Inlägg "Hello, I would like to express some problems we ..." | Analys av data från storskaliga molekylärbiologiska experiment (BB2490)

Hello,

I would like to express some problems we faced regarding the project. In our project, we need to use UPPMAX and SLURM heavily. We need to run cutadapt, tophat and cufflinks commands and these commands are really time consuming, it takes hours or even days to process our 15 million reads. However, we wait in the SLURM queue for maybe 15-20 hours at best to get our job ran.

So in short, we make progress very slowly, since UPPMAX is involved, and of course the data is huge. When I talked to other groups, they said that they were experiencing the same problems. On the other hand, we also want to present reasonable results in the poster session.

So, is it possible for you to extend the project deadline until March 20th or 21st?

The same thing happened with our project as well. We want to do some experiments with feature extraction and feature selection with our peptide sequences. When it comes to 2mer or 3mer feature set, the experiments usually take days to complete, in addition the waiting time for the submitted job to be executed is also too looong. I don't know if we could cover all of our expected experiments or not with this slow progress. So deadline extension would be great.

I also have problem trying install package FSelector in UPPMAX (something related to JVM when install dependency packages), both from source and from interactive mode. Can anybody help me out of this?

If you get this error when you run java command:

[gokcen@kalkyl1 tophat]$ java
Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.

you may try:

module unload java ; module load java

It worked for me.

We are having similair problems. We only have one program that takes time to run but we have had a lot of problems getting things running. And our first runs have ran out of time so didn't complete.