Hi Dieter,
Does projecting an I2 object with 30 degree spacing make sense?
Best,
Basil
> Am 01.11.2019 um 22:12 schrieb Dieter Blaas <[log in to unmask]>:
>
> Hi Takanori, thanks!
>
> xmipp_image_header reports: Dimensions : 40 x 1 x 8184 x 11520 ((N)Objects x (Z)Slices x (Y)Rows x (X)Columns)
> I used shrink=0
>
> and, as I wrote, picking worked with 2D-class averages and MPI=1
>
> I now tried: shrink=1, MPI=4, and a (non-downsampled) 3D volume as reference: same error!
>
> I also tried: shrink=0, MPI=1, and the same (non-downsampled) 3D volume as reference: it runs until:
>
> Projecting a 3D reference with I2 symmetry, using angular sampling rate of 30 degrees, i.e. in 0 directions ...
> + Automatically set the background diameter to -nan Angstrom
> + You can override this by providing --particle_diameter (in Angstroms)
> and then nothing happens!
>
> so, something seems to go wrong with the 3D-references and, independently, with MPI with my particular parameters/dataset....
>
> I am not sure what parameters I should use to try downsampling. Do I have to prepare the reference beforehand or can everything be done with suitable parameters in Autopicking?
>
> bw Dieter
>
> -----------------------------------------------------------------------
> Dieter Blaas,
> Max F. Perutz Laboratories
> Medical University of Vienna,
> Inst. Med. Biochem., Vienna Biocenter (VBC),
> Dr. Bohr Gasse 9/3,
> A-1030 Vienna, Austria,
> Tel: 0043 1 4277 61630,
> Fax: 0043 1 4277 9616,
> e-mail: [log in to unmask]
> ------------------------------------------------------------------------
>
> Am 01.11.2019 um 22:17 schrieb Takanori Nakane:
>> Hi,
>>
>> What is the micrograph size?
>>
>> Did you specify "shrink: 0"?
>> What happens if you down-sample your reference?
>>
>> Best regards,
>>
>> Takanori Nakane
>>
>> On 2019/11/01 20:15, Dieter Blaas wrote:
>>> Hi and thanks!
>>>
>>> image size and reference size are both 900x900x900. I tried 12 MPI and 4 MPI. I have 512 GB total memory and 128 cpus. 'free -h' reports:
>>>
>>> total used free shared buff/cache available
>>> Mem: 503G 32G 415G 8.4M 56G 469G
>>> Swap: 8.0G 7.9G 63M
>>>
>>> In the run with 4 MPI I then recorded 'free -h' each 10 seconds to a file. The last entry just after the process crashed is:
>>>
>>> Mem: 503G 32G 405G 8.5M 65G 469G
>>> Swap: 8.0G 7.9G 70M
>>>
>>> but it can be that it was recorded just before the run aborted or just after....
>>>
>>> bw Dieter
>>>
>>>
>>> ------------------------------------------------------------------------
>>> Dieter Blaas,
>>> Max F. Perutz Laboratories
>>> Medical University of Vienna,
>>> Inst. Med. Biochem., Vienna Biocenter (VBC),
>>> Dr. Bohr Gasse 9/3,
>>> A-1030 Vienna, Austria,
>>> Tel: 0043 1 4277 61630,
>>> Fax: 0043 1 4277 9616,
>>> e-mail: [log in to unmask]
>>> ------------------------------------------------------------------------
>>>
>>> Am 01.11.2019 um 20:44 schrieb Takanori Nakane:
>>>> Hi,
>>>>
>>>> What is the image size and the reference size?
>>>>
>>>> Aren't you running out of (CPU) memory?
>>>>
>>>> Best regards,
>>>>
>>>> Takanori Nakane
>>>>
>>>> On 2019/11/01 17:46, Dieter Blaas wrote:
>>>>> Hi,
>>>>>
>>>>> Running Autopick on a workstation without GPU I always receive the below error, regardless of the number of MPI processes I use:
>>>>>
>>>>> * MotionCorr/job010/Micrographs/FoilHole_12244404_Data_11211986_11211988_20191013_0752_fractions.mrc
>>>>> *Projecting a 3D reference with* I2 symmetry, using angular sampling rate of 30 degrees, i.e. in 0 directions ...
>>>>> -------------------------------------------------------
>>>>> *Primary job terminated normally, but 1 process returned**
>>>>> **a non-zero exit code. Per user-direction, the job has been aborted.**
>>>>> *-------------------------------------------------------
>>>>> And (in the other window):
>>>>>
>>>>> --------------------------------------------------------------------------
>>>>> *mpirun noticed that process rank 12 with PID 0 on node dorothy3 exited on signal 9 (Killed).**
>>>>> *--------------------------------------------------------------------------
>>>>>
>>>>> *when just using 1 CPU it runs apparently normally* but forever. The same occurs when using 2D references, so it seems to a general problem related to MPI
>>>>>
>>>>> Class2D runs without problems with e.g. 32 MPI processes and e.g. 2 threads.
