memory requirement

[lpantano]

Hi!

thanks a lot for work on this, it is very useful.

I was trying to work on the example data you have (thanks for doing that example).

I wanted to know how much memory it needs for that data, because I keep getting my job killed by memory issues and I went up to 64G. I want to make sure it is not a problem at my end (job actually not having that memory allocated) or if it really needs more than that.

Thanks a bunch!

[JonathanShor]

We've run it on a personal laptop with 16gb RAM without problem.

You could try setting a lower n_top_var_genes value, tho you should really be more than fine with 64gb with the default settings.

[lpantano]

Thanks for the information. I am going to try my personal laptop as well, maybe something is wrong with the cluster or python package installation.

Thanks a lot!

On May 16, 2018, at 10:54 AM, Jonathan Shor notifications@github.com wrote:

We've run it on a personal laptop with 16gb RAM without problem.

You could try setting a lower n_top_var_genes value, tho you should really be more than fine with 64gb with the default settings.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/JonathanShor/DoubletDetection/issues/94#issuecomment-389548778, or mute the thread https://github.com/notifications/unsubscribe-auth/ABi_HJL_Tmtj1LDfEmdFUc4YFjiJibk6ks5tzD26gaJpZM4UAK42.

[adamgayoso]

@lpantano Were you able to figure out the problem?

[lpantano]

Hi, Yes, at the end I tried in my computer and it works, but for some reason there is a memory leak or something when running in the linux cluster (I have a macosx). I created a conda environment to install this package, similar to what I did in my machine.

Not sure, if the problem is some package version or the linux machines or what.

thanks!