Affy exprs multicore processing

0

Entering edit mode

Fiona • 0

@fiona-12453

Last seen 7.7 years ago

Edinburgh

Hi

I am trying to run this code on 149 affy array samples

boxplot(as.data.frame(log2(exprs(raw.data))))

Currently it only seems to be using 1 core out of my 16 core processor, is it possible to run this multithreaded on the same processor to utilise the rest of the cores?

Thanks

affy exprs multicore • 1.3k views

ADD COMMENT • link 7.7 years ago Fiona • 0

1

Entering edit mode

This should be fast; exprs() extracts an existing matrix, so you've got something like

> m = matrix(runif(30000 * 149), 30000)
> system.time(boxplot(as.data.frame(log2(m))))
   user  system elapsed 
  7.696   0.144   8.577

Most of the time is spent in drawing the plot, which cannot be parallelized. Maybe something else is going on, e.g., 'raw.data' is not what you (or maybe I) think it is?

ADD REPLY • link 7.7 years ago Martin Morgan 25k

0

Entering edit mode

Thank you for the reply.

The drawing of the plot must be my issue then (I estimate it will take over 45 minutes to run). Thank you for clarifying that the process cannot be parallelised.

ADD REPLY • link 7.7 years ago Fiona • 0

0

Entering edit mode

Maybe if you clarified, by editing your question, what 'raw.data' was, and perhaps also adding the output of the sessionInfo() command, it might be more apparent why this takes so long for you. You could also try to repeat my 'simulation'; how long does that take?

ADD REPLY • link 7.7 years ago Martin Morgan 25k

Login before adding your answer.