dnbaker
commented
1 year ago Thanks for the file! I'll give it a shot.
How many sequences are in the file? Once you go past 100k, then the default all-pairs distance calculation gets pretty expensive because of the O(n^2) comparisons because of the n-choose-2 pairwise comparisons that are performed.
It's true - if you're performing sketch in --parse-by-seq mode, it doesn't have anywhere to put the sketches by default.
You can have it write them to disk by adding -o sketch_file.bin. With 65k sequences, though, I would expect it to run faster.
I notice that --cmpout is a folder path; it's possible that it's stuck trying to write to a directory.
I'll update you with progress. And I appreciate the smaller xz file, it's what I usually share with.
matnguyen
The following Fasta file never completes (has been running for 5 days). The following command was used:
dashing2_binaries/linux/v2.1.16/dashing2_savx2 sketch --cmpout ../ab_data/ -k 7 --parse-by-seq -p 8 ../ab_data/ab.faab.fa.xz.gz
If trying to only do sketching, it completes but does not output any sketch file
dashing2_binaries/linux/v2.1.16/dashing2_savx2 sketch -k 7 --parse-by-seq -p 8 ../ab_data/ab.fa --cache --prefix ../ab_data/Sorry about the weird file format, xz is better for compression but Github won't let me upload them.