IQ-TREE with the -fast option is s

[Roey Angel]

Hi,

I'm trying to obtain a tree from amplicon sequencing dataset with ~12,000 sequences and 600 variants.

According to the documentaion the -fast should approach FastTree in computation speed, yet while FastTree runs under 1-2 hrs on my machine, IQ-TREE has been claculating this tree for 5 days now and is still not done.

I ran the following command:

mpirun -np 96 iqtree -s Seqs_filteredalign -fast -nt AUTO

On a 96 thread server with 512 GB RAM.

Am I missing something?

Thanks in advance,

Roey

[Minh Bui]

Hi Roey,

Before looking futther, Just to be sure: with MPI version the executable is named iqtree-mpi (by default, except if you change the name). Can you please make sure that it is the MPI version of IQ-TREE? You can look at the first line of the log file, that prints the Copyright and also the type of executable. Just simply paste this line here.

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 view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/9e976c66-378b-4055-bd43-a97233261f6e%40googlegroups.com.

[Roey Angel]

hmm good point.

I did run iqtree and not iqtree-mpi however I do see all 96 cores active.

Do you think it's worthwhile killing the process at this stage?

Not sure where I got that from

On Monday, 13 January 2020 13:50:42 UTC+1, Minh Bui wrote:

Hi Roey,

Before looking futther, Just to be sure: with MPI version the executable is named iqtree-mpi (by default, except if you change the name). Can you please make sure that it is the MPI version of IQ-TREE? You can look at the first line of the log file, that prints the Copyright and also the type of executable. Just simply paste this line here.

Minh

On 13 Jan 2020, at 10:40 pm, Roey Angel <work.r...@gmail.com> wrote:

Hi,

I'm trying to obtain a tree from amplicon sequencing dataset with ~12,000 sequences and 600 variants.

According to the documentaion the -fast should approach FastTree in computation speed, yet while FastTree runs under 1-2 hrs on my machine, IQ-TREE has been claculating this tree for 5 days now and is still not done.

I ran the following command:

mpirun -np 96 iqtree -s Seqs_filteredalign -fast -nt AUTO

On a 96 thread server with 512 GB RAM.

Am I missing something?

Thanks in advance,

Roey

--
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 iqt...@googlegroups.com.

[Minh Bui]

The word “MPI” should appear on the first line, something like this:

IQ-TREE MPI multicore version 2.0-rc1 for Linux 64-bit built Jan  6 2020

If not, then please definitely kill your jobs. Right now they are acting as 96 separate IQ-TREE runs, which are overwriting output files of each other.

Minh

To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/0f6e29e5-5250-4917-8388-2f11e65197ba%40googlegroups.com.

[Roey Angel]

It took us some time because  we had to upgrade our server.

Now MPI works but immediately after the sequence composition test I get the following error from MPI:

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 12395 RUNNING AT anaerobe
=   EXIT CODE: 11
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault (signal 11)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions

According to MPI help this is a problem with IQTREE not with MPI.

Any idea where this could come from?

Thank again

Roey

On Tuesday, 14 January 2020 02:19:12 UTC+1, Minh Bui wrote:

The word “MPI” should appear on the first line, something like this:

IQ-TREE MPI multicore version 2.0-rc1 for Linux 64-bit built Jan  6 2020

If not, then please definitely kill your jobs. Right now they are acting as 96 separate IQ-TREE runs, which are overwriting output files of each other.

Minh

To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/0f6e29e5-5250-4917-8388-2f11e65197ba%40googlegroups.com.

[Minh Bui]

Hi Roey,

Sorry for the delay - I’m on annual leave.

Can you pls send to my personal email the log file and the data per further inspection?

Thanks,

Minh

To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/3cb6cd63-9207-4443-91e8-5cb7a1a50aa8%40googlegroups.com.