Add further parallelization by continuing immediately after querying trace cols in initial file now, instead of waiting to load the entire file before starting other threads.
↧