Entering edit mode
2.2 years ago
SushiRoll
▴
140
Hi all!
I'm intending to run snippy followed by snippy-core (basically snippy-multi) on a lot of NGS data (around 1000 isolates). I have read that these processes can crash if the RAM is not enough, right now I'm running a VM with 8Gb so I guess I will have to ask for an upgrade. The thing is that the guys administering the resources are a little picky so I would like to know if there is a way of calculating the needed RAM for my runs or if you could give me an estimate. I don't mind if it takes a while as long as it doesn't crash halfway.
As usual, thanks a lot for any suggestion/help!
https://github.com/tseemann/snippy
You may need to do this yourself since it could be dependent on the data. Try running a few samples and see how much RAM you need for each. I assume your samples are similar sized genomes? If you intend to run 1000 samples in some form of series of jobs then you will need to see how many you could start at the same time based on hardware you have access to.
Example on the page @Pierre linked used 16 cores for only ~4 min to complete one job, so it may not be so bad. We don't know how much RAM was used though.
Hi Pierre and GenoMax,
Thanks both for your answers. I have checked Torsten's Github page but as GenoMax said, it only states the number of cores used. The genomes should indeed be similar sized. In my case each run takes around 7 min. but I don't have 16 cores. I guess I can run several jobs increasing the input until it crashes. Theoretically I can have access to an important amount of resources but this needs to be justified, if I ask for something that is just overkill, they will just turn me down.