dimod icon indicating copy to clipboard operation
dimod copied to clipboard

Allow CQMs to be constructed in storage rather than memory

Open arcondello opened this issue 4 years ago • 1 comments

For very large CQMs, it would be helpful to have something like numpy.memmap. It would wrap a CQM file, as generated by CQM.to_file. Constraints and objectives could then be added directly to storage rather than being stored in memory.

arcondello avatar Jul 16 '21 19:07 arcondello

Changed the title because memory-mapped is not quite what we're talking about here.

arcondello avatar Jul 16 '21 19:07 arcondello