M1 Benchmarks/comparisons

[hawkeye_a]

I was a little disappointed with the lack of benchmark/comparisons info in yesterday's presentation. I'm am not sure what the 3X or 6X faster (compared to what exactly?) means. Lack of context in general. I miss the days of yore when Phil Shiller and Steve Jobs would do performance-shootouts on stage. I realize that the M1 is more of an "entry level" processor and number-crunching probably isn't the point.

I'm curious about comparisons between existing intel Macs(using intel binaries), M1(using M1 binaries) and M1(using intel binaries on Rosetta2)....
-Applying filters to large images
-Encoding video (H264 and HEVC)
-Copying/moving large files (SSD speeds).
-FPS in games

And any other "real world" comparisons.

Cheers

 

[EugW]

Apple posted the specs for the hardware comparison on their website. It's different for each test, but suffice it to say, that they're comparing mainly against the older entry level Mac mini, MacBook Air, and MacBook Pro that were replaced, and not the higher end Intel models which are still for sale.

 

[Wyvernspirit]

Apple posted the specs for the hardware comparison on their website. It's different for each test, but suffice it to say, that they're comparing mainly against the older entry level Mac mini, MacBook Air, and MacBook Pro that were replaced, and not the higher end Intel models which are still for sale.

Which only makes sense. They will start comparing to the water models when they have options to replace them. I am interested in what those will look like but I am expecting a different M chip for those models.

 

[senna95]

Take a look at this thread with Cinebench R23 scores

 

[pshufd]

I'd prefer Geekbench 5 as that's my preferred benchmark. It would be nice to have official Apple results or one of the review sites.

 

[Homy]

BG 3 at Ultra settings 1080p on M1 in this video at 6:45: https://developer.apple.com/videos/play/tech-talks/10859/

Intel UHD Graphics 630 in Mac Mini 2018: 3.150 GPixel/s, 25.20 GTexel/s, FP32 403.2 GFLOPS

M1 Mac Mini 2020: 41 GPixel/s, 82 GTexel/s, FP32 2.6 TFLOPS

A12Z FP32 1.1 TFLOPS
Radeon Pro 560X 2.056 TFLOPS
Radeon Pro 5300M 3.2 TFLOPS
Radeon Pro 5300 4.2 TFLOPS
Radeon Pro 580X 5.530 TFLOPS

 

[NewUsername]

Geekbench now has a universal binary as well, so we’ll just have to wait until the first reviews are out.

 

[radus]

Apple M1 - Benchmark, Test and Specs

Apple M1 - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

Alleged A14X Bionic Benchmarks Show Huge Performance Jump vs A12Z Bionic • iPhone in Canada Blog

The very first benchmarks of the Apple A14X Bionic have just hit the web.

www.iphoneincanada.ca

Geekbench 5 Single Core 1634
Geekbench 5 Multi Core 7220

.. if their information is valid

 

[pshufd]

Apple M1 - Benchmark, Test and Specs

Apple M1 - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

Alleged A14X Bionic Benchmarks Show Huge Performance Jump vs A12Z Bionic • iPhone in Canada Blog

The very first benchmarks of the Apple A14X Bionic have just hit the web.

www.iphoneincanada.ca

Geekbench 5 Single Core 1634
Geekbench 5 Multi Core 7220

.. if their information is valid

That's the problem. What we need is Anand or Tom's or someone to state that they have a device in hand and that they are running native benchmarks.

 

[EugW]

Apple M1 - Benchmark, Test and Specs

Apple M1 - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

Alleged A14X Bionic Benchmarks Show Huge Performance Jump vs A12Z Bionic • iPhone in Canada Blog

The very first benchmarks of the Apple A14X Bionic have just hit the web.

www.iphoneincanada.ca

Geekbench 5 Single Core 1634
Geekbench 5 Multi Core 7220

.. if their information is valid

Initially when they first posted their A14X results I was a little bit hopeful they could be real (with a healthy dose of skepticism), but now I just think their info is bogus. The scores they provide are exactly the same for M1 and A14X. Not just very close, but exactly the same. This basically never happens even on the same machine, unless you re-run the test 20 times or something. And they posted the result for M1 the same day M1 was announced, before anyone had the device in hand. Similarly, they posted the A14X results eons ago, before anyone had such a device in hand.

 

[UltimateSyn]

BG 3 at Ultra settings 1080p on M1 in this video at 6:45: https://developer.apple.com/videos/play/tech-talks/10859/

Intel UHD Graphics 630 in Mac Mini 2018: 3.150 GPixel/s, 25.20 GTexel/s, FP32 403.2 GFLOPS

M1 Mac Mini 2020: 41 GPixel/s, 82 GTexel/s, FP32 2.6 TFLOPS

A12Z FP32 1.1 TFLOPS
Radeon Pro 560X 2.056 TFLOPS
Radeon Pro 5300M 3.2 TFLOPS
Radeon Pro 5300 4.2 TFLOPS
Radeon Pro 580X 5.530 TFLOPS

Could be a dumb question but is the teraflops of GPU performance expected to scale linearly with the number of cores? If so, that would give a theoretical 16-core Apple GPU in the 'M1X' about ~5.2 TFLOPS, which I believe is in line with what the MBP 16"'s current high-end GPU option (5600M) currently provides. Someone please correct me if that's inaccurate.

 

[Andropov]

is the teraflops of GPU performance expected to scale linearly with the number of cores?

Yes.

 

[Saturn007]

Any benchmarks around that are based on using common or leading apps?

I've laughed about upgrades in the past where the benchmarks showed 50%, 100%, or more increases in speed that were barely noticeable in real world use.

I like the type of info that shows seconds to launch the laptop; seconds to open key apps; seconds to open a 250-page doc or 5,000 row spreadsheet; seconds to calculate a correlation of two variables and draw a scatter plot; seconds to convert from one video format to another; seconds to shut down entirely; etc.

 

[moosinuk]

Have a look at this thread: https://forums.macrumors.com/thread...d-on-a14.2266884/?post=29209940#post-29209940 and read the Anandtech article

 

[Homy]

Could be a dumb question but is the teraflops of GPU performance expected to scale linearly with the number of cores? If so, that would give a theoretical 16-core Apple GPU in the 'M1X' about ~5.2 TFLOPS, which I believe is in line with what the MBP 16"'s current high-end GPU option (5600M) currently provides. Someone please correct me if that's inaccurate.

As I've read many times it's true for GPUs so an iMac with 16-core GPU and 5.2 TFLOPS could be at least as good as Radeon 5500 XT with 5.3 TFLOPS.

 

[phobos]

Any benchmarks around that are based on using common or leading apps?

Cinebench is exactly that. You run it and you know exactly how your machine will perform/stack up when you render with Cinema 4D.

And now that it runs for a longer time (you can also set your own time) it takes in to account thermal throttling.

 

[Homy]

M1 GPU is on par with Radeon Pro 560X: https://browser.geek...compute/1802234

 

[vjson]

This is almost ridiculous, the M1 seems absolutely incredible even in a passively cooled MacBook Air! https://browser.geekbench.com/v5/cpu/4661599

Single core score is 33% better than my iMac i9, and my i9 (with 64GB RAM) only manages a 33% better multicore score despite having 10 cores compared to the M1's 4 high performance cores. Seems I could have saved a lot of of money by buying an M1 laptop along with an external display instead of an iMac i9!

 

[EugW]

This is almost ridiculous, the M1 seems absolutely incredible even in a passively cooled MacBook Air! https://browser.geekbench.com/v5/cpu/4661599

Single core score is 33% better than my iMac i9, and my i9 (with 64GB RAM) only manages a 33% better multicore score despite having 10 cores compared to the M1's 4 high performance cores. Seems I could have saved a lot of of money by buying an M1 laptop along with an external display instead of an iMac i9!

Note that the MacBook Air will throttle with extended workloads. If you really need that kind of performance on a regular basis for extended periods, you're better off with a MacBook Pro (or Mac mini).

 

[pshufd]

They may have reported performance on a Mini for the M1 which has the best cooling letting everyone assume that you get the same performance in the MacBook Air and MacBook Pro 13.

 

[Homy]

In graphic (not compute) tasks, the M1 could be on par with the Radeon Pro 570X (and way ahead of the 560X). The TBDR architecture of the M1 (for which Metal has been tailored) benefits graphics more than it benefits compute. Also, Apple GPUs can use 16-bit AND 32-bit numbers in shaders, for precision and to boost efficiency, which PC GPUs can't.

That's great! I suspected that since it can render more pixels/s:

M1 41 GPixel/s, 82 GTexel/s
Pro 560X 16.06 GPixel/s, 64.26 GTexel/s
Pro 570X 35.36 GPixels/s, 123.8 GTexel/s
Pro 580X 38.4 GPixels/s, 172.8 GTexel/s
Pro 5300 52.8 GPixels/s, 132 GTexel/s
Pro 5500 XT 56.22 GPixels/s, 168.7 GTexel/s
Pro 5700 86.4 GPixels/s, 194.4 GTexel/s
Pro 5700 XT 95.94 GPixels/s, 239.8 GTexel/s

 

[OriginalClone]

I was a little disappointed with the lack of benchmark/comparisons info in yesterday's presentation. I'm am not sure what the 3X or 6X faster (compared to what exactly?) means. Lack of context in general. I miss the days of yore when Phil Shiller and Steve Jobs would do performance-shootouts on stage. I realize that the M1 is more of an "entry level" processor and number-crunching probably isn't the point.

I'm curious about comparisons between existing intel Macs(using intel binaries), M1(using M1 binaries) and M1(using intel binaries on Rosetta2)....
-Applying filters to large images
-Encoding video (H264 and HEVC)
-Copying/moving large files (SSD speeds).
-FPS in games

And any other "real world" comparisons.

Cheers

Hoping I find out Monday. My air is scheduled for delivery on Monday but that could change. I’m gonna do some exports in premiere on the air (16gb ram) vs my 2018 15” mbp 32gb ram w/560x. I highly doubt it will match performance, but I’ll see how close or far it is.

 

[Kung gu]

Hoping I find out Monday. My air is scheduled for delivery on Monday but that could change. I’m gonna do some exports in premiere on the air (16gb ram) vs my 2018 15” mbp 32gb ram w/560x. I highly doubt it will match performance, but I’ll see how close or far it is.

Adobe Premiere is not native on the M1, so it’s not a fair comparison. I know cinebench is now native for cpu tasks, same with affinity is native too.

 

[wyrdness]

Hoping I find out Monday. My air is scheduled for delivery on Monday but that could change. I’m gonna do some exports in premiere on the air (16gb ram) vs my 2018 15” mbp 32gb ram w/560x. I highly doubt it will match performance, but I’ll see how close or far it is.

It'll be an interesting comparison, with Rosetta on the Air vs your MBP. As you say, I doubt that it will match performance, but I'd be curious to see how much faster the Air is ?

 

[hawkeye_a]

Hoping I find out Monday. My air is scheduled for delivery on Monday but that could change. I’m gonna do some exports in premiere on the air (16gb ram) vs my 2018 15” mbp 32gb ram w/560x. I highly doubt it will match performance, but I’ll see how close or far it is.

Just a thought..... maybe using Quicktime export would be a "fairer" comparison between the two?

Cheers