BUG: 1.6.5 crash on -madd LG4M and LG4X

[giddy....@gmail.com]

Hi guys,

FYI:

IQ-TREE multicore version 1.6.5 for Linux 64-bit built May  8 2018
Developed by Bui Quang Minh, Nguyen Lam Tung, Olga Chernomor,
Heiko Schmidt, Dominik Schrempf, Michael Woodhams.

Host:    ada (AVX2, FMA3, 503 GB RAM)
Command: iqtree -nt 6 -wsr -wsl -t BIONJ -mset WAG,LG,JTT -madd LG4M -pre H_mod -s H_mft_lin.fasta -redo
Seed:    39310 (Using SPRNG - Scalable Parallel Random Number Generator)
Time:    Sun Jun  3 00:54:13 2018
Kernel:  AVX+FMA - 6 threads (48 CPU cores detected)

.......

 72  JTT+F+R4      15653.513    218 31743.025    33334.425    32534.629
 73  JTT+F+R5      15630.945    220 31701.890    33378.442    32500.757
 74  JTT+F+R6      15628.662    222 31701.324    33469.395    32507.453
Model LG4M is alias for MIX{LG4M1,LG4M2,LG4M3,LG4M4}*G4
ERROR: modelmarkov.cpp:274: void ModelMarkov::init_state_freq(StateFreqType): Assertion `freq_type != FREQ_UNKNOWN' failed.
ERROR: STACK TRACE FOR DEBUGGING:
ERROR:
ERROR: *** IQ-TREE CRASHES WITH SIGNAL ABORTED
ERROR: *** For bug report please send to developers:

Input+log saved, awaiting your request.

Thanks for a great program!
Cheers,
Giddy

[Minh Bui]

Hi Giddy,

Thank you for this report. I could reproduce this crash with some example alignment here (so need to send your data files). It occurs whenever one uses LG4M or LG4X model. Interestingly there was no crash with v1.6.4. So we’ll prioritise to fix this and other issues ASAP.

Cheers,

Minh

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

[Karen Meusemann]

Hey Minh,

can you let me know when this is fixed pls?

I would urgently need it (since working with LG4M/X - (meanwhile I go on then probably with version 1.6.2 or 1.6.3...)

Best Karen

[Minh Bui]

Hi Karen and others, 

This issue is fixed in a local version here (just one line of code). I need to look at other issues as well. Thus, the next version will be released (hopefully early) next week.

Cheers

Minh