afford them.
- conk 5y agoWhere are these available? I just checked all the links on Nvidia's webpage for a 3060 and they are all out of stock...
- geraldwhen 5y agoGigabyte and strix 3090 are routinely in stock at Newegg. Msrp. The shortage is over
- kayoone 5y ago3090 is still 2400 EUR in Germany, pretty sure that's not MSRP