Glimpse2_phase (2.0.0) segfault - but only with large number of variants in reference panel.

[bgulko]

Needless to say, thank you for making V2 available, and on GitHub!

I am also running grch38 whole-chromosome (without chunk/ligation) and find the smaller chromosomes complete, but the larger ones quietly segfault. The boundary seems to be at around 650K variants in the reference panel.

G2 succeeds consistently when there are <613K variants and fails consistently then there are >667K variants (numerous trials, with same input BAM and sub-sampled reference VCF s).

There seems plenty available disk storage and memory - large successful runs took at most a few minutes and utilized <64GB memory.

My apologies in advance if a fix has been applied. If so, please let me know, and I'll move to the dev branch at 2.0.1.

Environment specs include:
Using downloaded binary (for Glimpse2_phase)
UBUNTU 22.04 LTS
AVX512 available,
256GB RAM.