>>>>>
>>>>> Did I do anything wrong during installation?
>>>>>
>>>>> Thanks, Dieter
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------
>>>>> Dieter Blaas,
>>>>> Max F. Perutz Laboratories
>>>>> Medical University of Vienna,
>>>>> Inst. Med. Biochem., Vienna Biocenter (VBC),
>>>>> Dr. Bohr Gasse 9/3,
>>>>> A-1030 Vienna, Austria,
>>>>> Tel: 0043 1 4277 61630,
>>>>> Fax: 0043 1 4277 9616,
>>>>> e-mail:[log in to unmask]
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>> Am 01.11.2019 um 08:42 schrieb Takanori Nakane:
>>>>>> Hi,
>>>>>>
>>>>>> You still should look at micrographs from all defocus ranges
>>>>>> when you optimise autopicking parameters (threshold etc).
>>>>>>
>>>>>> Best regards,
>>>>>>
>>>>>> Takanori Nakane
>>>>>>
>>>>>> On 2019/11/01 5:57, Dieter Blaas wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> does this mean that, if I already have a 3D map (from this very data set) and use it as a template for a second round of autopicking, I do not have to care about the below?
>>>>>>>
>>>>>>> bw Dieter
>>>>>>>
>>>>>>> ------------------------------------------------------------------------
>>>>>>> Dieter Blaas,
>>>>>>> Max F. Perutz Laboratories
>>>>>>> Medical University of Vienna,
>>>>>>> Inst. Med. Biochem., Vienna Biocenter (VBC),
>>>>>>> Dr. Bohr Gasse 9/3,
>>>>>>> A-1030 Vienna, Austria,
>>>>>>> Tel: 0043 1 4277 61630,
>>>>>>> Fax: 0043 1 4277 9616,
>>>>>>> e-mail: [log in to unmask]
>>>>>>> ------------------------------------------------------------------------
>>>>>>>
>>>>>>> Am 31.10.2019 um 17:22 schrieb Takanori Nakane:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> You are correct, if you perform Class2D on your manually
>>>>>>>> picked particles and generate templates for autopicking.
>>>>>>>> However, you should look at micrographs from all defocus ranges
>>>>>>>> when you optimise autopicking parameters (threshold etc).
>>>>>>>>
>>>>>>>> If you use your manually picked particles to train
>>>>>>>> other programs (e.g. neural-network based pickers),
>>>>>>>> you might need to train on low-defocus micrographs as well.
>>>>>>>> This depends on the program.
>>>>>>>>
>>>>>>>> Best regards,
>>>>>>>>
>>>>>>>> Takanori Nakane
>>>>>>>>
>>>>>>>>> Dear all,
>>>>>>>>>
>>>>>>>>> maybe this is a silly question but I thought I ask anyway.
>>>>>>>>> When I'm picking particles manually, does it matter if I pick particles
>>>>>>>>> accross all defocus values, or is it ok to stay on larger defocus values
>>>>>>>>> where particle are more identifiable?
>>>>>>>>> I was thinking that since we strongly low pass filter our 2D classes, it
>>>>>>>>> is ok to stick with larger defocus values. Am I correct?
>>>>>>>>>
>>>>>>>>> Thank you for your input.
>>>>>>>>> Vincent
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>>
>>>>>>>>> Vincent Chaptal, PhD
>>>>>>>>>
>>>>>>>>> MMSB -UMR5086
>>>>>>>>>
>>>>>>>>> Drug Resistance and Membrane Proteins Laboratory
>>>>>>>>>
>>>>>>>>> 7 passage du Vercors
>>>>>>>>>
>>>>>>>>> 69007 LYON
>>>>>>>>>
>>>>>>>>> FRANCE
>>>>>>>>>
>>>>>>>>> +33 4 37 65 29 01
>>>>>>>>>
>>>>>>>>> http://mmsb.cnrs.fr/en/
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ########################################################################
>>>>>>>>>
>>>>>>>>> To unsubscribe from the CCPEM list, click the following link:
>>>>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>>>>>>>>
>>>>>>>> ########################################################################
>>>>>>>>
>>>>>>>> To unsubscribe from the CCPEM list, click the following link:
>>>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>>>>>>
>>>>>>> ########################################################################
>>>>>>>
>>>>>>> To unsubscribe from the CCPEM list, click the following link:
>>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>> To unsubscribe from the CCPEM list, click the following link:
>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>>>>
>>>>
>>>
>>> ########################################################################
>>>
>>> To unsubscribe from the CCPEM list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>
>
> ########################################################################
>
> To unsubscribe from the CCPEM list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
|