alexg9010
commented
11 months ago HI @desmodus1984,
This issue is unrelated to methylKit and is something that you have to fix together with your HPC provider.
Based on googling your error it could be possible that the nodes on your cluster are using versions of OPENBLAS/BLAS libraries which are causing the crash. I found two threads with similar error messages they reported were able to fix the problem:
- https://groups.google.com/g/r-inla-discussion-group/c/8xctxzv4cW0
- https://bbs.archlinux.org/viewtopic.php?id=217539
Others have also reported that changing compiler flags fixed the problem:
Best, Alex
desmodus1984
Hi,
I am trying to run methylKit in a cluster. I was able to install methylKit but when I try loading it in R, I get the following error:
Loading required package: GenomicRanges Loading required package: stats4 Loading required package: BiocGenerics
Attaching package: ‘BiocGenerics’
caught illegal operation address 0x7fde4ea4cc8c, cause 'illegal operand'
Traceback: 1: strsplit(x, "\n[ \t\n]\n", perl = TRUE) 2: lapply(strsplit(x, "\n[ \t\n]\n", perl = TRUE), strsplit, "[ \t\n]", perl = TRUE) 3: strwrap(paste(same, collapse = ", "), indent = 4L, exdent = 4L) 4: .maskedMsg(sort(conflicts[[i]]), pkg = sQuote(pkg[i]), by = cpos[i] < lib.pos) 5: checkConflicts(package, pkgname, pkgpath, nogenerics, ns) 6: library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc, quietly = quietly) 7: .getRequiredPackages2(pkgInfo, quietly = quietly, lib.loc = c(lib.loc, .libPaths())) 8: library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc, quietly = quietly) 9: .getRequiredPackages2(pkgInfo, quietly = quietly, lib.loc = c(lib.loc, .libPaths())) 10: library(methylKit) An irrecoverable exception occurred. R is aborting now ... /scratch/slurmd/job11049733/slurm_script: line 16: 19054 Illegal instruction Rscript PCA.meth.R > SLURM_JOBID_NAME.out
Any suggestion on how to fix it to run on a compute node? With 50 samples it is using ~ 200 GB of Ram, and can at most run 4 cores. I will get soon 500 methylation samples and I will need to use a high-memory node, but as of now, I can only use the low-memory login nodes.
Thanks;