Changes between Version 4 and Version 5 of doc/app/cpp_options

Timestamp:

Jun 5, 2019 1:51:25 PM (6 years ago)

Author:

raasch

Comment:

--

doc/app/cpp_options

@@ -29 +29 @@
 ||{{{__parallel}}}          ||Enables usage of MPI. If you do '''not set''' {{{-D__parallel}}}, PALM runs in non-parallel mode on a single core.  ||
 ||{{{__rrtmg}}}             ||Enables usage of the Rapid Radiation Transfer Model library. See [wiki:/doc/tec/radiation#RRTMG] for further details.   ||
+||{{{__single_precision}}}  ||'''Attention: This is a test feature.''' \\Enables 32bit working precision for all REALs. Based on r4015, single precision works for the PALM core (cyclic horizontal boundaries, no additinal modules switched on) with FFT- and MG-pressure solvers. Other setups have not been tested so far. ||
 
 
@@ -40 +41 @@
 ||{{{-DMPI_2REAL=MPI_2DOUBLE_PRECISION}}}  ||Replaces strings {{{MPI_2REAL}}} by {{{MPI_2DOUBLE_PRECISION}}} wherever they appear in the code.   ||
 
-These preprocessor options are required, because by default the precision of {{{REAL}}} quantities in PALM is set to 64bit. You may define 32bit as the working precision (see routine [source:palm/trunk/SOURCE/mod_kinds.f90@3151#L69]). In such a case, you must omit the above preprocessor options. 
+These preprocessor options are required, because by default the precision of {{{REAL}}} quantities in PALM is set to 64bit. You may define 32bit as the working precision (see routine [source:palm/trunk/SOURCE/mod_kinds.f90@3151#L69]) by setting the preprocessor option {{{-D__single_precision}}}. In such a case, you must omit the above preprocessor options.