NewsPhotography

Benchmarking Performance: Lightroom on M1 vs Rosetta 2 vs Intel

Today Adobe publicly announced native support for Apple Silicon, which means you won’t have to emulate with Rosetta 2 to edit photos if you have a new M1-equipped Mac. But is there a noticeable improvement over emulation, or even over Intel-based machines? After testing, we have some answers.

The below test was run by DL Cade, former Editor of PetaPixel and current product testing specialist. I asked DL to take a new Apple MacBook Pro equipped with M1 as well as an Intel-based MacBook Pro and run Lightroom through a testing regimen that has become our “standard” at PetaPixel that tests performance on imports and exports from a Sony a7R IV and a PhaseOne 100-megapixel medium format.

The Test

DL ran 100 a7R IV photos and 150 PhaseOne images both through a timed import and export via three versions of the software: first on an ARM Macbook Pro using Lightroom with native support, next using that same ARM Macbook Pro running the Intel-emulated version of Lightroom, and lastly through an Intel-based fully-loaded MacBook Pro. It should be noted that the Intel-based Macbook Pro has twice the RAM of the M1 MacBook Pro.

Every benchmark is the average of three or more consecutive runs to adjust for any inconsistencies.

The Results

For imports, the ARM Mac version averaged 18 seconds for the 100 Sony a7R IV files, and 22 seconds for the 150 PhaseOne files. The emulated Intel version managed the same in 16 seconds and 22 seconds, while the Intel-based Mac did it in 16 seconds and 24 seconds.

Below is this data visually charted, with the ARM Mac at the top in red, the emulated version in orange, and the Intel version in yellow:

The next step involves timing how long each of the versions can export those same files in two different formats: 100% resolution JPEG in sRGB and 16-bit TIFF in AdobeRGB.

The ARM Version took 18 minutes and 41 seconds to export the a7R IV JPEGs and 20 minutes and 19 seconds to export the TIFFs. It took 31 minutes and 50 seconds to export the PhaseOne JPEGs and 38 minutes and 23 seconds to export the TIFFs.

The Intel-emulated version took 17 minutes and 45 seconds to export the a7R IV JPEGs and 21 minutes and 12 seconds to export the TIFFs. It took 34 minutes and 52 seconds to export the PhaseOne JPEGs and 45 minutes and 35 seconds to export the TIFFs.

The Intel machine took 18 minutes and 43 seconds to export the a7R IV JPEGs and 19 minutes and 36 seconds to export the TIFFs. It took 31 minutes and 23 seconds to export the JPEGs and 35 minutes and 42 seconds to export the TIFFs.

Below are all three results compared against one another with the ARM Mac at the top in red, the emulated version in orange, and the Intel version in yellow:

Conclusion

Comparing the ARM-optimized version to the Intel version running via emulation, we didn’t see any big performance improvements in our standard import/export benchmarks. Yes, native M1 Lightroom is faster, but it’s not by a hugely appreciable amount in most cases. There does seem to be a slight speed bump with some longer export tasks, but the most noticeable improvements are in the overall usability of the app: it runs smoother and is even lighter on the CPU, so the fans simply never have to turn on.

Comparing it to the Intel version running on an Intel Mac, it’s just so abundantly clear that the M1 chip is superior. Not only were the export and import times relatively comparable despite the Intel model having twice the RAM as the M1 model, the Intel MacBook Pro had its fans on full blast during every export run. Even a 20-second import could get the fans going. Meanwhile, the M1 MacBook Pro was whisper quiet the entire time.

So while there are some speed improvements on the M1 Macbook Pro when compared to emulation, most will find the biggest benefits in how efficiently the Macbook Pro with M1 runs Lightroom. The real speed tests for what Apple can do with its own silicon will come when they can devote the chip fully to power. Hopefully, we will see that in the next generation of iMacs.


Author: Jaron Schneider
Source: Petapixel

Related posts
AI & RoboticsNews

The show’s not over: 2024 sees big boost to AI investment

AI & RoboticsNews

AI on your smartphone? Hugging Face’s SmolLM2 brings powerful models to the palm of your hand

AI & RoboticsNews

Why multi-agent AI tackles complexities LLMs can’t

DefenseNews

US Army buys long-flying solar drones to watch over Pacific units

Sign up for our Newsletter and
stay informed!