Apple M4 Chip Benchmarked: 22% Faster Single-Core and 25% Faster Multi-Core Performance

[AleksandarK]

Yesterday, Apple launched its next-generation M4 chip based on Apple Silicon custom design. The processor is a fourth-generation design that brings AI capabilities and improved CPU performance. First debuting in an iPad Pro, the CPU has been benchmarked in Geekbench v6. And results seem to be very promising. The latest M4 chip managed to score 3,767 points in single-core tests and 14,677 points in multi-core tests. Compared to the M3 chip, which scores 3,087 points in single-core and 11,702 in multi-core tests, the M4 chip is about 22% faster in single-core and 25% faster in multi-core synthetic benchmarks.

Of course, these results are not real-world use cases, but they give us a hint of what the Apple Silicon design team has been working on. For real-world results, we have to wait a little longer to see reviews and results from devices such as MacBook Pro and MacBook Air, which should have better cooling and possibly better clocks for the chip.



View at TechPowerUp Main Site | Source

 

[Bruno Vieira]

Its a impressive jump to make in the same node and in this short amount of time since the M3

 

[R0H1T]

Not really, I'm thinking clocks are higher as well. Probably 5-15% just guesstimating based on how underwhelming Apple's IPC gains have been in the recent past.

 

[Tek-Check]

Its a impressive jump to make in the same node and in this short amount of time since the M3

Don't forget there are more cores on M4 CPU, plus higher clocks.

There is just as much as one could get by moving from N3E to N3B. There are no miracles.

I'd imagine IPC uplift is 5-8%. This doesn't translate into equal uplift in real life workloads. In some it could be more than on others. It's certainly more in Geek Geekbench 6.

 

[Daven]

Don't forget there are more cores on M4 CPU, plus higher clocks.

There is just as much as one could get by moving from N3E to N3B. There are no miracles.

I'd imagine IPC uplift is 5-8%. This doesn't translate into equal uplift in real life workloads. In some it could be more than on others. It's certainly more in Geek Geekbench 6.

The miracle is the passive cooling for such a powerful SoC.

 

[dorioku]

single core 22% faster , more 2 E core and N3E, but multi core only 25% faster?

 

[londiste]

Power? Frequencies?

 

[Tek-Check]

The miracle is the passive cooling for such a powerful SoC.

Well, we will need to wait and see how powerful passive cooling really is in daily workloads, when tested by reviewers.

 

[Nhonho]

Do you know why Apple still hasn't put an AV1 encoder in its CPUs?

 

[Chrispy_]

I wish people wouldn't use Geekbench. It's a very very very loose indicator of real-world performance, and a much better indicator of how well Geekbench has been tuned to accommodate specific architecture in the latest update.

Apple silicon is impressive, mostly by using the most expensive fabrication process available and ditching all the legacy architecture that makes it run anything other than the miniscule selection of Apple-compiled software run at a slower, less-efficient emulation speed. At the end of the day though, it's just your everyday ARM with an absolute shit ton of Apple marketing budget and weasel-wording behind it.

This latest M4 improvement seems to have murdered power efficiency, which was Apple Silicon's biggest prior advantage. It's clear that Apple are discovering the laws of physics just like AMD, Intel, and Nvidia did all those years ago...

 

[Daven]

I wish people wouldn't use Geekbench. It's a very very very loose indicator of real-world performance, and a much better indicator of how well Geekbench has been tuned to accommodate specific architecture in the latest update.

Apple silicon is impressive, mostly by using the most expensive fabrication process available and ditching all the legacy architecture that makes it run anything other than the miniscule selection of Apple-compiled software run at a slower, less-efficient emulation speed. At the end of the day though, it's just your everyday ARM with an absolute shit ton of Apple marketing budget and weasel-wording behind it.

This latest M4 improvement seems to have murdered power efficiency, which was Apple Silicon's biggest prior advantage. It's clear that Apple are discovering the laws of physics just like AMD, Intel, and Nvidia did all those years ago...

Geekbench is used because its possible to compare across different OSes and hardware.

