Apple M3 Pro Chip Has 25% Less Memory Bandwidth Than M1/M2 Pro

[UK-MacAddict]

Um, the 2020 iMac 27" is not even four years old yet. And can still be found in the refurb store.

So yes, Apple did make an item (27" Intel iMac) in their refurb store look obsolete, with those M3 graphics and spiel.

3 years in tech is ancient. Should be comparing the direct predecessor to the current chip. So M3 vs M2 comparison.

Why even compare it to the M1 it’s a pointless comparison since it’s already 2 generations old.

Same crap in the iPhones compare the 15 Pro to the 14 Pro. Nothing earlier than that.

 

[Kung]

I actually don’t like the new colour, but I totally get your point. I checked what it would cost for le to get a MBP with M3 Pro chip with specs I need, it’s 100€ more now. I’m definitely not regretting my MBP M2Pro. And I’ll be honest, I was getting a little nervous prior to the event.

Agreed. Of course, the actual tell will be when serious indepth testing is done on transfer rates and the like.

With that said, even with my veteran's discount and a trade-in, it'd be a $400 difference to upgrade to an M3 Pro with the same specs....so unless they have a huge fire sale @ some point in the future, OR indepth testing shows they magically started using SSDs with higher read/write rates or something (hah), I'll be just fine with my M2 Pro.

 

[JippaLippa]

I expected a little more from the M3 generation, especially given the new production process.
Still, I welcome the improvements.
Provided I'll be able to afford it, I'll switch from my M1 Ultra Mac Studio to an M3 Ultra in mid 2024.
The M1 Ultra has been good, however I always felt I needed 30-40% more power and the M3 Ultra will be perfect for it.

 

[wegster]

M5! Beware. It may decide to destroy your entire fleet of starships. Keep your phasers locked down.

Nah - Siri would try and then get confused.

 

[bradman83]

I, too, have read this elsewhere about the distinction between version three and four. However, that does not appear to be the case on the spec sheet see photo.

A little known fact about the M2 Mac Mini is that it supports dual displays over Thunderbolt. From Apple's spec sheet for the base M2 mini:

Simultaneously supports up to two displays:

• One display with up to 6K resolution at 60Hz over Thunderbolt and one display with up to 5K resolution at 60Hz over Thunderbolt or 4K resolution at 60Hz over HDMI

This is confusingly worded, but it's clear the M2 Mini's can support two displays over Thunderbolt or one over Thunderbolt and one over HDMI. While it doesn't explicitly talk about one port, there are plenty of user stories out there of people who have used Thunderbolt 4 docks to power two displays off of a single Thunderbolt port on a base M2 Mini without DisplayLink or other forms of trickery (something that did not work on the M1 Mini). So the M2 Mini meets Thunderbolt 4 specifications, whereas the Air and base MBP do not.

I suspect the base M2 doesn't have the I/O bandwidth to drive more than two displays, so on a laptop it will always be limited to one external display plus the internal display, whereas the Mini has to rely on external displays and can therefore properly implement the Thunderbolt 4 spec.

 

[UK-MacAddict]

Just another nit picky way Apple has found screw the customer.

 

[UK-MacAddict]

Whole thing stinks of rush job TBH.

Jumped the gun on Apple Silicon they’re already hitting a performance ceiling. Gains in performance across the board is negligible.

 

[JimmyG]

Taken together, it's presently unclear what real-world difference these changes make to M3 performance when pitted against Apple's equivalent precursor chips in various usage scenarios, especially given that the latest processors include new Dynamic Caching memory allocation technology which ensures that only the exact amount of memory needed is used for each task.

​

This opaqueness is not helped by the fact that Apple advertises the power of the new M3 Pro and M3 Max chips by repeatedly emphasizing comparisons to the M1 Pro and M1 Max, rather than the more recent M2 variants, against which performance gains appear more modest. Hopefully we will learn more in time when the first thoroughgoing third-party benchmarks become available.

Is it that folks did not pay attention during the live event where no less than 7 graphics were provided showing M1/M2/M3 performance graphics, or are those folks not actually looking at Apple's own web pages for this very information where they provide no less than 40 performance graphics along with testing footnotes?

