Question Incredible Apple M4 benchmarks...

Page 8 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Mopetar

Diamond Member
Jan 31, 2011
7,961
6,312
136
You could use SPEC to measure compilers as well by running the output binaries each produce on the same CPU to see which does the best job.

AT does a good job because they keep everything consistent to the highest extent possible to try to compare various CPUs to one and other. However, it's not the only way to test.
 

mr_roboto

Junior Member
May 15, 2024
2
13
36
They are preventing everyone from running Windows on their hardware. Virtualization is not the answer. All they need to do is support Windows on ARM in bootcamp mode with decent drivers and this bit of criticism goes poof!
I know this is an old post, but you've kept returning to this idea, and it's just nonsense.

Some core low level peripherals such as interrupt controllers need to have their drivers baked into the kernel - they can't be loadable modules. On Arm platforms, Windows assumes these are the de facto standard options common to most open-market Arm SoCs. Additionally, WoA is designed to be booted by UEFI firmware.

This hardware and firmware doesn't exist on Apple Silicon. Some of that's because Apple began designing Arm SoCs before any de facto or de jure Arm platform standards got rolling. Some of it's about Apple's requirements - UEFI isn't secure enough for their tastes, so they wanted to get away from it when they left x86 behind.

Apple can't change the Windows kernel or boot infrastructure. So, even if Microsoft wants to make Apple write most of the drivers (fair enough), there's no point in Apple doing so until Microsoft has poured the foundation. Apple has indicated they're open to working with Microsoft to get this done, Microsoft doesn't seem interested.

Do you forget that it was an Apple engineer who in his spare time, got MacOS working on x86 hardware, Steve Jobs saw it, flew to Japan to meet with Sony top dog and pitched an idea of MacOS running on Sony VAIO laptops?
The story you're referencing omits so much background that it can be misleading.

By the time Apple was reverse-acquired by NeXT, their OS (which was going to become Mac OS X) had already been ported to five different hardware platforms in total. Notably, one of these was the x86 PC.

Some "Rhapsody" (Apple's codename for OS X) developer prereleases still supported x86 PCs, and Jobs occasionally hinted in public that he'd gleefully abandon PowerPC if he didn't get what he wanted out of IBM and Motorola. By the time they released OS X in 2001, that kind of rhetoric had died down and they'd removed x86 support.

However, Apple also open sourced part of OS X as "Darwin". Darwin included most of the OS below the graphical user interface layer. They periodically shipped Darwin binary distributions corresponding to Mac OS X releases, and these distros continued to be dual-architecture PPC + x86.

That's how a single engineer was able to port OS X to x86: much of the work was already done and actively maintained by others. The rest was mostly already done, but not maintained.

If Apple isn't promoting the fact that their bootloader is unlocked, there's no guarantee that it will stay unlocked. Maybe they are curious to see how far hackers get with running a functional Linux on their hardware. If they get too close for comfort, there is nothing stopping Apple from locking everything down coz we all know how much Apple fears competition and they do literally everything under the sun to prevent anyone from getting in on their side of the fence.
You often let the mask slip and make it clear you're trolling. "we all know how much Apple fears competition"? Give me a break.

No, Apple doesn't run marketing campaigns to inform the general public that their bootloader is unlocked. Why would they?

That doesn't mean they haven't discussed it at all. They've framed it as a feature for security researchers and other developers who need to boot and test modified macOS kernels. This is possible as Apple still releases kernel source. In turn, since Apple isn't about to hand anyone else their private signing keys, they offer a way for end users to downgrade boot security for specific OS containers, making it possible to install and boot unsigned kernels.

These unsigned kernels don't have to be compiled from Darwin/XNU source code, of course. They can be anything. Apple knew this, and knew that it would lead to a Linux port (and more).

In practice, Apple has gone the opposite direction of obstruction. IIRC, it was about a year into the Apple Silicon Mac era when they added a bootloader feature which was utterly useless for themselves, but made life significantly easier for everyone working on booting non-Apple binary formats like Linux ELF.
 
Jul 27, 2020
17,168
11,026
106
Apple has indicated they're open to working with Microsoft to get this done, Microsoft doesn't seem interested.
Has Apple said so publicly or privately?
That's how a single engineer was able to port OS X to x86: much of the work was already done and actively maintained by others. The rest was mostly already done, but not maintained.

Thanks. I don't think that was revealed in the articles about Apple's x86 transition. The article I read (don't remember which one) made that lone engineer seem like a real star who single-handedly changed Apple's destiny with his efforts.

You often let the mask slip and make it clear you're trolling. "we all know how much Apple fears competition"? Give me a break.

Then I'm doing a bad job of trolling coz clearly more could be done by someone sufficiently motivated. Nah. I just speak whatever's on my mind. And thanks to that, you took it upon yourself to clear up a lot of stuff that I and probably many others here were not aware of. Thanks again.
 

roger_k

Member
Sep 23, 2021
90
174
76
Here you go. I've made it in the same format as in the link. Note that I quoted the correlation coefficient in my previous post but am showing the coefficient of determination (R^2) in the graph to match the link.

