DReichLab / EIG

Eigen tools by Nick Patterson and Alkes Price lab
Other
180 stars 60 forks source link

Unable to allocate units #103

Closed dp-0102 closed 7 months ago

dp-0102 commented 7 months ago

I'm trying to get a subset of populations from the AADR 1240k dataset, however when I run the convertf command, it aborts and shows an error message that says:

fatalx: Unable to allocate 5052887274 unit(s) for item Aborted

This memory allocation error appears regardless of how many populations I list in my poplist file. I've provided the information of my par file below

genotypename: 1240.geno snpname: 1240.snp indivname: 1240.ind outputformat: EIGENSTRAT genotypeoutname: subset.geno snpoutname: subset.snp indivoutname: subset.ind poplistname: pops.txt

MatthewMah commented 7 months ago

The current software unfortunately allocates the complete matrix of genotype values in memory and loads the input file, so memory requirements depend on the input file and not just the desired output. We will likely have additional tools available in the next month or two to work around this memory requirement issue.