Intel has improved much faster these last few years. The i7 has improved from 10th to 12th generation as much as from 4th to 10th generation.
Intel Core i7-12700K - ST 4057
Intel Core i7-10700K - ST 3072
Intel Core i7-4771 - ST 2193
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.
M2 is really just m1.1 (Update, I was wrong).
- Thread starter macguy360
- Start date
- Sort by reaction score
Intel Core i7-12700K - ST 4057
Intel Core i7-10700K - ST 3072
Intel Core i7-4771 - ST 2193
By the time you wiped the M5 off, I would be waving goodbye in my M7 Ultra MAX + Chip
Actually, because the performance increases compound, I think if the total improvement over 7 years was 37%, the average annual improvement was 4.6%. It’s geometric, not arithmetic. (1.046^7=1.37)
Also, it’s worth noting that there have been 2 years between M1 and M2, so the annual improvement by Apple is sqrt(1.1)=4.8%. Comparing generations doesn’t make sense, we need to compare over time.
And I don’t subscribe to the “it’s all about short term supply chain issues” theory. Intel has had short term issues over the past 7 years also.
All in all, I think the M1->M2 was a respectable update, and people thinking it was going to be more ground breaking than it was were probably underestimating how well optimized the existing architecture is. Apple didn’t trip, they kept pace, that’s what I expect. Over the long time horizon, I expect the gap between AS and Intel to slowly widen to Apple’s benefit, but not by 20% a year.
Besides, the real story of AS isn’t the cores at all, it’s the overall heterogenous architecture and SoC construction.
I've seen a lot of pointless petty arguments on the internet but this one is pretty goofy
Yes, and the perf per watt, in which the gap is already significant. They seem like details but current Intel offerings can either perform competitively or have competitive battery life, not both. As long as those desktop replacement laptops with discreet graphics are turned off the performance halves so they can achieve a somewhat similar battery life.
19 months
Will be interesting to see what Apples update cadence becomes when they don’t need to worry about supply chain delays
I was using Apple's self-reported 18% single-core improvement over 18 months to get the ~10% number, but with the official Ars Technica review out with actual GeekBench 5 scores (what I was comparing the iMacs with) it looks like overall single-core gains on that specific metric are indeed ~10% from M1 to M2.
Yep, they definitely played some long-overdue catch-up with Alder Lake, but given how much they increased TDP to achieve that I highly doubt that rate of improvement is sustainable. Intel's track record over the past 20 years is to release major architectural improvements with huge performance gains (Core 2 Duo, Core i3/i5/i7) and then make incremental improvements over the better part of the decade until the next architecture shift. Unless they've got some serious tricks up their sleeve, I don't expect the next few generations of Intel CPUs to break from that pattern.
I’m convinced that their TDP was directly related to their desire to just put up better total performance numbers than Apple’s solution at all costs. One unforeseen outcome from Apple Silicon could be Intel has TDP spikes at every generation.
I’m wondering when Apple is going to start to become willing to let their chips get hot. The fact that the Mac Studio Ultra basically just sits there at room temp suggests they have some headroom. I’ve heard indications that the AS parts aren’t designed to just clock up at higher voltages, I suspect they’re optimized for the clock they run at. So if they‘re willing to run hotter, it would probably mean not scaling the exact same design from MBA to MacStudio and beyond.I’m convinced that their TDP was directly related to their desire to just put up better total performance numbers than Apple’s solution at all costs. One unforeseen outcome from Apple Silicon could be Intel has TDP spikes at every generation.
IMHO, I would think Apple would also want to have software consistency that scales from their entry level all the way to their top of the line Macs. So I would think its quite unlikely that the high end Macs would have vastly better single core performance compared to their entry level Macs. The differentiating factor across the scale would just be more cores.
I guess I’m not sure what you mean…. Why would there be a problem with their high end models being faster than their entry models? Isn’t it more strange that they’re not?
Never. Intel/AMD’s have to get hot because they’re in a back and forth running the same code and one has to provide a faster result than the other. Apple’s completely out of that arena now, they only have to be more performant than the last released Mac of a given type. I’d imagine at some point, Intel and AMD processors could reach double digit gigahertz and pulling insane amounts of power. While Apple will chart out steady improvements year over year.
My personal thought has always been that iPhone chips COULD be clocked higher as I’m sure they bin chips like everyone. The big difference is that, since they don’t have to support a wide range of cheap solutions that they make for others, they just toss everything below, say 2 GHz. Some undoubtedly could run higher, but as Apple are writing the software the chips will eventually run (and writing the software that will create the software that others write), they know that 2GHz is not only a good speed for the particular chip, it’s also a speed that most of them will be able to easily hit, yielding more good chips per run. I can’t envision them ever being in a situation where an Intel processor doing something faster is a thing they ever need to be concerned about again.
That’s just because of what chip vendors that make general purpose solutions have trained folks to think. When you really think about it, if all the chips are built from the same basic foundations, shouldn’t they all offer similar single threaded performance? The cheapest Apple Silicon offers amazing single threaded performance because it’s a derivative of the top end with fewer corses. The cheapest Intel is intentionally created to not encroach on the single threaded performance of their high end.
It's Apple's SOC... they can call it whatever they want... since the beginning of the personal computer, processors have made incremental advances, so nothing to get worked up about... I have an M1 Max in my Mac Studio and an M1 in my MBA... both perform extremely well for my use, so I will "limp" along until I need a new computer and then buy what ever Apple makes available, regardless of what they call it.
That's cause they got complacent and when AMD and Apple came strong into the market they reacted; late reaction.
Good point, so the M series improved 11% per year: 1.11^(19/12) = 1.18
That’s pretty significant compared to the 4.6% Intel figure mentioned above.
that's not how it works.
Core Size are 100% identical and performance increased by 20% (18, but whatever). This comes because it's based of A15 and not A14. Architecture optimizations and Clockspeed, also bigger cache?!
The WHOLE SOC-Die increased in size. Because now there's a dedicated Media Engine, bigger Neural Engine, and maybe some IO changes for stuff like 24gb Ram.
All of those things do Nothing for the CPU performance.
Look at the M1 Pro vs. M1 Max. Huge Die Size difference and transistor count. But CPU Performance is absolutely completely identical, because the CPU Cores are identical.
Why would I want my desktop chip with room for much more thermal management to have the same single threaded performance as my phone? I understand how that’s easier for Apple, because they have less designs to manage, but how is it better for me? Perhaps Intel cripples their low end part to protect their high end, but Apple is probably limiting desktop performance by limiting to what can be done in a phone.
I say probably, because I obviously don’t know what the design limitations are, but it stands to reason that if relax some constraints (thermal limits) you open up some design possibilities.
Apple still competes with other PC vendors. Obviously some people will make some compromises to run MacOS and keep their familiar workflows, but in the end people have jobs to do and if Intel can do those jobs significantly faster then they will leave Apple— just as people are leaving Intel now to move to the M1 series laptops.
Not going to give that video a view, but based on the click-me thumbnail it looks like the M2 is running at 104C. I seem to remember that M1 ran at 100C. Are we supposed to get up in arms that M2 is running 1% hotter?
Remember, physics depends on Kelvin, not Celsius.
Of course anytime someone makes an ill advised snarky complaint about price they always seem to neglect that comparable windows machines seem to cost more, why is that? Maybe SSD, CPU, display quality, build quality don’t matter? I mean this is way more money than a cheap Chromebook or a cheap low power windows craptop