Workload benchmarks and cost comparisons

Ash1 · April 4, 2023

As a software engineer, my workload generally involves compiling huge chunks of code, in JS, python, C++, java etc... as well as tabbing between a ton of browser tabs and running multiple browsers in headless mode for testing purposes. For the purposes of this discussion, lets keep heavily threaded tasks like machine learning and encoding or rendering out of the comparison.
I have tried various different benchmarks and googling for hours to try and figure out the best option for me to improve my efficiency. My main question is this, for a relatively small budget, am I better of getting a Mac Mini (M2) for $599, which sounds really reasonable, or upgrading my existing platform to the latest AMD and getting a top of the line CPU etc..., which would possibly cost a bit more but would allow me to use my existing setup and decent upgrade for future proofing a rig that can possibly be used for gaming. On the other hand the efficiency and portability of the Mac Mini is quite desirable too.

A lot of developers talk of Mac's being better for development but it is hard for me to know this without buying all the parts and trying them out. Does anyone know of any reliable information that can help me make a good educated guess?

igormp · April 5, 2023

A Macbook for sure is a beast compared to other laptops, but for a desktop? Other than pricing, you can get better performance for your tasks with a 5700x or better, and get more ram for cheap. Only scenario where a Mac beats such a desktop is when you have tasks that are bottlenecked by memory bandwidth, which is usually not the case with code compiling.

This might give you an idea of what you're looking for:

https://www.phoronix.com/review/apple-m1-linux-perf/3

https://www.phoronix.com/review/apple-m1-linux-perf/5

Ash1 · April 5, 2023

9 hours ago, igormp said:

A Macbook for sure is a beast compared to other laptops, but for a desktop? Other than pricing, you can get better performance for your tasks with a 5700x or better, and get more ram for cheap. Only scenario where a Mac beats such a desktop is when you have tasks that are bottlenecked by memory bandwidth, which is usually not the case with code compiling.

This might give you an idea of what you're looking for:

https://www.phoronix.com/review/apple-m1-linux-perf/3

https://www.phoronix.com/review/apple-m1-linux-perf/5

Heey thanks for that, looks like an interesting set of results. I am slightly puzzled how the older M1 Asahi machine, with an older mother board, less cores and clock speed is beating the newer macOS version. Linux for the win of course, I know that for sure and use linux for development generally. Still wish there was an easier comparison between M2 Mac mini's and say a 7950X, would put my investments into better perspective.

igormp · April 6, 2023

3 hours ago, Ash1 said:

Heey thanks for that, looks like an interesting set of results. I am slightly puzzled how the older M1 Asahi machine, with an older mother board, less cores and clock speed is beating the newer macOS version. Linux for the win of course, I know that for sure and use linux for development generally. Still wish there was an easier comparison between M2 Mac mini's and say a 7950X, would put my investments into better perspective.

Should be mostly scheduler or filesystem-related stuff. Still, I wouldn't recommend using asahi as of today since it's not 100% stable, so I'd recommend only to use the MacOS results for comparison.