The best route is to run Geekbench5* on a Google cloud instance of your choice of machine, so that exactly the same machine configuration is available to anyone wanting to replicate your winning entry. Quoting
the rules:
Each program must run in less than 70'000/T hours on a machine using at most 10GB RAM and 100GB HDD for temporary files, where T is the machine's Geekbench5 score. No GPU usage.
This is because:
1) Every prize winner since
Starlit crashes on
my Ryzen machine. This is due to a bug in the LLVM compiler that they apparently can't fix.
2) Lengths of binaries do change slightly with different target CPU architectures -- and this can affect the payout if not the award decision.
However, if you would prefer to run on
my Intel laptop, I've recently upgraded it. But be forewarned: If it does not execute on that hardware, you'll probably have a delay in the award until either globe-trotting-in-retirement Matt Mahoney can get around to doing the test, or until you are able to get it to run on a Google cloud instance.
Moreover, it is possible there will be a change in the length of the binary that affects the judgement.
* We should probably update to Geekbench6 in the rules. I'll look into that. Until then, you'll have to run the legacy Geekbench5: