Changes between Version 213 and Version 214 of doc/app/runtime_parameters
- Timestamp:
- Feb 9, 2011 2:45:18 PM (14 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
doc/app/runtime_parameters
v213 v214 3 3 ==== [#output Data output] ==== 4 4 ==== [#run Run steering] ==== 5 ==== [#pgrid Processor grid ] ====5 ==== [#pgrid Processor grid / MPI settings] ==== 6 6 \\\\\\\\\\\\\\\\\\\\\\\\\\ 7 7 [=#output '''Data output:]\\ … … 1457 1457 }}} 1458 1458 \\\\ 1459 [=#pgrid '''Processor grid :]\\1459 [=#pgrid '''Processor grid / MPI settings:]\\ 1460 1460 ||='''Parameter Name''' =||='''[[../fortrantypes|FORTRAN]]\\[[../fortrantypes|Type]]''' =||='''Default\\Value''' =||='''Explanation''' =|| 1461 1461 |---------------- … … 1489 1489 Number of processors along y-direction of the virtual processor net.\\\\ 1490 1490 For further information see [#npex npex]. 1491 }}} 1492 |---------------- 1493 {{{#!td style="vertical-align:top" 1494 [=#synchronous_exchange '''synchronous_exchange'''] 1495 }}} 1496 {{{#!td style="vertical-align:top" 1497 L 1498 }}} 1499 {{{#!td style="vertical-align:top" 1500 .F. 1501 }}} 1502 {{{#!td 1503 Defines how MPI handles the exchange of ghost points. By default ({{{.F.}}}), asynchronous transfer with {{{MPI_ISEND}}} and {{{MPI_IRECV}}} is used. In case of '''synchronous_exchange''' = {{{.T.}}} {{{MPI_SENDRECV}}} is used instead. On most networks the asynchronous method will give better performance. 1491 1504 }}}