Hello,
My model contains ~2 million variables, ~1.4 million equality constraints and ~1 million inequality constraints, which results in a GDX output file size of ~140 MB. This is quite limiting in my working environment where GAMS is called a lot of times for different analyses. Compressing the GDX file is an approach that I have implemented. I would like to know if there are approaches that I can implement to help me further reduce the GDX file size even before compression?
Thanks,
Zamry
Zamry,
compressed gdx files are usually quite small compared to storing the same information in other file formats.
You do not mention how you compress the GDX file, so maybe it is of interest to you that there is a command line parameter GDXCOMPRESS which will cause GDX files to be created in compressed format. You do not need to care about decompression. GAMS can read those compressed GDX files.
I hope this helps!
Fred
Hi Fred,
Thanks for suggesting gdxcompress, it works!
What I have been implementing is zip compress the gdx output with some csv inputs. What I’m observing now is that the same compressed file size is obtained with and without gdxcompress. I guess this is expected?
Yes, I would expect that a compressed GDX cannot be compressed much further with zip.