IMHO, potential shoppers can find more than ample examples of numerous real-world application usage scenarios from personal to workplace to entertainment, etc., to help them make their purchasing decisions on Apple's product pages, you'll have to do the scrolling and viewing and reading on your time...

MacBook Pro - Apple

MacBook Pro

MacBook Pro laptop with M4, M4 Pro, and M4 Max chips. Built for Apple Intelligence. Up to 24 hours of battery life. Liquid Retina XDR display.

www.apple.com

...FYI, click the "Go deeper on M3 chips +" button.

Now, if anyone has a bone to pick about any of Apple's test results perhaps they can specify from which of the footnotes below that they would call into question. Otherwise, complaints and aspersions about Apple's, er, selectiveness as to which chips it has-or-has-not made comaprisons with is nothing more than unfounded misrepresentations of what is clearly visible and available for public viewing and inaccurate reporting.

TLDR: Apple has clearly shown test results vs M2 throughout its testing.

Footnotes from above link provided for clarity...

<<
1. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 512GB SSD. The wireless web test measures battery life by wirelessly browsing 25 popular websites with display brightness set to 8 clicks from bottom. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration. See apple.com/batteries for more information.

2. MacBook Pro with M3 chip is not available in Space Black.

3. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Redshift v3.5.18 tested using a 29.2MB scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

4. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 48GB of RAM, and 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM. Performance measured using select industry‑standard benchmarks. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

5. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Super Resolution tested with Photomator 3.1.2 using a 4.4MB image. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

6. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Prerelease Blender 4.0.0 tested by rendering a 47.1 MB 3D scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

7. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Final Cut Pro 10.6.9 tested using a complex 2-minute project with 4K ProRes 422 media. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

8. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Tested with Affinity Photo 2 v2.1.1 using the built-in benchmark version 21000. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

9. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Open source project built with Xcode 15.0 and Apple Clang 15.0.0. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

10 Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Tested with a selection of tasks using Microsoft Excel v16.77.1. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

11. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M2, 8-core CPU, 10-core GPU, and 24GB of RAM, production 13-inch MacBook Pro systems with Apple M1, 8-core CPU, 8-core GPU, and 16GB of RAM, and production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645 and 16GB of RAM, all configured with 2TB SSD. Tested with Logic Pro 10.7.9 with project consisting of multiple tracks, each with an Amp Designer plug-in instance applied. Individual tracks were added during playback until CPU became overloaded. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

12. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, both configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with MATLAB and Simulink R2023b v23.2.0 and Parallel Computing Toolbox using a vehicle dynamics model. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

13. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, both configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Redshift v3.5.18 tested using a 29.2MB scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

14. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Redshift v3.5.18 tested using a 29.2MB scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

15. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with MATLAB and Simulink R2023b v23.2.0 and Parallel Computing Toolbox using a vehicle dynamics model. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

16. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested using Adobe Premiere Pro (Beta) 24.1, and a project with multiple dialog-heavy video clips. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

17. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with a 258MB study size using prerelease Horos MD. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

18. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Adobe Photoshop 25.0.0 tested using the following filters and functions: color range, minimum, maximum, rotate image arbitrary, and content-aware fill. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

19. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, production 14-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, 32GB of RAM, and 8TB SSD, production 14-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, 32GB of RAM, and 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with Affinity Photo 2 v2.1.1 using the built-in benchmark version 21000. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

20. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Redshift v3.5.18 tested using a 29.2MB scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

21. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with MATLAB and Simulink R2023b v23.2.0 and Parallel Computing Toolbox using a vehicle dynamics model. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

22. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested using Adobe Premiere Pro (Beta) 24.1, and a project with multiple dialog-heavy video clips. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

23. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with a 258MB study size using prerelease Horos MD. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

24. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Adobe Photoshop 25.0.0 tested using the following filters and functions: content-aware scale, perspective warp, unsharp mask, motion blur, and bas relief. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

25. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 36GB of RAM, and 4TB SSD, as well as production 16-inch MacBook Pro systems with Apple M2 Pro, 12-core CPU, 19-core GPU, and 32GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Pro, 10-core CPU, 16-core GPU, and 32GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with Affinity Photo 2 v2.1.1 using the built-in benchmark version 21000. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

26. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, 128GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Final Cut Pro 10.6.9 tested using a 5-minute clip with 4K Apple ProRes 4444 media, at 3840x2160 resolution and 23.98 frames per second, transcoded to Apple ProRes 422. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

27. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Redshift v3.5.18 tested using a 29.2MB scene utilizing hardware-accelerated ray tracing on M3-based systems, and software-based ray tracing on all other units. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

28, Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with MATLAB and Simulink R2023b v23.2.0 and Parallel Computing Toolbox using a vehicle dynamics model. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

29. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with Affinity Photo 2 v2.1.1 using the built-in benchmark version 21000. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

30. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested with a 258MB study size using prerelease Horos MD. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

31. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Tested using Adobe Premiere Pro (Beta) 24.1, and a project with multiple dialog-heavy video clips. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

32. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD, and production 2.3GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics, 32GB of RAM, and 4TB SSD. Adobe Photoshop 25.0.0 tested using the following filters and functions: color range, minimum, maximum, rotate image arbitrary, and content-aware fill. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

33. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 14-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, and production 14-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, all configured with 8TB SSD. DaVinci Resolve Studio 18.6 tested using a 28-second project with Apple ProRes 4444 XQ UHD media and spatial noise reduction effect applied, on systems that supported the workload. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

34. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with MATLAB and Simulink R2023b v23.2.0 and Parallel Computing Toolbox using a vehicle dynamics model. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

35. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with Affinity Photo 2 v2.1.1 using the built-in benchmark version 21000. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

36. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested with a 258MB study size using prerelease Horos MD. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

37. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Tested using Adobe Premiere Pro (Beta) 24.1, and a project with multiple dialog-heavy video clips. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

38. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. Adobe Photoshop 25.0.0 tested using the following filters and functions: content-aware scale, perspective warp, unsharp mask, motion blur, and bas relief. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

39. Testing conducted by Apple in September and October 2023 using preproduction 16-inch MacBook Pro systems with Apple M3 Max, 16-core CPU, 40-core GPU, and 128GB of RAM, production 16-inch MacBook Pro systems with Apple M2 Max, 12-core CPU, 38-core GPU, and 96GB of RAM, production 16-inch MacBook Pro systems with Apple M1 Max, 10-core CPU, 32-core GPU, and 64GB of RAM, and production 2.4GHz 8-core Intel Core i9-based 16-inch MacBook Pro systems with Radeon Pro 5600M graphics with 8GB of HBM2, and 64GB of RAM, all configured with 8TB SSD. DaVinci Resolve Studio 18.6 tested using a 28-second project with Apple ProRes 4444 XQ UHD media and spatial noise reduction effect applied. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

40. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3, 8-core CPU, 10-core GPU, 8GB of RAM, and 512GB SSD. The wireless web test measures battery life by wirelessly browsing 25 popular websites with display brightness set to 8 clicks from bottom. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration. See apple.com/batteries for more information.

41. MacBook Pro with M3 Pro chip supports up to two external displays. MacBook Pro with M3 Max chip supports up to four external displays.

42. Testing conducted by Apple in September and October 2023 using preproduction 14-inch MacBook Pro systems with Apple M3 Pro, 12-core CPU, 18-core GPU, 18GB of RAM, and 1TB SSD. The wireless web test measures battery life by wirelessly browsing 25 popular websites with display brightness set to 8 clicks from bottom. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration. See apple.com/batteries for more information.

43. Claim refers to 16‑inch MacBook Pro.

44. Port configuration varies by model.

45. Trade-in values will vary based on the condition, year, and configuration of your eligible trade-in device. Not all devices are eligible for credit. You must be at least 18 years old to be eligible to trade in for credit or for an Apple Gift Card. Trade-in value may be applied toward qualifying new device purchase or added to an Apple Gift Card. Actual value awarded is based on receipt of a qualifying device matching the description provided when estimate was made. Sales tax may be assessed on full value of a new device purchase. In-store trade-in requires presentation of a valid photo ID (local law may require saving this information). Offer may not be available in all stores and may vary between in-store and online trade-in. Some stores may have additional requirements. Apple or its trade-in partners reserve the right to refuse or limit quantity of any trade-in transaction for any reason. More details are available from Apple’s trade-in partner for trade-in and recycling of eligible devices. Restrictions and limitations may apply.

