Real World M1 MacBook Pro Benchmark vs Real World question

Hi Folks,
I’ve been holding off buying a MacBook Pro because of the potentially updated versions on the horizon. My window for this is narrowing and I may have to just buy one of the current models. My initial skepticism is based on the max of 16gb of Ram. I do a lot of audio work, and on my intel machines, 16gb would not get it done.

When I run the benchmark software on my iMac (Late 2015, 3.2 GHz Quad-Core Intel Core i5, 32gb, 1T SSD), and compare it to the M1 MacBook Pro w/ 16gb, the M1 MacBook Pro scores 2x what the iMac does.

Is it fair to assume that the M1 will literally not only be faster than the iMac, but significantly so ?.

I"ve read a lot of about how fast these machines are, but in the real world crunching through large audio files in Logic, will the 16gb keep up ?

Thanks in advance

In my opinion (I have an M1 Air with 16GB and a 16" MBP with 64GB), there is no magic that makes 16GB enough on an M1 if it’s not enough on Intel. One of the things that I do with the MBP is run several large VMs that I use to simulate very high volume log data producers and consumers, and those VMs need a lot of RAM as buffers. For that use case 16GB isn’t enough. Your mileage may vary but “unified memory” isn’t a substitute for actual RAM if that’s what you need.

That being said, the M1 is fast and for everyday work the SSDs are so good that I don’t really notice the swapping that’s happening on my Air.

(Edited to add: I don’t do audio processing, so my experience may not match your reality :slight_smile: )

1 Like

Thanks for the input

1 Like

You could take advantage of the 14 day return window, buy a MBP and do a side-by-side comparison. Then let us all know :slightly_smiling_face:

I found the 16GB is fine for editing 4K video and recording and editing audio; I do videos around 90 mins and FCP has never once had any problems. I have also used Logic Pro with around 20 tracks and it was fine. This is on the MacBook Air.

The only time I’ve found the 16GB limited is dealing with large datasets using Python and Scala, then I’ve found it can easily run out of memory. I use AWS for this kind of work anyway, running on a cluster, as I can’t even fit most of the data I use on a home computer.


Thanks Rob. I guess I am trying to find out if things are improving with the M1 for this. I am wanting the M1 to be improving my workflow, speeding up my editing, processing rather than just getting through it. Logic says that 32gb works fine, so 16gb should be worse. The benchmark shows the M1 at 2x my iMAC, but there a number of posts suggesting that while the M1 is great, nothing is going to substitute for the lower ram. Very confusing :stuck_out_tongue:

My understanding is that if you do something that absolutely requires lots to be held in memory, with only virtual machines and huge datasets cropping up in posts that I’ve seen, you need lots of memory. There’s no getting around it.

If you do other things that can work well by utilising the SSD as a kind of cache, such as normal computing, or video editing or audio editing in apps that have been optimised for the M1 (e.g., Logic, FCP), then the M1 seemingly performs miracles and 16GB or even 8GB can suffice where on Intel it cannot. Plus it’s buttery smooth.


It depends on your workload. In workloads that only touch the CPU there is no difference, but for video and image based apps that are well optimised for apples Metal api on M1 there is quite a large memory saving compared to intel integrated GPUs.

On the intel iGPUs (when running macOS) the os would reserver a portion of your memory (2 to 4GB) that is only for the GPU. For apples that do a mixture of GPU and CPU compute on the same data set (video editors, image editors etc) they would end up with a copy of the image in the CPU addressable memory region and then need to make another copy over to the GPU addressable memory region.

On M1 CPU and GPU cores can share memory pointers just the same as 2 cpu cores can share memory there is not resolved portion of memory for the GPU so these pro apps can avoid having 2 copies of the image/video in memory.

Furthermore for GPU memory bottleneck applications if you have a 16GB M1 that means that the GPU has a lot more memory to play with than any laptop mac in the past (even top of the line 16" have never had this memory memory addressable for the GPU). When compared to the intel iGPU that limited GPU memory to 2 to 4 GB this is a massive difference.

There are some other changes to the memory subsystem such as inline memory compression support that do help a little bit but these are mostly there to improve bandwidth rather than increasing addressable capacity.

1 Like