No one has posted a MacOS benchmark here, but apparenly on arm64 it is truly much faster than the alternatives, I don't understand why but maybe the library is broken for amd64, or my code is broken, but then it would be weird.
No one has posted a MacOS benchmark here, but apparenly on arm64 it is truly much faster than the alternatives, I don't understand why but maybe the library is broken for amd64, or my code is broken, but then it would be weird.
did you check simd support on amd64?