[erwincoumans]

That is an impressive performance claim, almost 10 times faster than flatbuffers.

Where is the flatbuffers native C (or C++) implementation of the benchmark? Are memory allocations avoided/excluded in the benchmark?

[liuliu]

Seems the benchmark does read into native structs in both Flatbuffers and this in Go. I am confused by the performance claim too. Seems very similar design v.s. fbs, would be surprised if it is from format differences rather than codegen / Go implementation inefficiencies.