Correlation beween SPECInt and CB r20 results is purely coincidental in that they share the same confounder: CPU frequency. CB r20 is essentially a benchmark of L1D bandwidth + SIMD, and the CPU models you are looking at did not have any large changes in that domain.
 

Nothingness

Platinum Member
Jul 3, 2013
2,555
988
136
Correlation beween SPECInt and CB r20 results is purely coincidental in that they share the same confounder: CPU frequency. CB r20 is essentially a benchmark of L1D bandwidth + SIMD, and the CPU models you are looking at did not have any large changes in that domain.
Agreed. Are more recent versions of CB R20 harder on the memory side? Also a better support for AArch64 only came recently (it used to rely on compile-time translation of x86 intrinsics to NEON intrinsics IIRC).
 

roger_k

Member
Sep 23, 2021
90
174
76
Agreed. Are more recent versions of CB R20 harder on the memory side? Also a better support for AArch64 only came recently (it used to rely on compile-time translation of x86 intrinsics to NEON intrinsics IIRC).

The newest version (Cinebench 2024) has increased the scene size so that it does not fit in the cache of the most CPUs, and they also fixed a performance issue with their AVX emulation on Apple CPUs.

And of course, the newest Cinebench is being heavily criticized because suddenly Apple Silicon tops the charts, and that is obviously physically impossible, so the benchmark must be flawed.
 
Jul 27, 2020
17,168
11,026
106
The newest version (Cinebench 2024) has increased the scene size so that it does not fit in the cache of the most CPUs, and they also fixed a performance issue with their AVX emulation on Apple CPUs.
Which results did you see? URL?
 
Jul 27, 2020
17,168
11,026
106
That's not a huge lead. Apple can have it

Yes, a very power efficient lead but I invite Apple to make some porting magic happen so we can see more popular PC benchmarks on Apple Silicon and then we can have more data available to decide the Silicon Supremacy Battle.

13700K: https://pcpartpicker.com/list/8HZkt7

M3 10-core: https://www.bestbuy.com/site/apple-...b-latest-model-silver/6534764.p?skuId=6534764



I wonder how many cries of joy Apple users give out every time they run out of RAM or try to do advanced gaming that then turns into basic gaming.

Apple marketing is delusional.
 
Last edited:

Nothingness

Platinum Member
Jul 3, 2013
2,555
988
136
That's not a huge lead. Apple can have it

Yes, a very power efficient lead but I invite Apple to make some porting magic happen so we can see more popular PC benchmarks on Apple Silicon and then we can have more data available to decide the Silicon Supremacy Battle.

13700K: https://pcpartpicker.com/list/8HZkt7

M3 10-core: https://www.bestbuy.com/site/apple-...b-latest-model-silver/6534764.p?skuId=6534764

View attachment 99117

I wonder how many cries of joy Apple users give out every time they run out of RAM or try to do advanced gaming that then turns into basic gaming.

Apple marketing is delusional.
Are you really basing your judgement of M3 on Best Buy? Really?
 

Eug

Lifer
Mar 11, 2000
23,720
1,252
126
That's not a huge lead. Apple can have it

M3 10-core: https://www.bestbuy.com/site/apple-...b-latest-model-silver/6534764.p?skuId=6534764

View attachment 99117

I wonder how many cries of joy Apple users give out every time they run out of RAM or try to do advanced gaming that then turns into basic gaming.
? M3 is the entry level Mac chip. Not that gaming will be top tier on any Mac but if you want faster GPU, you’d get M3 Max or something instead.

Also, why are you linking Best Buy’s purchase guide anyway? That’s a little strange.

As for Cinebench, it doesn’t run on iPads. You won’t see M4 until it comes to the Mac.
 
Jul 27, 2020
17,168
11,026
106
Thanks for the corrections but just amusing to learn that "enormous" files on Apple hardware require just 24GB RAM:



I'm so glad Apple marketing isn't responsible for food industry marketing. We would be in the same plight as North Koreans.
 

Nothingness

Platinum Member
Jul 3, 2013
2,555
988
136
Enlighten me then.
This and other threads should have taught you plenty about Apple CPU and where Apple machines are good or not. Ample evidence has been provided that Apple is doing very well for performance (though not enough to compete against x86 CPU with many cores for rendering), and obviously bad for memory capacity in base models.

There simply is no need to link some random shop marketing blurb, you should know much better. I'd even say citing Apple marketing is pointless. We all know what to think of marketing here.
 
Reactions: Eug and Hitman928

Hitman928

Diamond Member
Apr 15, 2012
5,423
8,331
136
Correlation beween SPECInt and CB r20 results is purely coincidental in that they share the same confounder: CPU frequency. CB r20 is essentially a benchmark of L1D bandwidth + SIMD, and the CPU models you are looking at did not have any large changes in that domain.

I believe you missed the point.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |