admixture mem error, but not out of RAM

Question

[dec986]

I'm attempting to run admixture:

admixture tmp_bcftools_norm.bed 27 --supervised

the entire directory contains the files:

-rw-r--r--. 1 703404669 domain users  48G May 27 16:21 tmp_bcftools_norm.bed
-rw-r--r--. 1 703404669 domain users 2.3G Jun  1 09:43 tmp_bcftools_norm.bim
-rw-r--r--. 1 703404669 domain users  47K May 27 16:20 tmp_bcftools_norm.fam
-rw-r--r--. 1 703404669 domain users  910 May 27 16:21 tmp_bcftools_norm.log
-rw-r--r--. 1 703404669 domain users 2.0G May 27 12:45 tmp_bcftools_norm.map
-rw-r--r--. 1 703404669 domain users 793G May 27 13:17 tmp_bcftools_norm.ped
-rw-r--r--. 1 703404669 domain users  14K May 27 13:17 tmp_bcftools_norm.pop
-rw-r--r--. 1 703404669 domain users  15G May 27 12:30 tmp_bcftools_norm.vcf.gz
-rw-r--r--. 1 703404669 domain users 2.3G May 27 16:20 tmp_bcftools_norm_with_XYM.bim

but this results in the following message to STDERR:

terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc

normally, I would think that this error message means that RAM is running out. However, only 17% of RAM is used.

I had another error from admixture ADMIXTURE: "Invalid chromosome code. Use integers!" so I removed X, Y, and MT chromosomes from the bim file. I don't know if the discrepancy in chromosomes is what's causing this error, but I'm asking because it took a long time to regenerate all of those files, and I don't want to regenerate the files if the discrepancy in chromosomes isn't the problem.

[ATpoint]

However, only 17% of RAM is used.

How did you check that?

[dec986]

I checked via top -d1

[ATpoint]

Is this a local machine you are running on or a cluster/HPC/server?

[dec986]

this is a server. There are no other significant processes running