Verbosity for progress in search

[luciaurcelay]

Is there any way to see the realtime progress of search in gpu mode by adding some flag like verbosity? I am dealing with a very big database (195M sequences split in chunks) and I would like to estimate how much time it's going to take. Thanks!

Currently used command:

mmseqs search "$QUERY_DB" "$TARGET_DB" "$OUTPUT" "$TMP" \
    --gpu 1 \
    --threads 16 \
    --max-seqs 100

[milot-mirdita]

All the individual MMseqs2 modules should each show a progress bar. Could you paste the current terminal output here?

Also just a warning, but --max-seqs affects sensitivity heavily. This parameter specifies the number of hits passed from the prefilter to the alignment stage, not the number of hits presented to the user at the end.

[luciaurcelay]

Sure, here is the current terminal output (the job is currently running)

ungappedprefilter /home/databases/chunk_00_gpu /home/databases/chunk_00_gpu /scratch/b114cb
24/mmseqs_tmp/2850147/0/tmp_0_0/11359787563462428536/pref_0 --sub-mat 'aa:blosum62.out,nucl:nucleot
ide.out' -c 0 -e 0.001 --cov-mode 0 --comp-bias-corr 1 --comp-bias-corr-scale 1 --min-ungapped-scor
e 15 --max-seqs 150 --db-load-mode 0 --gpu 1 --gpu-server 0 --gpu-server-wait-timeout 600 --prefilt
er-mode 1 --threads 16 --compressed 0 -v 3

[========

Progress bar shows but there is not an ending character to see how long the bar is, also it would be very helpful to print the number of steps remaining in the progress bar.

Regarding --max-seqs, I run two search (gpu) experiments using 20 and 50 max sequences, and the computation time did not change significantly, here are the terminal outputs:

ungappedprefilter /home/databases/chunk_00_gpu /home/databases/chunk_00_gpu /scratch/b114cb24/mmseqs_tmp/2849156/0/tmp_0_0/14998127590645166376/pref_0 --sub-mat 'aa:blosum62.out,nucl:nucleotide.out' -c 0 -e 0.001 --cov-mode 0 --comp-bias-corr 1 --comp-bias-corr-scale 1 --min-ungapped-score 15 --max-seqs 20 --db-load-mode 0 --gpu 1 --gpu-server 0 --gpu-server-wait-timeout 600 --prefilter-mode 1 --threads 16 --compressed 0 -v 3 

[=================================================================] 3.00M 10h 57m 8s 962ms

ungappedprefilter /home/databases/chunk_00_gpu /home/databases/chunk_00_gpu /scratch/b114cb24/mmseqs_tmp/2849160/0/tmp_0_0/10251076232596653917/pref_0 --sub-mat 'aa:blosum62.out,nucl:nucleotide.out' -c 0 -e 0.001 --cov-mode 0 --comp-bias-corr 1 --comp-bias-corr-scale 1 --min-ungapped-score 15 --max-seqs 50 --db-load-mode 0 --gpu 1 --gpu-server 0 --gpu-server-wait-timeout 600 --prefilter-mode 1 --threads 16 --compressed 0 -v 3 

[=================================================================] 3.00M 10h 52m 43s 693ms

Does this make sense?

Thanks!