CPU ST Performance: Not Much Change from M1

Apple didn’t talk much about core performance of the new M1 Pro and Max, and this is likely because it hasn’t really changed all that much compared to the M1. We’re still seeing the same Firestrom performance cores, and they’re still clocked at 3.23GHz. The new chip has more caches, and more DRAM bandwidth, but under ST scenarios we’re not expecting large differences.

When we first tested the M1 last year, we had compiled SPEC under Apple’s Xcode compiler, and we lacked a Fortran compiler. We’ve moved onto a vanilla LLVM11 toolchain and making use of GFortran (GCC11) for the numbers published here, allowing us more apple-to-apples comparisons. The figures don’t change much for the C/C++ workloads, but we get a more complete set of figures for the suite due to the Fortran workloads. We keep flags very simple at just “-Ofast” and nothing else.

SPECint2017 Rate-1 Estimated Scores

In SPECint2017, the differences to the M1 are small. 523.xalancbmk is showcasing a large performance improvement, however I don’t think this is due to changes on the chip, but rather a change in Apple’s memory allocator in macOS 12. Unfortunately, we no longer have an M1 device available to us, so these are still older figures from earlier in the year on macOS 11.

Against the competition, the M1 Max either has a significant performance lead, or is able to at least reach parity with the best AMD and Intel have to offer. The chip however doesn’t change the landscape all too much.

SPECfp2017 Rate-1 Estimated Scores

SPECfp2017 also doesn’t change dramatically, 549.fotonik3d does score quite a bit better than the M1, which could be tied to the more available DRAM bandwidth as this workloads puts extreme stress on the memory subsystem, but otherwise the scores change quite little compared to the M1, which is still on average quite ahead of the laptop competition.

SPEC2017 Rate-1 Estimated Total

The M1 Max lands as the top performing laptop chip in SPECint2017, just shy of being the best CPU overall which still goes to the 5950X, but is able to take and maintain the crown from the M1 in the FP suite.

Overall, the new M1 Max doesn’t deliver any large surprises on single-threaded performance metrics, which is also something we didn’t expect the chip to achieve.

Power Behaviour: No Real TDP, but Wide Range CPU MT Performance: A Real Monster
Comments Locked

493 Comments

View All Comments

  • _crazy_crazy_ - Tuesday, October 26, 2021 - link

    yes the simple way to explain is packaging .. while intel and amd are working on cpu/apu that need to comunicate with external ram (not in the cpu die) and need to account a multitude of other peripherals and components apple just stuck everything in the cpu core thats where most of the optimisations come from in a simple way , its another way to do things ..
    but then comes the reverse that we have yet to know and thats durability because in normal hardware if a ram fails you can simply replace it in apple case... well its all e-waste
    its a way of doing things that works with apple since its pretty much a closed ecosystem
  • coolfactor - Tuesday, October 26, 2021 - link

    Good answer, but it's a SoC (System on a Chip), not just a "CPU Core" as you described it. This is an important distinction.

    One reason that computer hardware fails is due to heat, so if the system is way more efficient than traditional discrete components, it won't get as hot and therefore should not fail as much. A more reliable architecture in many ways.
  • _crazy_crazy_ - Tuesday, October 26, 2021 - link

    thanks for the correction i blame lack of my morning coffee xD

    well heat is one reason for hardware failure but there are a lot more and assuming a normal failure rate in consumer electronics of around 10% in the first year of use it's still a lot since you cannot replace anything from these machines eveything is soldered and not that easy to salvage
  • web2dot0 - Tuesday, October 26, 2021 - link

    When was the last time a consumer PC failed? Just more baseless fear mongering.
  • vladx - Tuesday, October 26, 2021 - link

    Indeed and at least PCs can be fixed without paying a fortune each time.
  • ComputeGuru - Tuesday, October 26, 2021 - link

    Very good point vladx. Hopefully Apple offers an excellent replacement warranty for those rare times when something on the SoC package will fail and render the whole system dead.
  • zodiacfml - Tuesday, October 26, 2021 - link

    TSMC 5nm process and really large chips. Apple has the luxury to make large chips because they have plenty loyal customers to buy their pricey products. in fairness to Apple, their M1 desktop computing products are better value than their iphones.
  • kwohlt - Tuesday, October 26, 2021 - link

    I don't know how Intel / AMD can actually respond. If they take the very-large-physical-chip approach, this will certainly help performance for their higher SKUs, but would absolutely kill the costs of their low end chips that would need to share a socket. As long as Intel / AMD offer a lot of SKUs that can be mixed and matched with other components, they will always be at an efficiency disadvantage.

    As long as that gap isn't too large, I guess it'll just have to be something people will accept.
  • vladx - Tuesday, October 26, 2021 - link

    They don't need to respond, software support is 100x more important than hardware performance.
  • OreoCookie - Tuesday, October 26, 2021 - link

    I think Intel makes more than enough profit per chip to eat up the larger cost of larger dies. Intel's problem is that they cannot produce chips with as many transistors with competitive efficiency as they have been stuck on a worse process node. To stay competitive in terms of performance, they have been pushing clockspeeds to the limit, which kills battery life.

    AMD can respond more easily, because they can stitch together suitable chiplets. In the future they could also add RAM in the same fashion that Apple does or they already do for their GPUs (HBM). So they have the expertise and the technology to do that.

    Lastly, one small comment: there are over 1.5 billion active iOS devices out there. The image that Apple customers are comprised of “loyal customers”/“fanbois” is outdated, the user base is way too large for that.

Log in

Don't have an account? Sign up now