46. Testing conducted by Apple in April and May 2023 using preproduction 15‑inch MacBook Air systems with Apple M2, 8-core CPU, 10-core GPU, 8GB of RAM, and 256GB SSD. The Apple TV app movie playback test measures battery life by playing back HD 1080p content with display brightness set to 8 clicks from bottom. Battery life varies by use and configuration. See apple.com/batteries for more information.
>>

 

[EvilMonk]

For once my MacBook Pro 16” M1 Max 32c/32Gb/1Tb doesn’t have much to envy the new one… the M2 Max coming so short after was a punch in the guts but this one don’t hurt as much knowing i got the MBP 16” M1 Max at launch…

 

[brooklyn360]

My workplace supplies my laptop so don’t much of a say. It’s a good specced machine but of course has MDM. My m1 Mac mini works perfectly and I really don’t want to buy a PC for personal use as I prefer the Mac UI/UX and longevity

Well, then you are going to miss out on gaming. There is no need to buy a high spec MBP or Mac Studio, it will not be supported for gaming. If you M1 mac mini does everything you need then that's fine. You can always get a Nintendo Switch or Xbox and connect to your tv.

 

[Mcdevidr]

The M1 was so far ahead of its time that it was ridiculous. Especially for single threaded performance, which Intel/AMD couldn't even touch at the time.

Now, Intel and AMD have really caught up on that front, and are even outperforming the M2 in single threaded performance on some of their higher end SKUs. Unfortunately for Intel, this comes at a pretty dramatic cost for power consumption, as they've relied on raising clock speeds through the roof to do it.

The competition really heated up for Apple in terms of raw performance (without regard to power consumption) though.

The intel chips supposedly can outperform apple on single thread. But take the 13th gen i9 in the razer book 16. In a game I play that is heavily single core dependent and cache bound the base m2 performs almost in par because the damn i9 won’t maintain a boost even over 4ghz. Now in a desktop sure. But this happens on intel’s low power skus as well in the game, the CPU’s can not maintain any kind of boost and a $500 open box Mac mini gets me the same performance (speed in game, graphics of course worse) as a $4000 laptop that is loud and heavy.

 

[AppliedVisual]

I am still convinced that Meteor lake is gonna make Apple look foolish for switching away. A lot of the power advantage came from the smaller node Apple was using vs the old 14nm intel chips apple was using. Much of that is solved with meteor lake which is the first chip built on Intel 4. Intel is expected to have a node advantage over the competition in a few years as well with the release of the 20A and 18A nodes. The loss of dedicated GPU support is still a huge loss, especially for products like the Mac Pro which substantially behind even NVIDIA's consumer mid range cards, and orders of magnitude slower than their high end cards.

Nothing is going to make Apple look foolish for switching. Intel has been underperforming for some time now and allowed the competition to swoop in. Apple built their own solutions to serve their own products, and did an exceptional job. There was no advantage to stick around... With Apple now in control of their own chip design, if they do tap Intel again in the future, it will be for production/ fab services.

I do think Meteor Lake will turn it around for Intel. Excited to see what their new platforms offer. I'm not expecting miracles though.

Discrete GPUs are a definite advantage to the PC world. I don't think people give Apple enough credit for their GPU designs though. Their GPU cores are quite impressive, really. Just very expensive and limited in quantity for most Mac configurations. It's too bad they're not making dedicated GPUs for the Mac Pro that stack hundreds of their GPU cores into a PCIe4 card.

 

[Stewart Gooderman]

A question to the experts on this blog. In July, I bought a 16" MBP with 32GB memory and a 4TB SSD which forced me into a Max processor (I originally just wanted a Pro processor). Now, the M3 36GB/4TB SSD machine allows me to buy the Pro chip rather than the Max for $200 less than what I paid for my machine. Is there much difference in performance between an M2 Max and an M3 Pro?

 

[Deanster]

It's always a pleasure when BOTH the 'it's INCREMENTAL' and the 'only better than earlier gens!' people get upset at the same time. M1 spoiled us with a huge jump in capability across the line. Anyone who imagined there'd be 30%+ annual improvements from there hasn't been paying attention. ~15% per generation is doing well, and works out to something like a 50% increase on a three year cycle, which is just about right to really feel the improvement.

