Restarting a simulation from last data
******************* This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address. ******************* Dear all, I understand Nektar has a restart capability which involves renaming the last .fld output file as a .rst file. Then, configure the FUNCTION 'InitialConditions' accordingly. In case of large files, the .fld file is partitioned in files with names P0000000.fld, P0000001.fld, etc and the Info.xml file is generated. How do we generate the start file from the parallel partition? I assume a single .fld file is needed for restarting a simulation. Thank you for your help. Kind regards, Silvia
Hi Silvia, I normally use FieldConvert to reduce the number of files in a .fld folder before downloading flies from supercomputer, mpirun -np <nprocs> FieldConvert test.xml test.fld reconstruct-test.fld. To get a single *.fld file, <nprocs> = 1 should be used. Hope it helps! Kind regards, Chengjiao ===================================== 任珵娇；Chengiiao Ren， 博士研究生；Ph.D Candidate， 建设工程学部；Faculty of Infrastructure Engineering， 大连理工大学； Dalian University of Technology， 大连，中国 116024；Dalian，CHINA 116024. 电话，Phone：86-18504285808，61-413077330 电子邮箱，Email：dlut_renchengjiao@163.com; chengjiao.ren@outlook.com; ===================================== From: Silvia Ceccacci (HDR) Sent: Thursday, 13 August 2020 12:42 PM To: nektar-users@imperial.ac.uk Subject: [Nektar-users] Restarting a simulation from last data This email from silvia.ceccacci@hdr.mq.edu.au originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list to disable email stamping for this address. Dear all, I understand Nektar has a restart capability which involves renaming the last .fld output file as a .rst file. Then, configure the FUNCTION 'InitialConditions' accordingly. In case of large files, the .fld file is partitioned in files with names P0000000.fld, P0000001.fld, etc and the Info.xml file is generated. How do we generate the start file from the parallel partition? I assume a single .fld file is needed for restarting a simulation. Thank you for your help. Kind regards, Silvia
Hi Chengjjao, That works perfectly. Thank you very much. Silvia ________________________________ From: Chengjiao Ren <dlut_renchengjiao@163.com> Sent: Friday, 14 August 2020 12:37 AM To: Silvia Ceccacci (HDR) <silvia.ceccacci@hdr.mq.edu.au>; nektar-users@imperial.ac.uk <nektar-users@imperial.ac.uk> Subject: RE: [Nektar-users] Restarting a simulation from last data Hi Silvia, I normally use FieldConvert to reduce the number of files in a .fld folder before downloading flies from supercomputer, mpirun -np <nprocs> FieldConvert test.xml test.fld reconstruct-test.fld. To get a single *.fld file, <nprocs> = 1 should be used. Hope it helps! Kind regards, Chengjiao ===================================== 任珵娇；Chengiiao Ren， 博士研究生；Ph.D Candidate， 建设工程学部；Faculty of Infrastructure Engineering， 大连理工大学； Dalian University of Technology， 大连，中国 116024；Dalian，CHINA 116024. 电话，Phone：86-18504285808，61-413077330 电子邮箱，Email：dlut_renchengjiao@163.com<mailto:dlut_renchengjiao@163.com>; chengjiao.ren@outlook.com<mailto:chengjiao.ren@outlook.com>; ===================================== From: Silvia Ceccacci (HDR)<mailto:silvia.ceccacci@hdr.mq.edu.au> Sent: Thursday, 13 August 2020 12:42 PM To: nektar-users@imperial.ac.uk<mailto:nektar-users@imperial.ac.uk> Subject: [Nektar-users] Restarting a simulation from last data This email from silvia.ceccacci@hdr.mq.edu.au originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list<https://spam.ic.ac.uk/SpamConsole/Senders.aspx> to disable email stamping for this address. Dear all, I understand Nektar has a restart capability which involves renaming the last .fld output file as a .rst file. Then, configure the FUNCTION 'InitialConditions' accordingly. In case of large files, the .fld file is partitioned in files with names P0000000.fld, P0000001.fld, etc and the Info.xml file is generated. How do we generate the start file from the parallel partition? I assume a single .fld file is needed for restarting a simulation. Thank you for your help. Kind regards, Silvia
Dear Chengjiao, Silvia, It is possible just to copy or move the whole parallel directory containing the partitioned files to a new name (i.e. mv file.fld file.rst ) and then the code knows how to read the partitioned files from this directory to restart the run. @Chengjiao: This is a better approach than making it into a single file since otherwise each processor has to read the whole file when restarting that can lead to memory errors. There is a relatively new feature of using hdf5 output files where I believe the output appears as a single file but can be read in smaller chunks to also address this issue. Cheers, Spencer. Spencer Sherwin FREng, FRAeS Head of Aerodynamics Section, Director of Research Computing Service, Professor of Computational Fluid Mechanics, Department of Aeronautics, South Kensington Campus, Imperial College London, SW7 2AZ, UK Phone: +44 (0)20 7594 5052 http://www.imperial.ac.uk/people/s.sherwin/ On 13 Aug 2020, at 16:37, Chengjiao Ren <dlut_renchengjiao@163.com<mailto:dlut_renchengjiao@163.com>> wrote: Hi Silvia, I normally use FieldConvert to reduce the number of files in a .fld folder before downloading flies from supercomputer, mpirun -np <nprocs> FieldConvert test.xml test.fld reconstruct-test.fld. To get a single *.fld file, <nprocs> = 1 should be used. Hope it helps! Kind regards, Chengjiao ===================================== 任珵娇；Chengiiao Ren， 博士研究生；Ph.D Candidate， 建设工程学部；Faculty of Infrastructure Engineering， 大连理工大学； Dalian University of Technology， 大连，中国 116024；Dalian，CHINA 116024. 电话，Phone：86-18504285808，61-413077330 电子邮箱，Email：dlut_renchengjiao@163.com<mailto:dlut_renchengjiao@163.com>; chengjiao.ren@outlook.com<mailto:chengjiao.ren@outlook.com>; ===================================== From: Silvia Ceccacci (HDR)<mailto:silvia.ceccacci@hdr.mq.edu.au> Sent: Thursday, 13 August 2020 12:42 PM To: nektar-users@imperial.ac.uk<mailto:nektar-users@imperial.ac.uk> Subject: [Nektar-users] Restarting a simulation from last data This email from silvia.ceccacci@hdr.mq.edu.au<mailto:silvia.ceccacci@hdr.mq.edu.au> originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list<https://spam.ic.ac.uk/SpamConsole/Senders.aspx> to disable email stamping for this address. Dear all, I understand Nektar has a restart capability which involves renaming the last .fld output file as a .rst file. Then, configure the FUNCTION 'InitialConditions' accordingly. In case of large files, the .fld file is partitioned in files with names P0000000.fld, P0000001.fld, etc and the Info.xml file is generated. How do we generate the start file from the parallel partition? I assume a single .fld file is needed for restarting a simulation. Thank you for your help. Kind regards, Silvia _______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk<mailto:Nektar-users@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/nektar-users
participants (4)
-
Chengjiao Ren
-
Sherwin, Spencer J
-
Silvia Ceccacci (HDR)
-
ztdepyahoo