David @mleku - 1y
so, this is with the pubkeys encoded as raw binary in their intermediate stage in the encode/decode test, once again, a further boost, this time nearly 20 more ops in the same time as before ``` goos: linux goarch: amd64 pkg: mleku.net/nostrbench cpu: AMD Ryzen 5 PRO 4650G with Radeon Graphics BenchmarkBinaryEncoding/event2.MarshalJSON-12 14 72789590 ns/op BenchmarkBinaryEncoding/event2.EventToBinary-12 176 6811022 ns/op BenchmarkBinaryEncoding/easyjson.Marshal-12 64 18677195 ns/op BenchmarkBinaryEncoding/gob.Encode-12 19 61512279 ns/op BenchmarkBinaryEncoding/binary.Marshal-12 9 115865839 ns/op BenchmarkBinaryDecoding/event2.UnmarshalJSON-12 14 88401592 ns/op BenchmarkBinaryDecoding/event2.BinaryToEvent-12 100 12255334 ns/op BenchmarkBinaryDecoding/easyjson.Unmarshal-12 57 23192256 ns/op BenchmarkBinaryDecoding/gob.Decode-12 5 223011989 ns/op BenchmarkBinaryDecoding/binary.Unmarshal-12 127 8656310 ns/op ``` again no improvement in the decode, but i'll get to that after i've exorcised the strings from the event type #devstr #benchmarks #nostr
humbug... it now isn't decoding to the canonical form correctly, i bet it's going raw binaries better fix that first (i don't think this affects the binary/runtime data tho)
