Rarefaction.single processors parameters remove

Hi Schloss,

Thanks for your nice working of mothur.

It is said that the speed of rarefaction.single has been increased.

However, when I calculate the rarefaction of the diversity index, it seems slower with the removing of processors parameters.

rarefaction.single(shared=otus.shared,calc=sobs-chao-shannon,groupmode=f,freq=100,processors=5)

Do you suggest that we go back to the old version for quicker calculation?

No, you should never use an older version of mothur. How long do the following commands take to run?

rarefaction.single(shared=otus.shared,calc=sobs-chao-shannon,groupmode=f,freq=100,processors=5)

and

rarefaction.single(shared=otus.shared,calc=sobs-chao-shannon,groupmode=f,freq=100,processors=1)

when using 5 processors, it took 8 minutes.
when using 1 processor, it took 40 minutes.
And the new version of mothur 1.40.5 is exactly the same as 1.30.2

Sorry, I’m confused… What are you expecting the result to be? The default is processors=1 and one would expect 1 processor to take 5 times longer than using 8, which is what you are seeing.

I mean in the new version, there is no “processor” parameters. Thus, the speed becomes slow.
So should I keep using older version (which has the processor) parameter?
At the beginning, I thought the new version would be faster. But it is not.
Sorry for unclear description.

In the rewrite of mothur’s threaded code for version 1.40.0, we removed the processors option from the rarefaction.single command. We have plans to address this in a later release. You can follow the progress here, https://github.com/mothur/mothur/issues/546.