 

[bencrutz]

Its a impressive jump to make in the same node and in this short amount of time since the M3

impressive due to the SME, the highest jump, look at the details of the GB score.
CPU IPC increase is actually meh

hint: object detection

 

[Nater]

And still can't run any apps I use. So. Meh.

 

[Geofrancis]

geekbench is a joke.

I want to see some classic benchmarks that actually use the CPU cores and not the bunch of accelerators attached to it.

I bet if someone ran something like Pifast or wprime it wouldnt look so good.

 

[Guwapo77]

M4 has a great performance increase of 22-25%, but its nothing compared to Zen 5's 10% increase. *insert clown emoji*

 

[phints]

Yet another glorified tablet CPU leveraging the most expensive lithography known to man. Locked down doesn't work with Vulkan, GPUs, or much else. Pass.

 

[Fourstaff]

Stages of grief:
Apple chips are inferior <- we are past this already
Benchmarks are rubbish <- we are here
I don't use their software anyway <- some of us are already here
They are using the latest and most expensive nodes, that's unfair comparison <- one of us is here
Apple chips are actually good <- next step

/S

 

[G777]

So much copium in the comments

 

[Chrispy_]

Stages of grief:
Apple chips are inferior <- we are past this already
Benchmarks are rubbish <- we are here
I don't use their software anyway <- some of us are already here
They are using the latest and most expensive nodes, that's unfair comparison <- one of us is here
Apple chips are actually good <- next step

/S

Apple chips are actually good. They're state-of-the-art ARM.

My issue is with Geekbench being used.

Most of the M1, M2, M3, M4 performance can be attributed to Apple forcing developers to recompile their software specfically for their silicon. They ditched compatibility to gain performance which is a perfectly acceptable trade-off for Apple's walled garden. ARM performance and IPC is a known quantity; Apple are bolstering it with high-speed interfaces to fast storage and fast RAM, a decent power budget and state of the art manufacturing for the best performance and the best performance/Watt. A lot fhe M-silicon's success is down to Apple's software ecosystem and clout with Developers, but this also isn't just your basic tablet; These things have high-end platform componentry and a level of integration that x86 hardware vendors will never get close to. Even if it's just ARM, it's the best example of ARM you'll ever see.

 

[Steevo]

single core 22% faster , more 2 E core and N3E, but multi core only 25% faster?

Boost clocks and more power gating, no doubt also in their very own special workload that fits their architecture perfectly.

 

[AnotherReader]

Stages of grief:
Apple chips are inferior <- we are past this already
Benchmarks are rubbish <- we are here
I don't use their software anyway <- some of us are already here
They are using the latest and most expensive nodes, that's unfair comparison <- one of us is here
Apple chips are actually good <- next step

/S

Apple's chips are excellent, but the point about nodes is a valid one. When the shoe was on the other foot, RISC vendors used to cry ad nauseam about Intel's node advantage.

 

[R0H1T]

Well tbf AMD users, like me, did as well. It's a valid argument to show why X is better than Y, also Apple's chips have a slightly bandwidth advantage in that form factor. Will be great to see Strix Point(Halo?) with up to 256bit LPDDR5x memory to extract the best out of AMD's biggest advantage in the space i.e. their GPU.

 

[Fouquin]

Not really, I'm thinking clocks are higher as well. Probably 5-15% just guesstimating based on how underwhelming Apple's IPC gains have been in the recent past.

IPC gains have sucked because it's been the same cores since M1. They haven't done any drastic changes except to L2 caches, fabric/SLC bandwidth, and updates to various fixed function accelerators. The new GPU architecture was rolled out with M2, mildly gimped. The un-gimped GPU was rolled out with M3. Now we have both an improved and fully un-gimped GPU and a real core architecture improvement on both big and LITTLE cores.

 

[Space Lynx]

Well, we will need to wait and see how powerful passive cooling really is in daily workloads, when tested by reviewers.

true, m3 MacBook air throttles like a mother fucker. lol

Apple really should have bought the AirJet company and utilized it in its design

 

[R0H1T]

IPC gains have been minimal for much longer than, IIRC they've more or less crawled since 2018(19?) with the A series doing similar stuff.