[3dem] Problem with Relion multi-body refinement

[Eric Hanssen]

HI Takanori, 
The sampling was 0.9375 for most and a couple at half of that 0.46875
Regards
Eric


-----------------
Assoc. Prof Eric Hanssen

Head - Advanced Microscopy Facility 
Honorary Principal Fellow - Department of Biochemistry and Molecular Biology

President Australian Microscopy and Microanalysis Society

Bio21 Molecular Science and Biotechnology Institute
30 Flemington Road - The University of Melbourne - Victoria 3010 - Australia
email: ehanssen at unimelb.edu.au | Office: +61 3 83442449 | Microscope: +61 3 83442509 
Web: www.microscopy.unimelb.edu.au

   




-----Original Message-----
From: Takanori Nakane [mailto:tnakane at mrc-lmb.cam.ac.uk] 
Sent: Friday, 15 March 2019 4:07 PM
To: Eric Hanssen; De Biasio, Alfredo (Dr.)
Cc: 3dem at ncmir.ucsd.edu
Subject: Re: [3dem] Problem with Relion multi-body refinement

Hi,

What is the angular sampling at the crashing iteration?

Best regards,

Takanori Nakane

On 2019/03/15 0:28, Eric Hanssen wrote:
> Hi Takanori,
> I have had a similar issue on several datasets. I updated this morning as requested. it is still happening and at the same iteration. I have tried on 3 datasets (same protein and same box size, different particle numbers and resolution) and it is consistent, it keeps happening.
> Box size 344, 2 or 3 bodies, 4 GTX1080 with 12GB
> 
>    3: MPI_ERR_TRUNCATE: message truncated
>    3: MPI_ERR_TRUNCATE: message truncated
> in: /home/bio21em1/relion/src/mpi.cpp, line 307
> terminate called after throwing an instance of 'RelionError'
> [odin:202188] *** Process received signal ***
> [odin:202188] Signal: Aborted (6)
> [odin:202188] Signal code:  (-6)
> [odin:202188] [ 0] /lib64/libpthread.so.0[0x375c60f7e0]
> [odin:202188] [ 1] /lib64/libc.so.6(gsignal+0x35)[0x375be32495]
> [odin:202188] [ 2] /lib64/libc.so.6(abort+0x175)[0x375be33c75]
> [odin:202188] [ 3] /usr/lib64/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x12d)[0x3762abea7d]
> [odin:202188] [ 4] /usr/lib64/libstdc++.so.6[0x3762abcbd6]
> [odin:202188] [ 5] /usr/lib64/libstdc++.so.6[0x3762abcc03]
> [odin:202188] [ 6] /usr/lib64/libstdc++.so.6[0x3762abcd22]
> [odin:202188] [ 7] /usr/local/bin/relion_refine_mpi(_ZN7MpiNode16report_MPI_ERROREi+0x102)[0x48b052]
> [odin:202188] *** End of error message ***
> 
> Regards
> Eric
> 
> 
> -----------------
> Assoc. Prof Eric Hanssen
> 
> Head - Advanced Microscopy Facility
> Honorary Principal Fellow - Department of Biochemistry and Molecular Biology
> 
> President Australian Microscopy and Microanalysis Society
> 
> Bio21 Molecular Science and Biotechnology Institute
> 30 Flemington Road - The University of Melbourne - Victoria 3010 - Australia
> email: ehanssen at unimelb.edu.au | Office: +61 3 83442449 | Microscope: +61 3 83442509
> Web: www.microscopy.unimelb.edu.au
> 
>     
> 
> 
> 
> 
> -----Original Message-----
> From: 3dem [mailto:3dem-bounces at ncmir.ucsd.edu] On Behalf Of Takanori Nakane
> Sent: Friday, 15 March 2019 1:05 AM
> To: De Biasio, Alfredo (Dr.)
> Cc: 3dem at ncmir.ucsd.edu
> Subject: Re: [3dem] Problem with Relion multi-body refinement
> 
> Hi,
> 
> We recently fixed a bug in multi-body refinement.
> Could you please try the latest version (3.0.2) on our GitHub?
> 
> Best regards,
> 
> Takanori Nakane
> 
>> Dear all,
>>
>> I am running Multi-Body Refinement on a two-body complex and I get the
>> following error message after the 8th iteration:
>>
>>   3: Message truncated
>>    3: Message truncated, error stack:
>> PMPI_Wait(219).....: MPI_Wait(request=0x7ffea6199c88,
>> status=0x7ffea6199d10) failed
>> MPIR_Wait_impl(100): fail failed
>> do_cts(649)........: Message truncated; 409319840 bytes received but
>> buffer size is 409172384
>> srun: error: compute5: task 3: Segmentation fault (core dumped)
>>
>> If I continue the job from the 8th iteration, it gets to completion but
>> the final maps of the 2 bodies have completely different masks from the
>> initial (correct) masks and are highly anisotropic.
>>
>> Could you, please, help?
>>
>> Thank you,
>>
>> Alfredo De Biasio
>> LISCB, Leicester
>>
>>
>>
>> _______________________________________________
>> 3dem mailing list
>> 3dem at ncmir.ucsd.edu
>> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
>>
> 
> 
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
> 
>