The algorithm is somewhat primitive, but seems to work quite well in practice. It starts off using the chunk size defined by chunkSize, and if that chunk takes less than 10 seconds, it doubles the size. If a chunk takes more than 30 seconds, it halves the chunk size. (Of course, looking out for minimums and maximums.)
This adds dynamic chunk sizing to upchunk.
There are three new options:
The algorithm is somewhat primitive, but seems to work quite well in practice. It starts off using the chunk size defined by
chunkSize
, and if that chunk takes less than 10 seconds, it doubles the size. If a chunk takes more than 30 seconds, it halves the chunk size. (Of course, looking out for minimums and maximums.)