Closed edef1c closed 2 years ago
The first c->chunk_size_min - CA_CHUNKER_WINDOW_SIZE bytes don't contribute to the result, so we don't need to hash them. This improves chunker throughput by about 16% (216 MiB/s -> 250MiB/s) in my benchmarks.
c->chunk_size_min - CA_CHUNKER_WINDOW_SIZE
The first
c->chunk_size_min - CA_CHUNKER_WINDOW_SIZEbytes don't contribute to the result, so we don't need to hash them. This improves chunker throughput by about 16% (216 MiB/s -> 250MiB/s) in my benchmarks.