Top intel mobile processor appears to be https://browser.geekbench.com/v5/cpu/10431820
M1 Max gives 9% boost for single-core and 34% for multicore, with similar or larger (?) TDP -- Intel is 35-45 W, M1 Max is 60W but I assume some of it (a lot?) goes to GPU. Impressive, but probably wouldn't be called revolution if came from Intel.
M1 only uses around 15-18w with all 8 cores active according to Anandtech's review (with E-cores using 1/10 the power, that's equivalent to just a little less than 4.5 big cores. I'd guess 30w for all cores would be the upper-end power limit for the cores.
Intel's "TDP" is a suggestion rather than reality. Their chips often hit 60-80 watts of peak power before hitting thermals and being forced to dial back.
That figure is after subtracting idle power. It’s right there on the chart.
The workload is described as "Compute MT" (i.e. multithreaded), compared with "Memory MT (i.e., memory intensive Multi-threaded workload), with a third bar on the chart being the average of the two.
It seems to me that the "Compute MT" is the closest to the measurement that Intel or AMD would describe as TDP. The M1 is a ~25+W chip, not a 10W chip, if we’re going to make a fair comparison with Intel or AMD equivalents. It’s still a great CPU of course, but it’s not magic pixie dust.
Compute MT is using both CPU and GPU together. To compete, any other CPU needs to use a dedicated GPU.
Also, at idle, the DRAM and SSD are powered off. Samsung 980 uses 4.5 watts of power when active. I'd guess a similar power consumption applies here.
LPDDR4x ranges from 0.6v to 1.8v This should give 2-3w per 8gb chip at full-power. Anandtech's article agrees with ST Memory adding 4.2w of power over their other ST benchmarks.
Now, 26.8 - 4.5 - 4.2 = 18.1w at PEAK.
For just CPU loads, we get 22.3 - 4.5 - 4.2 = 13.6w.
I don’t think your logic is based on sound assumptions:
You can’t completely turn off DRAM at idle if it’s to store data - it has to self-refresh continuously. Some power savings are available by running it in a power saving mode, but you can’t get around the leakage of the DRAM cells, so the extra power used to drive the DRAM hard is not the full DRAM power usage, but a fraction of it. (LPDDR5 is nifty in that you can tell it to power down parts of the chip that aren’t currently being used.)
The SSD isn’t being hit by either workload: why would a CPU benchmark hit disk? The SSD is going to remain idle & be part of the 4W that Anandtech assigns to the system idle power drain.
Where does it say that Compute is CPU + GPU? (I’m happy to believe it, but I don’t see it in the article & they already break out the GPU power usage seperately using a different benchmark.)
(It would really help if Anandtech told us what the benchmarks were - some CPU intensive benchmarks don’t hit memory at all, they just hammer a few cache lines. It’s entirely possible for a pure computational benchmark to leave most of the LPDDR5 in deep sleep!)
If by MT Compute they mean specifically the Geekbench 5 compute benchmark then that is a Metal / GPU benchmark, so I will concede this point in that case :)
Even conceding that the Compute benchmark is really a GPU benchmark, it’s still the case comparing a CPU benchmark load against the TDP of another processor isn’t entirely correct, as the TDP is meant to be a power ceiling (although Intel has rather breached this in recent years - AMD TDPs are more honest I’m told.) - if we accept the 22.3W of the average workload benchmark, then the TDP for the chip is still going to be more than that & that’s the appropriate point of comparison.
LPDDR4x has various power levels based on whether it's actually being used or just keeping data from corrupting. The latter is much less memory intensive and runs at much lower voltages. You may have a point about SSD depending on how much read/write is happening.
Check out this comparison of the passively-cooled M1 Macbook Air vs both Intel and AMD Surface 4 machines.
TL;DR -- AMD loses 25-50% of their performance as soon as the plug is removed. Their actual power usage spikes at 40w (for a 10-25w U-series CPU). No matter how you cut it, Peak M1 power usage is about half for similar performance.
Insult to injury, the M1 not only has far better performance, but it's also cheaper.
But I think it’s maybe twice (three times tops?) as good as the competition on a power/perf basis, not the 10x some people keep claiming when they compare the 10W claimed by Apple against the 100+W TDP of similarly performing AMD/Intel CPUs.
Incidentally, assuming the benchmark they’re running is Geekbench (seems likely at this point), I’ve just run a pass of both CPU & Compute GeekBench5 benchmarks with iotop running in the background & neither did any disk IO of any significance during the run. I think the peak throughput I saw was 150kb/s and that might have been something else on the desktop (I didn’t bother killing anything). Most of the time there was no IO at all.
So, if it was GeekBench they were running (seems plausible) then I think we can assume that the SSD is idle & that the Compute benchmark is really a GPU benchmark.
At first glace yeah, but that's 45W at 2.5Ghz, but TDP isn't the power rating of the CPU. That benchmark lists the Intel CPU as up to 4.9Ghz, I would say it's actual power draw was closer to the 100W mark for CPU only.
4.9 GHz should be single-core max freq, not applicable in multicore benchmark. All-core turbo is lower, but also AFAIK there is a limit on time it run on this freq (tau for PL2, i think around half a minute by default).
Is there any reason why the Intel CPU wouldn’t run at 4.9 GHz for the single core benchmark though - whilst the M1 would be limited to a much lower frequency which it can sustain for much longer?
I think, but I'm not 100% sure, that limits are on overall TDP, i.e., single-core workloaad can run on turbo indefinitely. Then aalso I'd assume benchmarks are much longer than any reasonable value of tau which means it integrates out (i.e., Intel performance may be higher in short-term).
Edit: that is probably one of the reasons single-core gap is smaller
I wonder how much a performance hit not having "hyperthreading" would be for the AMD cpu. Back when the openBSD guys disabled it I did the same, and I havent really noticed much difference in the workloads I have.
Considering the m1 lacks it it seems it is not necessary for good multithreading performance, and the implications of it has always been a can of worms.
Yes, and the M1 is consuming less power at low loads, enabling longer battery lives. What other explanation do you think there could be for the long battery lives of the M1 laptops?
Depends on how you look at it. What's impressive about the M1 is that it gets highly respectable Geekbench scores while also being able to deliver 20 hours of battery life in a small form factor laptop. If there is an Intel chip that can do that, where are the laptops with comparable battery life?
We can nerd out about the exact definition of TDP all day, but ultimately these benchmarks are only useful insofar as they tell us something about real world performance and power consumption.
Well, I hear a lot of claims that M1 is significantly outperforming Intel (either by raw computing power or per watt). All I saying that both are at level of maybe single generation improvement, nice but not revolutionary.
Of course they also those low-power cores that allow long battery life, which is very nice, but thats a separate story
What do you mean "claims"? M1 has been out for a year. There are countless real-world demonstrations. Apple left Intel in the dust. Intel doesn't know how to make a cool mobile chip.
I mean that top Apple chip outperforms top Intel chip by 9% percent single core performance and 34% multicore on similar TDP. M1 is very energy efficient, and MBA is very cool cheap typewriter, but Apple hasn't outperformed Intel on high load.
So yeah, basically Apple has chip that slightly outperform Intel at high load and also can be very low power. Performance-wise (that what the parent message was about) the latter doesn't really matter.
>So yeah, basically Apple has chip that slightly outperform Intel at high load and also can be very low power.
This is what people are impressed by.
Given your 'typewriter' comment, you're clearly not interested in a lightweight laptop with decent performance and an extremely long battery life. But a lot of people are very interested in exactly this combination of features. And to be serious for a moment, the M1 Macbook Air has more than enough performance for the majority of people.
This is really an example of how Apple 'gets it'. Even if you're right that Intel is only slightly behind technically, they're way behind in terms of the overall product experience. You just can't make an M1 Macbook Air clone with any current Intel chip.
My OP talked exactly about claims (and there are many at least around me) of outperforming in terms of raw power.
I never argued with the fact that ARM allows to stick a couple of very low-power cores, while x86 doesn't.
In terms of "lightweight laptop with decent performance" MBP (the discussion is about new stuff after all) is less interesting, because it's heavier, more expensive, and probably doesn't provide much added value. M1 was revolution (mostly because someone took ARM and put in popular laptop and took the burden of software transition), those are much less so.
I can see myself getting MBA if I'll need something small and cheap and there won't be decent competition (maybe Google's Chromebooks? Or something from Samsung?), but not MBP (mostly because I dislike Apple's UX and I think competition is decent). I'd consider if it was say outperforming Intel by 50% in single core tasks.
Now that you mention, I remember my grandfather used to sit us around the fire and regale us with stories of how he used to edit 60fps 8k video on his old typewriter. I pulled it out of the attic just now and ran geekbench, and you're right, my new MBP is basically the same.
Thanks bud for bringing so much insight to this conversation!
Well the geekbench score didn't say anything about it being a revolution, that was your take. If Intel was capable of keeping their chips cool, while maintaining high perf, they would have been called a revolution too.
If you look up other laptops with the same processor you get a very large variation. The Lenovo you point to has a 280 watt power brick, maybe the results aren’t as good when it’s not plugged in.
