avail_mem not used for picard about chipseq

[ewels]

So basically, this code is no longer used for anything:

chipseq/main.nf

Lines 479 to 491 in 5f67d82

	if( task.memory == null ){
	log.warn "[Picard MarkDuplicates] Available memory not known - defaulting to 6GB ($prefix)"
	avail_mem = 6000
	} else {
	avail_mem = task.memory.toMega()
	if( avail_mem <= 0){
	avail_mem = 6000
	log.warn "[Picard MarkDuplicates] Available memory 0 - defaulting to 6GB ($prefix)"
	} else if( avail_mem < 250){
	avail_mem = 250
	log.warn "[Picard MarkDuplicates] Available memory under 250MB - defaulting to 250MB ($prefix)"
	}
	}

You can see that previously we ran picard with that as the memory argument (java -Xmx${avail_mem}m), but that was dropped when we switched to the bioconda installation: 4bbf815#diff-28b7bea57b04f15beeb4c1de0ee78382L470

It's no big deal - just that we may as well delete the avail_mem snippet if it's not needed, or if we can specify the available memory then perhaps thats better? Not sure if it actually helps performance though. I think the bioconda script does a bunch of that anyway?

[apeltzer]

Sorry, maybe I missed the discussion: Was there an issue using the one provided by picard in bioconda? Or what does this refer to?

[apeltzer]

Okay, I missed that part. The bioconda recipe just adds a custom wrapper that makes it possible to run the Java Jar file with the specified amount of memory (e.g. see this script here which passes the command line parameters to the executable JAR file: https://github.com/bioconda/bioconda-recipes/blob/master/recipes/picard/picard.sh )

Not sure about the performance gain here either - also I found that "limiting Java applications" in terms of CPU/Memory isn't working as expectedly all the time. Some do quite a bunch of threading that looks to some schedulers as surpassing their limits unfortunately...

[ewels]

x-ref nf-core/methylseq#45

Required code already present in the rnaseq pipeline, so should just be a copy & paste:

    script:
    if( !task.memory ){
        log.info "[Picard MarkDuplicates] Available memory not known - defaulting to 3GB. Specify process memory requirements to change this."
        avail_mem = 3
    } else {
        avail_mem = task.memory.toGiga()
    }
    """
    picard -Xmx${avail_mem}g MarkDuplicates \\

[drpatelh]

Fixed in #61