cudamapper error using bam output

[RichardCorbett]

Hi folks, I'm trying cudamapper with this command: /opt/GenomeWorks-2021.02.2/bin/cudamapper PAG33026_pass_concat.fastq.gz ../hg19a.fa -B > cudamapper_2021_02_02_GM24385.bam

which outputs quite a lot of output to stdout/stderr, from which I have reported the unique lines below:

Initialized GenomeWorks logger with log level ERROR
-C / --target-indices-in-host-memory not set, using -Q / --query-indices-in-host-memory value: 10
-c / --target-indices-in-device-memory not set, using -q / --query-indices-in-device-memory value: 5
 Query file: PAG33026_pass_concat.fastq.gz, number of reads: 9983679
Target file: ../hg19a.fa, number of reads: 84
Programmatically looking for max cached memory
Using device memory cache of 24899308094 bytes
Device 0 took batch 1 out of 1790 batches in total
[E::sam_hrecs_update_hashes] Duplicate entry "5" in sam header
print_sam: could not add header value
[E::sam_hrecs_update_hashes] Duplicate entry "5" in sam header
print_sam: could not add header value
[E::sam_hrecs_update_hashes] Duplicate entry "5" in sam header
print_sam: could not add header value
...
[E::sam_hrecs_update_hashes] Duplicate entry "19" in sam header
print_sam: could not add header value
[E::sam_hrecs_update_hashes] Duplicate entry "22" in sam header
print_sam: could not add header value
[E::sam_hrecs_update_hashes] Duplicate entry "5" in sam header
print_sam: could not add header value
[E::bgzf_flush] File write failed (wrong size)
terminate called after throwing an instance of 'std::runtime_error'
  what():  ERROR, print_sam: could not write alignment
Aborted (core dumped)

In practise I think I get the "Duplicate entry "N" " error for each read that is processed.

[RichardCorbett]

I get the same warnings/errors and result when using SAM output (-S).

[RichardCorbett]

Running without -B or -S seems to work without error: /opt/GenomeWorks-2021.02.2/bin/cudamapper PAG33026_pass_concat.fastq.gz ../hg19a.fa -d 8