Performance
The current README measurements describe WIPC as mostly framing overhead plus stream behavior:
- small encode paths in the millions of ops/s
- decode as a zero-copy view
- copy paths close to memcpy speed
- round-trip throughput limited more by Node stream backpressure than by framing itself
Run the repo benchmarks directly:
sh
npm run bench