Tweet
The bowtie2-build indexer lists several performance options. I am unsure about what some of these do, and about whether they can be used to increase the performance (speed up) just the index building or if they will help make the actual aligning process faster. Bowtie uses less than 5 GB of memory while aligning, I am happy to let Bowtie use more memory (up to about 16 GB) if it will make the alignment process faster.
bmax/bmaxdivn: These two are related to each other. For bmax is it says, "Default (in terms of the --bmaxdivn parameter) is --bmaxdivn 4." So the default is to set bmaxdivn to 4, and if I set bmaxdivn lower it will use more memory but make the alignment faster?
dcv: "A larger period yields less memory overhead, but may make suffix sorting slower, ... Default: 1024." So if I set this to 256 will it make my alignment faster?
offrate: "the indexer will mark every 2^<int> rows. Marking more rows makes reference-position lookups faster, but requires more memory to hold the annotations at runtime. The default is 5 (every 32nd row is marked;". I could set this to 1 or 2 to get a faster alignment?
ftabchars: "A larger <int> yields a larger lookup table but faster query times. The ftab has size 4^(<int>+1) bytes. The default setting is 10". I could try 12 or 14 to get a faster alignment?
bmax/bmaxdivn: These two are related to each other. For bmax is it says, "Default (in terms of the --bmaxdivn parameter) is --bmaxdivn 4." So the default is to set bmaxdivn to 4, and if I set bmaxdivn lower it will use more memory but make the alignment faster?
dcv: "A larger period yields less memory overhead, but may make suffix sorting slower, ... Default: 1024." So if I set this to 256 will it make my alignment faster?
offrate: "the indexer will mark every 2^<int> rows. Marking more rows makes reference-position lookups faster, but requires more memory to hold the annotations at runtime. The default is 5 (every 32nd row is marked;". I could set this to 1 or 2 to get a faster alignment?
ftabchars: "A larger <int> yields a larger lookup table but faster query times. The ftab has size 4^(<int>+1) bytes. The default setting is 10". I could try 12 or 14 to get a faster alignment?