Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel Alder Lake vs. Apple M1
- Thread starter leman
- Start date
- Sort by reaction score
I’ve seen it go both ways on single threaded, which is a bit odd because in theory that shouldn’t be the case. It should only be the same or better (higher memory throw put and larger caches - the only thing I can think of is slightly longer latency on some memory ops).
However I agree that single threaded power of the base M1 makes it the far more economical choice for non-Pro users. That’s why I’m saying Apple forced the Pro/Max to go big to differentiate itself in throughput (both CPU and GPU) rather than latency. As an aside this is also why I’m not a fan of a lot of the threads on here bemoaning or celebrating the demise of the previous MBPs. It’s less to do with Ives or the lack of him and more to do with Apple Silicon seismically shifting Apple’s product stack around. I suspect with the M2 and revamped Airs and maybe baseline MBPs we’ll see more of this.
Is not M1 Max using up to 95W? Are you implying that Alder Lake uses 950W. Where did you get this bogus number?
95W is for the CPU and GPU combined. The CPU alone is about 40W (can be pushed to 43W). For single core, the M1 is about 5W. Also he’s exaggerating (slightly
) for effect.I haven't done the math, but I believe he is talking about M-series CPU core performance against Alder Lake CPU performance. 95W is the full SoC, if I'm not mistaken.
So do we think it make sense to compare Intel halo water cooled desktop product vs apple laptop chip ? Do we think apple will have a halo desktop part as well ?
But that's not the case. The Intel CPU only runs at these very high power levels above 200W when it's fully loaded with sustained CPU-heavy workloads. But for those workloads it is significantly faster than the M1 Max. For example, in Cinebench R23 the i9-12900K gets ~2,000 single/27,000 multi, whereas the M1 Max gets ~1,500 single/12,000 multi.
But that desktop CPU is a high-end enthusiast model. When you go to the 12700 or 12600 you get significantly lower power consumption and they are still fast.
For sure. But it's not miraculous.
I think it makes more sense to compare the M1 Pro and M1 Mac to the inevitable 12800H, 12850H, 12900H, 1295H and 12980HK CPUs whenever Intel gets them out the door and whatever discrete GPU is attached.
Yeah, and I have a suspicion that a couple of the trolls have created new screen names.
They aren't different. If M1 can do the same job as an i9 with half the power, you can fit two of them in a system designed to support i9 power consumption.
Cores throttle when they overheat, portable packages get hot to the touch, breakers trip, cooling is the largest cost in a data center or super computer. At every level, you can get more performance total if you can get more performance per watt. Efficiency is performance.
That's an oversimplification at best. M1 Max is already at the limit of one chip die size. You can’t double it.
We'll see what Apple does with the iMac (Pro?) and Mac Pro that get the workstation class Apple Silicon SoCs.
That's an oversimplification at best. You can have more than one chip in a machine.
Do you think the MacPro is going to be 8 performance and 2 efficiency cores?
Why not compare to the i5 and i7 12th gen cpus? The i5 pulls 125watts and still beats the m1 in single core. There will be 700 dollar laptops out soon that will go head to head with 3k Mac books
I don’t fully disagree but this is not a good example as Cinebench doesn’t seem to use cores fully on macOS/M1 - unclear which interaction is the problem as some people said it still doesn’t do it on Intel Macs either but I haven’t verified - too old a Mac anyway. Windows/x86 doesn’t seem to have this problem so the perf/W gap weirdly actually widens further than it should in favor the M1 Max even though absolute performance is lower than it should be. At any rate other multicore benchmarks have them far closer in performance with power usage in the Intel chip still being significantly higher. It depends on the workload. (Floating point being a particular strength of Apple’s design apparently)
Anandtech estimated that the M1 Max was 6-7x more power efficient than Tiger Lake in multicore. That’s too much ground to make up, even for ADL and even if you put it on the same TSMC node as the M1 Max. Yeah the M1 Max won’t be 10x more power efficient than Alder Lake mobile, okay. But it is still going to be significant.
Last edited:
If that's the case the reported power consumption numbers for the M1 when running Cinebench are probably also not correct.
Well, according to the tweat posted earlier in the the thread the i9-12900K still gets a ~25,000 MT score in Cinebench when the PL2 power is capped at ~150W ...
We'll see soon. According to Gelsinger the laptop parts will be revealed later this quarter.
There were some oddities with the software reporting but it’s backed up by wall power measurements too. The M1 Max CPU pulls 50-60W at wall going full throttle. On Cinebench, it’s pulling less than 40W at the wall. TGL meanwhile pulled 106W at the wall near its top for CPU only.
Actually it is pretty miraculous, we finally have real desktop level performance in a laptop with near-to-complete silent operation and battery life that actually lasts (way) more than 90 minutes under stress. When has this ever happened before in any laptop? Hint: It's never.
Lol, no.
Yeah I also have no idea why people keep assuming that Apple has no strategy for 'real' desktop, it's really bizarre. Everyone did it with the first M1's, then Apple released M1 Max which is 4x M1, and there have already been several rumours that M1 equipped Mac Pro's will be 40 core CPU + 128 core GPU which is 4x current M1 Max. So its a pretty safe assumption iMac Pro's should then slot roughly between the two and probably offer 2x M1 Max with 16-20 core CPU + 48-64 core GPU, which will obviously obliterate Intel.
You’re not going to see an Intel H-Series in a $700 Windows laptop. Right now, you’re also comparing a desktop CPU that wastes 125w of electricity to best a 20w CPU in single core benchmarks. Is that something Intel should be proud of? They need 6x the power to beat Apple?
What do you mean the 12th gen i5 is like 200 dollar cpu and the m1 chip does not pull 20 watts. The i5 pulls 125 with its gpu also running. Look it up it breaks 1800 single core in geek bench 5.
That's not comparing intels alder lake Mobil cpus. The desktop i5 maxes at 125 watt full load. The mobile will be alot better
The m1 which is what a $700 laptop would actually be competing against (and such a laptop would rarely if ever use a $200 cpu) is indeed a 20W cpu - 30 including the GPU. ?♂️