MMseqs precomputing alignment too slow

[hanrthu]

Hi，I am using the recommended MMseqs2 alignment method from Colabfold pipeline. However, the script stucks for more than 24 hours for one chain (about 200 amino acids long). When I check the resources, I found that mmseqs has a high memory usage but a really low cpu usage. Is it normal? How can I speed up the process? Thank you!