Hi!

I am using DESeq2 on an in-house server which is shared among a handful of users and has not a cueing system. I want to set a limit on the number of nodes DESeq2 uses, for which I used MulticoreParams as follows:

parallelParams <- MulticoreParam(workers = 4,progressbar = TRUE, log = TRUE)
register(parallelParams,default=TRUE)
parallelParams

Which prints the following:

class: MulticoreParam
  bpisup: FALSE; bpnworkers: 4; bptasks: 2147483647; bpjobname: BPJOB
  bplog: TRUE; bpthreshold: INFO; bpstopOnError: TRUE
  bpRNGseed: ; bptimeout: 2592000; bpprogressbar: TRUE
  bpexportglobals: TRUE; bpforceGC: TRUE
  bplogdir: NA
  bpresultdir: NA
  cluster type: FORK

However, when I run DESeq2 and inspect the CPU work using htop I see that it is using all of the available resources, not only the 4 nodes I asked for.

I have also tried the following mentioned on the DESeq2 vignette:

dds<-DESeq(dds,parallel=TRUE,BPPARAM=parallelParams)

but the result is the same.

I am running my code inside a singularity container pulled from docker biocontainers (tag 1.34.0--r41h399db7b_0). Running sessionInfo() inside the container prints:

> sessionInfo()
R version 4.1.1 (2021-08-10)
Platform: x86_64-conda-linux-gnu (64-bit)
Running under: Debian GNU/Linux 10 (buster)

Matrix products: default
BLAS/LAPACK: /usr/local/lib/libopenblasp-r0.3.18.so

locale:
 [1] LC_CTYPE=C.UTF-8    LC_NUMERIC=C        LC_TIME=C          
 [4] LC_COLLATE=C.UTF-8  LC_MONETARY=C       LC_MESSAGES=C.UTF-8
 [7] LC_PAPER=C          LC_NAME=C           LC_ADDRESS=C       
[10] LC_TELEPHONE=C      LC_MEASUREMENT=C    LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
[1] compiler_4.1.1

Thank you in advance,

Sebastian

It’s not DESeq2/BiocParallel causing this but the multithreaded BLAS library. It is indeed an annoyance and I have always wondered why multithreading with all cores is a default in R. Anyway, you can use the RhpcBLASctl package to disable that multithreading. I always run this on top of any script:

RhpcBLASctl::omp_set_num_threads(1)
RhpcBLASctl::blas_set_num_threads(1)