Difference between revisions of "Main Page/PIO"
From Nekcem
Jump to navigationJump to search| Line 17: | Line 17: | ||
---- | ---- | ||
Several advanced parallel I/O approaches were developed. Performance for those approaches | Several advanced parallel I/O approaches were developed. Performance for those approaches | ||
| − | are detailed in "Parallel I/O performance for application-level checkpointing on the Blue | + | are detailed in "Jing Fu et. al, Parallel I/O performance for application-level checkpointing |
| − | Gene/P http://www.mcs.anl.gov/~mmin/ICS2011.pdf. | + | on the Blue Gene/P" http://www.mcs.anl.gov/~mmin/ICS2011.pdf. |
| Line 25: | Line 25: | ||
: '''param(81) = 5''': collective IO for N processors to multiple M-files (mpi-binary-NM-xxx.vtk. <br> | : '''param(81) = 5''': collective IO for N processors to multiple M-files (mpi-binary-NM-xxx.vtk. <br> | ||
| − | : '''param( | + | : '''param(81) = 6''': reduced-blocking IO for N processors to 1 file with M writers (mpi-binary-NM1-xxx.vtk). <br> |
| − | : '''param( | + | : '''param(81) = 6''': reduced-blocking IO for N processors to 1 file with M writers (mpi-ascii-NM1-xxx.vtk). <br> |
| − | : '''param( | + | : '''param(81) = 8''': reduced-blocking IO for N processors to multiple M-files with M writers (mpi-NMM-xxx.vtk). <br> |
| Line 35: | Line 35: | ||
Note that param(82) and param(83) need to be set correctly in *.rea file. | Note that param(82) and param(83) need to be set correctly in *.rea file. | ||
| − | param(82) = number of | + | '''param(82)''' = number of output files, <br> |
| + | '''param(83)''' = max number of fields to be written. <br> | ||
Revision as of 11:45, 12 June 2011
This is the document page for parallel I/O library developed for NekCEM.
Background
- File Format
- Binary (used for production, compact size), or ASCII (used for debugging, human-readable)
Usage Introduction:
For traditional approaches based on one file per processor, users can use #1,2,3
in .rea file for a specific example, by parameter 81.
These three approaches produce one file per processor and we will refer them as old libraries.
- param(81) = 1: use nek5000 old_output format (ASCII, fld format, xxx-.fld).
- param(81) = 2: use Fortran I/O library (ASCII, VTK format, ascii-xxx-.vtk).
- param(81) = 3: use C-POSIX I/O libraries (binary, VTK format).
Several advanced parallel I/O approaches were developed. Performance for those approaches are detailed in "Jing Fu et. al, Parallel I/O performance for application-level checkpointing on the Blue Gene/P" http://www.mcs.anl.gov/~mmin/ICS2011.pdf.
- param(81) = 4: collective for N processors IO to 1 file (mpi-binary-N1-xxx.vtk)
- param(81) = 5: collective IO for N processors to multiple M-files (mpi-binary-NM-xxx.vtk.
- param(81) = 6: reduced-blocking IO for N processors to 1 file with M writers (mpi-binary-NM1-xxx.vtk).
- param(81) = 6: reduced-blocking IO for N processors to 1 file with M writers (mpi-ascii-NM1-xxx.vtk).
- param(81) = 8: reduced-blocking IO for N processors to multiple M-files with M writers (mpi-NMM-xxx.vtk).
Note that param(82) and param(83) need to be set correctly in *.rea file.
param(82) = number of output files,
param(83) = max number of fields to be written.
