The effort and long time it took Go to get to something that 3-6x times slower than other, better languages should be an important reminder to everyone assuming it belongs to the same weight class as Rust, C# or Java.
And yet, this is something that can be trivially done in C#, C++ and Rust (albeit C# has the best UX with crossplat SIMD API introduced in .NET 7, with C++ close second with its own take on this being in preview). Java OTOH manages to be in the same category by having extremely advanced JIT that allows it to have comparable codegen quality even though it lacks comparable SIMD API for now (Panama vectors are problematic currently), so benchmarks implementations using it are forced to do SWAR.
My main gripe is of course an extremely common misconception about Go's speed which it just does not have the moment you write anything sufficiently advanced or want to express a particular problem in a terser way than writing thousands of open coded loops.