Memory issue when processing large protein. Is my swap not being leveraged

I am running predictions for 2500 residue protein on a system with 80GB RAM, 8TB diskspace, and 1TB swap running Alphafold v2.2 via docker. The process fails with the message "RuntimeError: Resource exhausted: Out of memory while trying to allocate 51935261664 bytes." The ram was upgraded to 128GB, but the job dies with the same error.

docker stats: shows the MEM LIMIT is not including available swap. I am likely mistaken, but I thought it would... Is my swap not being leveraged by the container? How can I modify the docker image, such that when launched the container leverages swap (which I previously thought was available by default)? If I am off target, any advice would be appreciated...

Note: I have another system with 512GB RAM that no problem completing the calculations for the same protein.