Thanks, Michael. I've created a tiny project with those files here: https://github.com/cpmech/go-fast-math-experiments
The output of go test -run=XXX -bench=. -benchtime=10s (in the xmath package) is: BenchmarkPowStd-32 300000000 49.8 ns/op BenchmarkPowP-32 2000000000 8.85 ns/op BenchmarkPowFI-32 2000000000 6.31 ns/op BenchmarkPowStd10-32 100000000 237 ns/op BenchmarkPowP10-32 500000000 39.1 ns/op BenchmarkPowFI10-32 300000000 44.6 ns/op BenchmarkPowStd20-32 30000000 549 ns/op BenchmarkPowP20-32 200000000 97.4 ns/op BenchmarkPowFI20-32 200000000 93.8 ns/op BenchmarkPowStd50-32 10000000 1584 ns/op BenchmarkPowP50-32 50000000 304 ns/op BenchmarkPowFI50-32 50000000 249 ns/op BenchmarkPowStd100-32 5000000 3529 ns/op BenchmarkPowP100-32 20000000 692 ns/op BenchmarkPowFI100-32 30000000 522 ns/op BenchmarkPowStd200-32 2000000 9221 ns/op BenchmarkPowP200-32 10000000 1623 ns/op BenchmarkPowFI200-32 20000000 1124 ns/op Your table-based function (PowFI) wins overall with ~7x speedup vs mine (PowP) with ~5x speedup (compared to the std math which is non-integer-specific, so, not very fair... but OK as reference). Cheers. Dorival -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