This is of course why they pitch the improvement over earlier gens, like last intel models (of which there are MANY still in use, of course), and M1's (many), vs M2's (plenty, but a slightly narrower market given M1/2's offered simultaneously for a long period.

I'm a launch day M1 Pro MBP buyer, and while I'm a heavy Lightroom user and a medium Final Cut Pro user, I'm very happy with the speed/power use balance of my MBP. I'll be interested to see how the various levels of M3 compare to my M1 Pro, but I don't expect that the M3 will be ready to beat my M1 Pro at a lower price or that the M3 Pro would meaningfully beat an M1/2 Max at a lower price, either of which would be interesting. I'll also look and see what M3 Max brings to the table, but I'd be more than a little surprised if it offers enough more power to get me looking at upgrades.

 

[macjaffa]

yep, upselling at its finest
anyone buying, you either go M3 or M3 Max
M3 Pro is basically useless unless you want the GPU power

So…. Not useless then? 😂

 

[macjaffa]

Some posters in other places have suggested that this is due to a combination of yield issues + changing the bus width of the RAM + available SKUs of RAM chips.

The M3 Pro and lower-tier Max appear to be using triple-channel RAM as opposed to the quad-channel of previous years.

The tradeoff that Apple seems to have made here is lower bandwidth in favor of more memory (6 or 12 GiB per channel rather than 4 or 8). Whether that is a good tradeoff remains to be seen.

It is also worth mentioning that -- when looking at the die shots from the keynote -- the M3 Pro chip appears to be a completely different chip design that is more similar to the M3 in layout than a cut-down M3 Max.

Notably, the M3 Max retains the cut line in its die that was used in previous generations to make a Pro chip (by cutting off the extra GPU cores). This, and the fact that the Max chip MacBooks will only be available later, may also point to yield issues that required a new design for the Pro chip.

Has anyone managed to overclock a m1/m2 yet by getting a processor that failed the pro bin but still had the transistors aka old school overclocking when people would get lower speed procs that were just failed higher tier procs or something

 

[zapmymac]

A little known fact about the M2 Mac Mini is that it supports dual displays over Thunderbolt. From Apple's spec sheet for the base M2 mini:

Simultaneously supports up to two displays:

• One display with up to 6K resolution at 60Hz over Thunderbolt and one display with up to 5K resolution at 60Hz over Thunderbolt or 4K resolution at 60Hz over HDMI

This is confusingly worded, but it's clear the M2 Mini's can support two displays over Thunderbolt or one over Thunderbolt and one over HDMI. While it doesn't explicitly talk about one port, there are plenty of user stories out there of people who have used Thunderbolt 4 docks to power two displays off of a single Thunderbolt port on a base M2 Mini without DisplayLink or other forms of trickery (something that did not work on the M1 Mini). So the M2 Mini meets Thunderbolt 4 specifications, whereas the Air and base MBP do not.

I suspect the base M2 doesn't have the I/O bandwidth to drive more than two displays, so on a laptop it will always be limited to one external display plus the internal display, whereas the Mini has to rely on external displays and can therefore properly implement the Thunderbolt 4 spec.

Very interesting. My base M2 is only ~2 weeks old, still finding interesting tidbits ✌️

 

[Biro]

A question to the experts on this blog. In July, I bought a 16" MBP with 32GB memory and a 4TB SSD which forced me into a Max processor (I originally just wanted a Pro processor). Now, the M3 36GB/4TB SSD machine allows me to buy the Pro chip rather than the Max for $200 less than what I paid for my machine. Is there much difference in performance between an M2 Max and an M3 Pro?

I don’t think those benchmarks are out yet. But… I’ll offer this: After the M2 processors came out, the M1 Pro was still faster than the base M2. I can’t guarantee anything, but I think you’re fine with the M2 Max.

 

[kjvmartin]

The mid-range models are too good, so they are nerfing them to push people to spend more on the top end models. Clearly done to boost the revenues in the short term in the face of declining sales (via increased revenue per device).

As I have said before (and will say again), this strategy does work well if you are popular and make good products, but eventually you will push people past the point where they are willing to spend more and sales/revenue will fall off a cliff.

They got me at my limit. I feel like two categories of buyers.

1. Uneducated shoppers who want the laptop to match their phone. It could have any specs or no specs at all as long as they can finance it for less than their phone bill. They tend to buy a new computer when their old one is nonfunctioning or breaks. If they have excess cash, they'll get a base MBPro for the prestige factor.

2. Smart buyers. Myself included, studies the specs to make sure it will last long enough if we decide not to upgrade. Waits for the deal, and sells their well maintained prior Apple laptop to defray the cost.

Smart buyers generally bought the M1 14" or 16" to meet our needs. I don't think we're going to have much interest in the M3 at these prices and configurations. Apple lost their target market this round, I think. The value factor is just so far gone. Holding onto my M1 Pro. Fantastic device. I used to upgrade every other iteration. Now I think if this one died, do I really need a Mac anymore?

 

[barracuda156]

And yet still M1 macbook 8gb bought at 800$ kick ass of everything in photoshop, capture one... People buy smart and what you truely need

As it turns out, ads about 8 GB being sufficient are far from reality. They are not sufficient. (And no, I am not running 3 VM simultaneously, just compiled flang-17.)

On a side note, it is laughable that my PowerMac from 2005 has twice as much RAM as MBA M1 from 2020.

 

[anshuvorty]

On a side note, it is laughable that my PowerMac from 2005 has twice as much RAM as MBA M1 from 2020.

Yeah, but how much FASTER is the RAM on the M1 MBA compared to your aging PowerMac from 2005? Sometimes quality >>>>> quantity, as in this case.

 

[hagjohn]

Early M3 Max benchmarks.

https://forums.macrumors.com/threads/m3-max-chip-around-as-fast-as-m2-ultra-in-early-benchmark-results.2409580/

Don't some of you feel really foolish now?

 

[brgjoe]

That's the Max chip. It looks like that's the direction Apple wants us to go with the pricing. I mean, it doesn't look like there is a hella difference between the M2 Pro and the M3 pro chips. If anything, looks like the M2 Pros beat the M3 Pros in certain areas. But yeah, the M3 Max chip is definitely powerful.

But I am going to wait until real world test results happen on the actual machines before I make my purchasing decision.

 

[radellaf]

Well, then you are going to miss out on gaming. There is no need to buy a high spec MBP...

That's just part of what we've always had to put up with to avoid Windows. Kinda been the curse of all the computers I've owned, including a TRS-80 CoCo vs the ubiquitous C64.
That said, my old iMac has done OK for the games I play (mostly Blizzard's), and the M3 should, at least for now, let me max out the graphics.
No doubt a gaming PC, or even ordinary PC with a sweet-spot-value GPU would beat it out... but I can't afford both.

What I really can't figure out is whether 24GB RAM would be better than 16GB. Every single RAM video talks about Photo and video editing, neither of which I expect to be doing beyond iMovie and a few layers in Photoshop or Affinity. I'm inclined to spring for it. It's the 2TB SSD price that really kills me - but it's just not worth trying to shrink down or offload my boot drive contents from a 2015 2TB fusion drive computer to fit an (oddly) smaller internal drive on a new one.

If the iMac M3 had an option for the M3 Pro with almost double the GPU cores, I'd jump on it, though. I probably should upgrade every 3 years instead of every 7... but I don't. Some of that is left over dread from how it can be, but rarely is any more, trying to move over apps and data to a new computer. As long as migration assistant does its magic, you're golden. But, sometimes it doesn't, and the days before it existed are all too burned into my memory. Also, 7 years used to seem like forever. Now, 3 years seems like yesterday, and one year cycles are like "I _just_ did that".

 

[VisceralRealist]

The M3 Pro really got shafted. I just realized in the event they said it was 20% faster than M1 Pro, which is the same thing they said about the M2 Pro. Meaning that the M3 Pro is 0% faster than the M2 Pro. It's the same. To call it "Scary Fast" and then not improve the CPU performance of the Pro chip at all (in fact, to deliberately limit it by removing two performance cores) is pretty shady. Just another upselling strategy.