Comparing CPUs for Development
Apple M3, M2, Threadripper and Ryzen
It is hard to come by development benchmarks for CPUs. It’s always gaming. In a Hackernews discussion, someone mentioned Geekbench 6 Clang Benchmark. For my future reference and for everyone interested, I put my research here.
“The Clang workload uses the Clang compiler to compile the Lua interpreter, a popular open-source language interpreter. It models the use case of developers building their code and the just-in-time compiling that general users can encounter on their systems”
Perhaps not the best compiler benchmark for developers, but we need to take what we’ve got.
For now this are kind of random numbers, I searched for a CPU and then tried to find the best performing system
| CPU | Cores | klines/sec | $ | speed/$ |
|---|---|---|---|---|
| 7995WX | 96 | 421,2 | 20000 | 0,021 |
| 7970X | 32 | 376,6 | 9000 | 0,042 |
| M2 Ultra | 24 | 233,9 | 9000 | 0,026 |
| 7950x | 16 | 230,3 | 2500 | 0,092 |
| 14900K | 24 | 215,3 | 2500 | 0,086 |
| M3 Max | 16 | 196,5 | 6000 | 0,033 |
It looks like the best tradeoff CPU if you love compilation speed and money, is a Ryzen 7950x (3-4x better than Apple Silicon). The worst is the 96-core Threadripper. If you have the money but don’t want to pay an absurd amount, you might get a Threadripper 7970X. The best laptop with nearly the speed of a desktop 7950x is a Mac Pro with a M3 Max.
