|
|
|
|
|
by Karellen
1203 days ago
|
|
Wait, the benchmark to find the 35th fibonacci number took 1.077s for the Zig VM vs 1.657s for the Rust VM? I realise that these VMs are going to be totally idiomatic Zig/Rust, with the most straightforward implementation possible, and little-to-no performance tuning, but even so - that's gotta be a typo, right? Or it's actually finding `fib(350)`? Or each "run" is actually finding `fib(35)` 100 (1000?) times? |
|
(I eyeballed the assembly to make sure the native implementation was recursive, but that's the extent of my rigor - consider these napkin numbers.)