Raman calculation failed

[Sophie]

Hi all,

I am doing Raman calculations for a bisphenol molecule using nwchem 6.8. But I met a problem which I could not figure it out what the error would be.

It said:

=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 142223 RUNNING AT compute-10-29.localdomain
=   EXIT CODE: 9
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES

I tried to increase the memory but the same problem still occurred at the same stage.

Here is the output of the calculation:

https://drive.google.com/file/d/1gknKyzosiMARzpb1E_OtpLjSvsIwZiRT/view?usp=sharing

Thank you in advance for any help and your time.

Best,
Sophia

[Edoardo Aprà]

Files are not open for access

[sophia sagala]

Oops, sorry about that. 

Please try the following link for the output file:

https://drive.google.com/file/d/1gknKyzosiMARzpb1E_OtpLjSvsIwZiRT/view?usp=drivesdk

Thank you,

Sophie

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/04ca6b8d-39a6-4923-a8d0-d16bc2b8ca5fn%40googlegroups.com.

[Edoardo Aprà]

"EXIT CODE: 9" means that the job was terminated with signal 9, that
is the job was killed
There could be several reasons why the job was killed, either it ran
out of time or if asked more memory than the available memory.
Were you running under a batch queuing system (e.g. PBS, Slurm)?
You might want to ask the system administrators of the system to get
more information.
The only thing you might try to change from the input side is the memory line.

[Sophie]

Hi Edo,

Thanks for your explanation. Yes, I was running under a batch queuing system, and had also discussed this matter to the system administrators. It is due to reaching its memory limit.

Regarding changing the memory line in the input file, as you suggested, I had also tried that before. So I tried from memory 5000 mb to memory 9000 mb, and I still faced the same problem that the job got terminated. and when I checked the output file, it is stopped at the same stage, i.e. solvent accessible surface ...

Below is the link for the output file with memory 9000 mb:

https://drive.google.com/file/d/1Bf7O736qTdEwlnO_82VeFL-krPrKCuHW/view?usp=sharing

I wonder if I put the correct command for the memory line in the input side or not? Another question was what is the maximum memory that I can put in the input side?

Thank you,

Sophie

[Edoardo Aprà]

How many cores and how much memory do you have on the nodes of the cluster you are using?

[sophia sagala]

The cluster I am currently using has 192 GB of RAM and 36 cores per node. 

On Mon, Dec 14, 2020, 11:37 PM Edoardo Aprà <edoard...@gmail.com> wrote:

How many cores and how much memory do you have on the nodes of the cluster you are using?

--
You received this message because you are subscribed to the Google Groups "NWChem Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nwchem-forum...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/2dc89474-9b11-43a6-8804-0701807ebc5en%40googlegroups.com.

[Edoardo Aprà]

I would use 4 GB of memory or less.

192*0.75/36=4

[Sophie]

Thank you, Edo for your suggestion.

Could you please update me what 0.75 corresponds to, so I could understand better?

Best,

Sophie

[Edoardo Aprà]

75% of the available memory. Since the operating system and other processes might be using memory, it's better to put an upper limit of 75-80% to the memory requested by NWChem

[sophia sagala]

Hi Edo,

I tried again the Raman calculation by using 2GB and the calculation stopped again at the solvent accessible surface stage.

In the error file of the job id, it is written :

compute-4-10.72484Exhausted 1048576 MQ irecv request descriptors, which usually indicates a user program error or insufficient request descriptors (PSM2_MQ_RECVREQS_MAX=1048576)

I wonder if the problem was due to the memory again or due to the parameters for Raman calculation that I applied in the input file?

Meanwhile, I also have tried IR calculation (task dft freq) on this molecule under similar conditions (b3lyp-d3 / 6-311++G**), it worked fine without errors. I have also tried Raman calculation on this molecule gas phase by using standard default memory of NWChem, and I still faced a problem that the calculation stopped upon "Solving response equations for perturbing field".

Here is the output for the Raman calculation by placing 2GB memory limit:

https://drive.google.com/file/d/1Umrv0Ym4dWwzu9TsUf-OkT1_R44Uwvn2/view?usp=sharing

I really appreciate your kind help as I am still learning in this area.

Regards,

Sophie

--
You received this message because you are subscribed to a topic in the Google Groups "NWChem Forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/nwchem-forum/c7aWOSbfvs4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to nwchem-forum...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nwchem-forum/9de1d1ad-34d2-4449-8f41-f869e14f1a44n%40googlegroups.com.

[Edoardo Aprà]

Please have a look at

https://github.com/nwchemgit/nwchem/issues/284#issuecomment-743984830

Setting the env. variable

PSM2_MEMORY=large

seems the fix for this kind of PSM2/Omnipath problems