Apple M2 CPU & GPU Benchmarks Surface on Geekbench

[Uskompuf]

The recently announced Apple M2 processor which is set to feature in the new MacBook Air and 13-inch MacBook Pro models has been benchmarked. The processor appeared in numerous Geekbench 5 CPU & GPU tests where the chip scored a maximum single-core result of 1919 points and 8928 points in multi-core representing an 11% and 18% CPU performance improvement respectively from the M1. The chip brings significant GPU performance increases achieving a Geekbench Metal score of 30627 points which is a ~42% increase from the M1 partially due to a larger 10-core GPU compared to the 8-core GPU on the M1. These initial numbers largely align with claims from Apple of an 18% CPU and 35% GPU improvement over the original M1.



View at TechPowerUp Main Site | Source

 

[Nephilim666]

Ahh I see the cash truck arrived safely at Geekbench HQ.

 

[DeathtoGnomes]

not much to say here except, GO APPLE!.

 

[Richards]

Sheep are still gonna buy this editing toy that can't play games

 

[Daven]

Sheep are still gonna buy this editing toy that can't play games

Sigh. Please explain why this is a toy compared to something like a Lenovo laptop with only Intel integrated graphics and Linux OS.

PS Something that plays games is universally known as an actual toy.

 

[siluro818]

Sheep are still gonna buy this editing toy that can't play games

I hope that was a joke, because it was a really funny one xD

 

[holyprof]

Too bad that awesome piece of hardware will end up exclusively in closed, anti-consumer Apple devices.
Here's a hint for you Apple - if you don't want to open your ecosystem, at least give a chance to OEMs to build linux or windows laptops based on the M CPUs. That would definitely put pressure on Intel and AMD. But enough of this nonsense ... Microsoft would never adapt Windows to M based devices, look how well Windows on ARM is doing (not). And as much as I hate Microsoft, I must admit that Windows on ARM failure is not MS fault, it's just too hard to crate new ecosystem that's competitive with existing ones: x86 with Linux and Windows, Android on ARM and Apple iOS/MacOS.

 

[bonehead123]

Nice performance numbers, but did they fix the underlying, unpatchable hardware vulnerability that was recently discovered ?

 

[londiste]

look how well Windows on ARM is doing (not). And as much as I hate Microsoft, I must admit that Windows on ARM failure is not MS fault, it's just too hard to crate new ecosystem that's competitive with existing ones: x86 with Linux and Windows, Android on ARM and Apple iOS/MacOS.

ARM lacks what x86 has - standardization and manufacturer support. And by that I do not mean the base ISA itself but bunch of adjacent parts - firmware/BIOS/UEFI and getting stuff to boot, driver support in any wide manner and maybe slightly ISA extensions as well although last ARM versions have been pretty nicely rounded with default extensions.

As for M2 scores, when comparing to M1 scores they do not seem to have made all that much progress in CPU. CPU scores are slightly faster but looking at some better M1 scores - in mac mini for example - singlecore result is 10% better and multicore 15% better while stated clocks are 3.5 vs 3.2 GHz, a 9% difference in favor of M2. GPU is both bigger and definitely benefits from all that additional memory bandwidth so that 35% improvement is very nice.

 

[SOAREVERSOR]

Sigh. Please explain why this is a toy compared to something like a Lenovo laptop with only Intel integrated graphics and Linux OS.

PS Something that plays games is universally known as an actual toy.

For one, just like the original M1 if you get a laptop with it you only get the laptop screen and one external monitor. I wouldn't call something that can't support multiple desktops a work item.

I'm not sure if it just full out can't or if apple isn't allowing it to hype their pro line with the more expensive chips, which actually come with the ports you'd need for work. And I know their are hoops you can jump through to kinda do it but you aren't getting multiple 4k 60hz from those jumps.

 

[Lionheart]

Sigh. Please explain why this is a toy compared to something like a Lenovo laptop with only Intel integrated graphics and Linux OS.

PS Something that plays games is universally known as an actual toy.

That so called Lenovo intel based laptop would be a quarter of the price also, therefore it's affordable, also by your toy description, all adults are still playing with toys lol.

 

[randomName]

I work with both,
13" HP ultrabook with i5 8250U (cost when purchased 2,2k eur), 16GB RAM, 512GB NVMe
and
13" MacBook Pro M1 (cost when purchased 1,9k eur), 16GB RAM, 512GB NVMe.

I only use single monitor, so thats all i need and they both suit me there.
M1 does the same workload as HP much faster at much lower power consumption.
Speaking of games. I don't play much, but one game that runs on both of them i can compare.
both laptops closed, using external monitor and measuring only laptop+charger power draw.

on HP, lowest settings hardly reaching 60fps, consuming 30W
on M1, mid to high settings easily running 60fps, consuming 15W

I am not an apple fanboy, but to me, apple is winning by far for my workloads

 

[SOAREVERSOR]

ARM lacks what x86 has - standardization and manufacturer support. And by that I do not mean the base ISA itself but bunch of adjacent parts - firmware/BIOS/UEFI and getting stuff to boot, driver support in any wide manner and maybe slightly ISA extensions as well although last ARM versions have been pretty nicely rounded with default extensions.

As for M2 scores, when comparing to M1 scores they do not seem to have made all that much progress in CPU. CPU scores are slightly faster but looking at some better M1 scores - in mac mini for example - singlecore result is 10% better and multicore 15% better while stated clocks are 3.5 vs 3.2 GHz, a 9% difference in favor of M2. GPU is both bigger and definitely benefits from all that additional memory bandwidth so that 35% improvement is very nice.

GPU still officially does not support dual external monitors. You have to go to the Max or Pro for that.

 

[mechtech]

Does PAC-MAN bug affect this chip??

 

[Daven]

I work with both,
13" HP ultrabook with i5 8250U (cost when purchased 2,2k eur), 16GB RAM, 512GB NVMe
and
13" MacBook Pro M1 (cost when purchased 1,9k eur), 16GB RAM, 512GB NVMe.

I only use single monitor, so thats all i need and they both suit me there.
M1 does the same workload as HP much faster at much lower power consumption.
Speaking of games. I don't play much, but one game that runs on both of them i can compare.
both laptops closed, using external monitor and measuring only laptop+charger power draw.

on HP, lowest settings hardly reaching 60fps, consuming 30W
on M1, mid to high settings easily running 60fps, consuming 15W

I am not an apple fanboy, but to me, apple is winning by far for my workloads

This. Everyone’s usage case differs. All companies are proconsumer and want to sell product. The rest of individual customer sentiment regarding products they don't use is informed or uninformed opinion. It’s your hard earned money and at least in the US, you have many choices on how to spend it.

 

[ERazer]

Hate all you want on Apple's walled garden, apple has control over hardware/software this new chip will be highly efficient with their OS.

Not like Microsoft had to pressure OEM's about spinning drives. https://www.techpowerup.com/295683/...-to-stop-use-of-hdds-as-boot-drives-from-2023

 

[SunMaster]

I work with both,
13" HP ultrabook with i5 8250U (cost when purchased 2,2k eur), 16GB RAM, 512GB NVMe
and
13" MacBook Pro M1 (cost when purchased 1,9k eur), 16GB RAM, 512GB NVMe.

I only use single monitor, so thats all i need and they both suit me there.
M1 does the same workload as HP much faster at much lower power consumption.
Speaking of games. I don't play much, but one game that runs on both of them i can compare.
both laptops closed, using external monitor and measuring only laptop+charger power draw.

on HP, lowest settings hardly reaching 60fps, consuming 30W
on M1, mid to high settings easily running 60fps, consuming 15W

I am not an apple fanboy, but to me, apple is winning by far for my workloads

Considering the i5 8250U was released in 2017 and the M1 Pro in 2021 it would be truly embarrasing if the M1 wasn't faster. Intel's i5 series has 4 cores, the M1 has 8. Hardly an apple to apple comparison.

 

[usiname]

Also 8250u is 14nm, m1 5nm

 

[SOAREVERSOR]

Hate all you want on Apple's walled garden, apple has control over hardware/software this new chip will be highly efficient with their OS.

Not like Microsoft had to pressure OEM's about spinning drives. https://www.techpowerup.com/295683/...-to-stop-use-of-hdds-as-boot-drives-from-2023

I don't mind the walled garden so much, I do mind some things within it though.

There's nothing wrong about having complete control over your hardware and software stack and that comes with advantages. But it also leads to artificial crippling, where apple cripples various products to force you into a high price range. With M series chips this gets bad. We know the M1 supported more than two external monitors because people jumped through hoops to make it so, but officially it does not and it's a PITA and they can take it away. So dual external monitors and standard stuff (USB, HDMI, SD card) are all mashed onto the higher end pro models that use the Pro and Max chips.

