GLIMPSE2_phase using more CPU than expected

[alshai]

When I run GLIMPSE2_phase with 1 thread, I expect around 100% CPU usage, but instead I am seeing >900%. The HPC scheduler, which is expecting 1cpu worth of usage, sees this and is killing my job as a result. After digging in, the spike in CPU happens right just before the commands finishes, so my guess is that it's related to BCF writing and htslib.

Command:

/bin/GLIMPSE2_phase --baseq 18 --bam-file <BAM> --reference <BIN> --mapq 35 --output <OUT> --threads 1

Nextflow trace:

nextflow.trace/v2
realtime=750057
%cpu=924
cpu_model=Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
rchar=1009579922
wchar=15785321
syscr=28684
syscw=3461
read_bytes=0
write_bytes=0
%mem=2
vmem=1609940
rss=1403428
peak_vmem=2150172
peak_rss=1817836
vol_ctxt=117
inv_ctxt=2334

Using the docker image constructed from the provided Dockerfile (commit: d84ee2338a79c18061b499ef0e2820db4972451c).

Please let me know if any other information is required.