when I use GPU to accelarate colabfold_search , it slower than colabfold_search by using cpu #986
Comments
yank666
changed the title
|
My guess is that you specify The search script will actually set the parameter to the optimal value by default |
thanks for your reply, it did work , but it still cost about 10 min , when I set db-load-mode to be 2, Is that normal about one query cost 10 min when using GPU accelarate? |
|
In my experience it depends on how much RAM you have. Around 128GB I was able to search a query under a minute. With 64GB 12 minutes (same GPU) |
how many vCPU do you have ? i have enough RAM of 240 GB , but i only have about 28 vCPU |
|
|
thank you ， it did work , but it still cost about 10 min per query, colabfold_search had the gpu benchamrk? Or are there any performance base lines that can be referenced? |
|
Could you upload the full terminal output please? |
|
=
this is full terminal output createdb /root/result/out9/query.fas /root/result/out9/qdb --shuffle 0 Converting sequences ungappedprefilter /root/result/out9/qdb /ai/sequencedb/colabfold_db/uniref30_2302_db.idx /root/result/out9/tmp/5640223795513751843/pref_0 --sub-mat 'aa:blosum62.out,nucl:nucleotide.out' -c 0 -e 0.1 --cov-mode 0 --comp-bias-corr 1 --comp-bias-corr-scale 1 --min-ungapped-score 15 --max-seqs 10000 --db-load-mode 0 --gpu 1 --gpu-server 0 --gpu-server-wait-timeout 600 --prefilter-mode 1 --threads 64 --compressed 0 -v 3 Index version: 16 Index version: 16 Index version: 16 Index version: 16 subtractdbs /root/result/out9/tmp/5640223795513751843/pref_tmp_1 /root/result/out9/tmp/5640223795513751843/aln_0 /root/result/out9/tmp/5640223795513751843/pref_1 --threads 64 --e-profile 0.1 -e 0.1 --compressed 0 -v 3 Remove /root/result/out9/tmp/5640223795513751843/aln_0 ids from /root/result/out9/tmp/5640223795513751843/pref_tmp_1 Time for processing: 0h 0m 0s 1ms Index version: 16 Merging the results to /root/result/out9/tmp/5640223795513751843/aln_1 Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Index version: 16 Index version: 16 subtractdbs /root/result/out9/tmp/5640223795513751843/pref_tmp_2 /root/result/out9/tmp/5640223795513751843/aln_1 /root/result/out9/tmp/5640223795513751843/pref_2 --threads 64 --e-profile 0.1 -e 0.1 --compressed 0 -v 3 Remove /root/result/out9/tmp/5640223795513751843/aln_1 ids from /root/result/out9/tmp/5640223795513751843/pref_tmp_2 Time for processing: 0h 0m 0s 1ms Index version: 16 Merging the results to /root/result/out9/res Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Index version: 16 Index version: 16 Index version: 16 Index version: 16 Time for processing: 0h 0m 0s 2ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 3ms Time for processing: 0h 0m 0s 1ms ungappedprefilter /root/result/out9/prof_res /ai/sequencedb/colabfold_db/colabfold_envdb_202108_db.idx /root/result/out9/tmp3/13310131947556013529/pref_0 --sub-mat 'aa:blosum62.out,nucl:nucleotide.out' -c 0 -e 0.1 --cov-mode 0 --comp-bias-corr 1 --comp-bias-corr-scale 1 --min-ungapped-score 15 --max-seqs 10000 --db-load-mode 0 --gpu 1 --gpu-server 0 --gpu-server-wait-timeout 600 --prefilter-mode 1 --threads 64 --compressed 0 -v 3 Index version: 16 Index version: 16 Index version: 16 Index version: 16 subtractdbs /root/result/out9/tmp3/13310131947556013529/pref_tmp_1 /root/result/out9/tmp3/13310131947556013529/aln_0 /root/result/out9/tmp3/13310131947556013529/pref_1 --threads 64 --e-profile 0.1 -e 0.1 --compressed 0 -v 3 Remove /root/result/out9/tmp3/13310131947556013529/aln_0 ids from /root/result/out9/tmp3/13310131947556013529/pref_tmp_1 Time for processing: 0h 0m 0s 1ms Index version: 16 Merging the results to /root/result/out9/tmp3/13310131947556013529/aln_1 Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Index version: 16 Index version: 16 subtractdbs /root/result/out9/tmp3/13310131947556013529/pref_tmp_2 /root/result/out9/tmp3/13310131947556013529/aln_1 /root/result/out9/tmp3/13310131947556013529/pref_2 --threads 64 --e-profile 0.1 -e 0.1 --compressed 0 -v 3 Remove /root/result/out9/tmp3/13310131947556013529/aln_1 ids from /root/result/out9/tmp3/13310131947556013529/pref_tmp_2 Time for processing: 0h 0m 0s 1ms Index version: 16 Merging the results to /root/result/out9/res_env Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Index version: 16 Index version: 16 Index version: 16 Index version: 16 Time for processing: 0h 0m 0s 2ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 3ms Time for processing: 0h 0m 0s 1ms Merging the results to /root/result/out9/final.a3m Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms [=================================================================] 1 0s 8ms Time for processing: 0h 0m 0s 2ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms Time for processing: 0h 0m 0s 1ms |
|
If you build the databases with I guess this is something I should revisit in the code base. |
Uh oh!
There was an error while loading. Please reload this page.
One sequence query will take about one hour and --gpu-server 1 doesn't make sense
env：
mmseqs (support gpu):
MMseqs2 Version: b804fbe
here is my step:
log
The text was updated successfully, but these errors were encountered: