performance of writer turned out to be extremely slow (writing <100MB file with validation + several minutes which was previously maybe 30s), observed that chunking is very finely granularized #276
checking the execution flow turned out that 99% of the time in the NXS writer is spent in not doing I/O, so
yes auto chunker choices are not optimal but this is not the culprit, let's check with Rubels improvements
[x] Was found out to be substantially dependant on verification, fixes in several places have been implemented but more work would be beneficial to make parsing faster in the future
checking the execution flow turned out that 99% of the time in the NXS writer is spent in not doing I/O, so yes auto chunker choices are not optimal but this is not the culprit, let's check with Rubels improvements