There are several coding practices that we should adhere to when programing for Enzo. Some are style, some are more important for the health of the code (and other Enzo users’ projects/sanity).
Remember that other programmers will read your code¶
“Everyone knows that debugging is twice as hard as writing a program in the first place. So if you’re as clever as you can be when you write it, how will you ever debug it?” –Brian Kernighan “The Elements of Programming Style”, 2nd edition, chapter 2
float is double¶
One must constantly be wary of the possibility of built in C types to be re-defined to higher precision types. This is outlined in Variable precision in Enzo.
Allocating arrays with
Enzo has a significant issue with memory fragmentation. This is due
to grid and particle storage arrays being constantly created,
destroyed, and re-created with slightly different sizes. The most
successful solution to this problem has been to limit the sizes of
arrays to powers of 2. When compiling Enzo with log2alloc-yes,
new is overloaded such that arrays are always created with sizes
that are the nearest power of 2. Thus, it is important to keep in
mind that arrays created with
new will usually be slightly
larger than you think.
Unlike Enzo’s C and C++ routines, Fortran files (.F and .F90) do not re-define the built-in ‘integer’ and ‘real’ types, so all variables and constants must be defined with the appropriate precision. There are pre-defined type specifiers that will match Enzo’s C and C++ precision re-definitions, which should be used for all variables that pass through the C/Fortran interface. This is discussed in detail in Variable precision in Enzo.
Header files must be included in the correct order. This is due, among other
things, to the redefinition of float which is done in
macros_and_parameters.h. This must be done before Enzo headers, but after
external libraries. The order should be as follows:
#include "ErrorExceptions.h" #include "svn_version.def" #include "EnzoTiming.h" #include "performance.h" #include "macros_and_parameters.h" #include "typedefs.h" #include "global_data.h" #include "units.h" #include "flowdefs.h" #include "Fluxes.h" #include "GridList.h" #include "ExternalBoundary.h" #include "Grid.h" #include "Hierarchy.h" #include "LevelHierarchy.h" #include "TopGridData.h" #include "communication.h" #include "CommunicationUtilities.h"
Access data in the BaryonField array as is described in the page on Accessing Data in BaryonField.
Accessing the Hierarchy¶
The hierarchy should be traversed as described in Getting Around the Hierarchy: Linked Lists in Enzo.
The enum construct in C has no standardized size, which can cause problems when using 64 bit integers. Direct integer assignment should be used instead. This also has the added advantage of making explicit the values of parameters that are also used in parameter files. The typical idiom should be:
#ifdef SMALL_INTS typedef int hydro_method; #endif #ifdef LARGE_INTS typedef long_int hydro_method; #endif const hydro_method PPM_DirectEuler = 0, PPM_LagrangeRemap = 1, Zeus_Hydro = 2, HD_RK = 3, MHD_RK = 4, HydroMethodUndefined = 5;