I would like to improve the speed of my code. It generates an output file
for the results in CSV format which may be very large (10-200MB). Each line
of the output file contains a record which is calculated and written
sequentially using WRITE #.
I would like to improve the speed of this process by improving the speed of
the File Output which seems to be the main bottleneck.
I thought about "writing to memory" and then dumping results in big chunks
at regular intervals could improve the speed but I am not sure how to
implement it to respect the CSV formatting that I am getting using the WRITE
# command.

Any suggestions, ideas?