--make-grm-part not correctly splitting up GRM

Hello, 
I'm trying to construct a GRM using a large dataset (>10k individuals, >140million sites). I am splitting up the process into 200 parts using the --make-grm-part option. 

My code is:


for i in {1..200}; do gcta64 --bfile INPUT --make-grm-part 200 $i --thread-num 10 --out testOUT; done

Parts 1 through 196 seem to finish just fine. But parts 197 through 200 result in segfaults. It looks like, starting with part 197, the program is trying to construct a GRM for people that do not exist in the dataset (see output below).  You can see that the program is trying to compute the GRM for individuals 10406-10432, although there are only 10418 individuals total in the dataset. The same issue occurs for parts 198,199, and 200. Any ideas what is going on here?

Options:

--bfile INPUT
--make-grm-part 200 197
--thread-num 10
--out testOUT

The program will be running on up to 10 threads.
Note: GRM is computed using the SNPs on the autosome.
Reading PLINK FAM file from [/path/to/INPUT]...
10418 individuals to be included from FAM file.
10418 individuals to be included. 0 males, 0 females, 10418 unknown.
Reading PLINK BIM file from [path/to/INPUT]...
146103976 SNPs to be included from BIM file(s).
Computing the genetic relationship matrix (GRM) v2 ...
Subset 197/200, no. subject 10406-10432
27 samples, 146103976 markers, 281313 GRM elements