1.9 TB is insufficient memory for whole genome alignment between two 5 GB genomes.

[altuffin]

After running minimap2 -ax asm5 --eqx with two fasta files that are both hifiasm assemblies scaffolded to the same reference via ragtag, minimap crashes with the following output:

[M::mm_idx_gen::25.290*1.51] collected minimizers
[M::mm_idx_gen::27.468*2.00] sorted minimizers
[M::main::27.469*2.00] loaded/built the index for 7 target sequence(s)
[M::mm_mapopt_update::28.725*1.96] mid_occ = 500
[M::mm_idx_stat] kmer size: 19; skip: 19; is_hpc: 0; #seq: 7
[M::mm_idx_stat::29.548*1.93] distinct minimizers: 73790276 (82.84% are singletons); average occurrences: 2.207; average spacing: 9.970; total length: 1623510972
[morecore] insufficient memory
/var/spool/slurm/job310283/slurm_script: line 14: 1752956 Aborted                 minimap2 -ax asm5 --eqx -t 8 ${ref} ${query} > $SCRATCH/${output}.sam

This occurs after 2.5 CPU hours with sjobacct showing 300 GB of memory used before it fails. The input ref and query are 5 and 5.1 GB in size, consisting of 7 chromosomes each. I'm trying to run syri, hence the --eqx tag. I initially tried this with more threads and less memory but kept getting the same issue.

Are there any options I should change to try and get this to work?