508.namd_r.flags

These are flags that appear in the benchmark sources, categorized as to whether they are portability candidates, mandatory, or forbidden.

]]>

Include an "erfc" routine. Use on systems which do not have the C99 erfc routine available. Add the Alt Viral Tensor reduction tags. Add Power PC specific assembly and alignment instructions. Used internally to enable energy calculation. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. If defined, includes several include files already included elsewhere. Define FALSE to be 0. Use the "FAST" methods. (default) Defines the extension of the FEP routine names. Not used in SPEC version. Use the "FULL" method. Enable the "FULL" method. Use the "int" extension for FEP routine names. INT_MAX is defined in "limits.h" and is used to check whether to typedef "int32" as an "int" or a "short". Guard so include file doesn't get included twice. Use the "les" extension for FEP routine names. Guard so include file doesn't get included twice. A HACK to re-sort FullAtom list used in Parallel IO. When every home patch processor receives its atoms list for a patch, the atoms inside this patch may not sorted according to hydList value to save space, use anonymous union data structure to share the space of "langevinParam" to store "hydList" from an InputAtom and then sort the atom list. The "langevinParam" value is not initialized until home patch creation. Use the merge elect routines. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Enable CUDA version of NAMD. Not available in SPEC CPU version. Enables Atom Sort. Basic Idea: Determine the line between the center of masses of the two patches. Project and then sort the lists of atom along this line. Then, as the pairlists are being generated for the atoms in the first atom list, use the sorted list to only add atoms from the second list that are between +/- ~cutoff from the atoms position on the line. Do not allow the use of SSE intrinsics. If no set, then the restrict keyword is defined to blank. However, restrict is not used anywhere in the source. If not zero, defines DEFINE_CHECK_WATER_SEPARATION macro. Not used in SPEC CPU version. Defined internally for the NonBonded pair interaction type. Defined internally for the NonBonded Self interaction type. Determines which Nonbonded type to use. Defined internally. Add a call to "CkNetworkProgress" routine to check the network's progress. Is routine is not defined in the SPEC CPU version. "NO" is internally defined as 0. Guard so include file doesn't get included twice. "ONE" is set internally to 1.00000000000. Internal macro which gets set when NBPAIR is used. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. "PI" macro set internally to 3.141592653589793. Enables the placement new operator. Enables the placement new operator in the global namespace. Use the "pprof" extension for FEP routine names. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Guard so include file doesn't get included twice. Internal macro which gets set when NBSELF is used. Internal macro used to define the short method. Guard so include file doesn't get included twice. Use a simpler method to determine the pair list. Only use the slow method. Disables "FAST". Defines SMALLRAD macro to 0.0005. Defines SMALLRAD2 macro to 0.0005*0.0005. Defines the STRINGNULL macro to '\0'. Interanally defined marco. Use the "tabener" extension for FEP routine names. Guard so include file doesn't get included twice. Enables a "main" routine within the Vector.h files which tests the basic functionality of the vector class. Use the "ti" extension for FEP routine names. Define the TWOPI macro to 2.0 * PI. Guard so include file doesn't get included twice. Include a version of erfc when running on Windows. Please use -DSPEC_NEED_ERFC instead. Enables the erfc routine. Please use -DSPEC_NEED_ERFC instead. Define the ZERO macro to 0. Set compiler specfic alignment attributes when using SSE intrinsics. SSE intrinsics are disabled in the SPEC CPU version. Set compiler specfic alignment attributes when using SSE intrinsics. SSE intrinsics are disabled in the SPEC CPU version. Set compiler specfic alignment attributes when using SSE intrinsics. SSE intrinsics are disabled in the SPEC CPU version. Enable SSE intrinsic routines. Internally defined macro. Internally defined macro.