There are plenty of prosumer situations where the M2 air or M2 pro would make more sense from a usage standard but also want dual monitors. But they force you up a bracket.

 

[Shatun_Bear]

I work with both,
13" HP ultrabook with i5 8250U (cost when purchased 2,2k eur), 16GB RAM, 512GB NVMe
and
13" MacBook Pro M1 (cost when purchased 1,9k eur), 16GB RAM, 512GB NVMe.

I only use single monitor, so thats all i need and they both suit me there.
M1 does the same workload as HP much faster at much lower power consumption.
Speaking of games. I don't play much, but one game that runs on both of them i can compare.
both laptops closed, using external monitor and measuring only laptop+charger power draw.

on HP, lowest settings hardly reaching 60fps, consuming 30W
on M1, mid to high settings easily running 60fps, consuming 15W

I am not an apple fanboy, but to me, apple is winning by far for my workloads

I know you bought these systems a while back but that Intel 8th gen series of CPU were a bad idea when you bought it let alone now. Ancient 14nm CPU.

Compare that Pro M1 vs an ultrabook with a Ryzen 9 6900HS, which can be had for under $2K, for your workloads.

You'd see generational lift in performance depending on what you do of course.

 

[mama]

I will await a proper analysis and testing before coming to a view. Geekbench is a nothing, particularly where Macs are concerned.

 

[Minus Infinity]

ARM lacks what x86 has - standardization and manufacturer support. And by that I do not mean the base ISA itself but bunch of adjacent parts - firmware/BIOS/UEFI and getting stuff to boot, driver support in any wide manner and maybe slightly ISA extensions as well although last ARM versions have been pretty nicely rounded with default extensions.

As for M2 scores, when comparing to M1 scores they do not seem to have made all that much progress in CPU. CPU scores are slightly faster but looking at some better M1 scores - in mac mini for example - singlecore result is 10% better and multicore 15% better while stated clocks are 3.5 vs 3.2 GHz, a 9% difference in favor of M2. GPU is both bigger and definitely benefits from all that additional memory bandwidth so that 35% improvement is very nice.

Well of course not much progress. M1 is based on A14 cores and M2 is just the A15 cores even using the same NN processor. The only huge upgrade is the support for far faster memory and probably the main reason for the large GPU uplift. I just hope the M3 isn't another rehash using the A16 cores which is said to be a rather weak upgrade. I would be more excited about how Meteor Lake let alone Arrow Lake go, as the main focus is on mobile after Raptor Lake. AMD's Phoenix (Point) will also be pretty darn good and then things really heat up with Zen 5 + RDNA 4 APU's.

 

[Readlight]

More transistors, faster package extraction.

 

[randomName]

I know you bought these systems a while back but that Intel 8th gen series of CPU were a bad idea when you bought it let alone now. Ancient 14nm CPU.

Compare that Pro M1 vs an ultrabook with a Ryzen 9 6900HS, which can be had for under $2K, for your workloads.

You'd see generational lift in performance depending on what you do of course.

HP laptop was bought in 2019. Its the time when ryzen mobile could only be seen on the horizon, but in our country 100% of ultrabooks were intel for like 1 more year. And only HP had USB-C charging. The i7 model would cost additional 200 or 300 eur just for 100 or so MHz.

As we all know, intel was stagnating so it wasnt a surprise to find 2017 CPUs in laptops.

I now checked what can we get today. The best of ryzen is Lenovo thinkpad x13 with R7 5850U for almost 1,8k eur (16GB ram, 512GB nvme)
On the intel part, its the Dell latitude with i7-1185G7 for 1,7k eur. And to nobodys surprise is again a 4 core cpu.

Speaking of M1 being 8 core. If system information is showing cpu usage correctly, then i would say it never uses those 8 cores all together. Maxes out at 50% total cpu, which is probably 4 performance cores working. So technicaly is works as 4 core cpu.

 

[claes]

Speaking of M1 being 8 core. If system information is showing cpu usage correctly, then i would say it never uses those 8 cores all together. Maxes out at 50% total cpu, which is probably 4 performance cores working. So technicaly is works as 4 core cpu.

You need to use something like istat or istatistica to see per core usage. Activity monitor is pretty useless for monitoring performance unless you’re debugging.