summaryrefslogtreecommitdiff
path: root/lib
Commit message (Collapse)AuthorAge
* Fix for PR316 - can now use LIBS=... in configuration process.goodale2000-04-17
| | | | | | | | | Removed checking for wchan.h, as we don't need it after all. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1551 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added a couple of lines to recognise the NEC SX5 machine.goodale2000-04-17
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1549 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Now havegoodale2000-04-17
| | | | | | | | | | | | | | | | | | HAVE_WCHAN_H and HAVE_SCANDIR for checking if things are available. Also a rather more verbose message if the architecture isn't in the known-architecture list. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1548 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added missing libraries for Lahey/Fujitsu compilerallen2000-04-16
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1546 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Ian's changes to remove bold text when prompt=noallen2000-04-16
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1545 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixed Cactus/302allen2000-04-16
| | | | | | | (Treatment of free format fortran) git-svn-id: http://svn.cactuscode.org/flesh/trunk@1543 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fix for Cactus/313 (comments in free fortran format)allen2000-04-16
| | | | | | | | | | | This isn't a perfect solution, but fixes the reported case. It will still break with CCTK_WARN(0,"blah blah") ! CCTK_WARN(0,"very long line ..... ") Complete fix will be added soon .... git-svn-id: http://svn.cactuscode.org/flesh/trunk@1542 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Warning flags and debugging flags for T3Eallen2000-04-16
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1541 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Support for Lahey/Fujitsu F90 compiler for Linuxallen2000-04-15
| | | | | | | Supplied by Eric Schnetter git-svn-id: http://svn.cactuscode.org/flesh/trunk@1539 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Explicit cast to void * to prevent warnings.goodale2000-04-13
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1537 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Protecting #defines of inline and const against c++ inclusion.goodale2000-04-13
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1536 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Changed to use the xlc and xlC compilers by default.goodale2000-04-13
| | | | | | | | | MPI=NATIVE now works (at least on the Argonne SP2). Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1535 17b73243-c579-4c4c-a9d2-2d5706c11dac
* First versionallen2000-04-12
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1534 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added since now needed with the new config.guess and config.sub.goodale2000-04-12
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1533 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Latest config.guess and config.sub from the autoconf cvs server. This willgoodale2000-04-12
| | | | | | | | | probably temporarily break things on some platforms, this should be fixed soon. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1532 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added for Maui SP2.goodale2000-04-12
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1531 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixed a typo in configuration of HDF5.tradke2000-04-05
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1529 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Changes from Ian for running in noninteractive modeallen2000-04-05
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1527 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Parse the interface.ccl for constant sized arrays on each processor.allen2000-04-04
| | | | | | | | | | | | | | | | Notation is [DISTRIB=CONSTANT|DEFAULT] DEFAULT will distribute the arrays in the default way (at the moment decided by the driver) across processors CONSTANT will put a constant sized array on each processor This is implemented now by PUGH git-svn-id: http://svn.cactuscode.org/flesh/trunk@1518 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Copy parameter files to examples directory rather than thornparfilesallen2000-04-04
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1517 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Configuration file for PACX MPI.tradke2000-04-04
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1516 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Use CCTK_MPI instead of MPIallen2000-04-04
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1515 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Use the NATIVE_MPI_XXX variables from the known-archs stuff heretradke2000-04-04
| | | | | | | | | | to finally set the MPI libs, libdirs, and includedirs. Also test if MPI=NATIVE was selected but the system doesn't provide a native MPI. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1514 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Renamed shell variables describing the system's native MPI implementationtradke2000-04-04
| | | | | | | | into NATIVE_MPI_XXX. If a system doesn't have a native MPI it just doesn't set it here. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1513 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Changedallen2000-04-04
| | | | | | | | | | | | | | | | | #define MPI to #define CCTK_MPI CCTK_MPI #define MPI MPI (this one will be removed in beta 8) Now external packages which the flesh knows about will be defined with a CCTK_, to indicate this, but this won't be made fully consistent until the beta8 when there will be improvements to how the external packages are treated. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1511 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Now if the :: delimiting a description of the range is missing, the rangegoodale2000-04-04
| | | | | | | | | doesn't end up blank. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1510 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixing compiler warningsallen2000-03-30
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1506 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Changing names of 'hidden' parameters - i.e. unwanted parameters fromgoodale2000-03-30
| | | | | | | | | | | | | shared implementations. These names are now unique only within this header file, rather than globally. This fixes the rather nasty problem of whenever you added or removed a thorn which has a restricted parameter, every thorn which shared a parameter from any other thorn could potentially be rebuilt ! Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1504 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Changing comment in built header files so it doesn't change dependinggoodale2000-03-30
| | | | | | | | | on the name of the first thorn requesting it. This fixes PR 293 Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1503 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Declare the Fortran wrapper prototype arguments for the dimensions to betradke2000-03-28
| | | | | | | | | 'const int *'. Thanks Tom. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1501 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Search explicitely in /usr/lib/ for libz.a,tradke2000-03-27
| | | | | | | | | | and don't add it to the LIBDIRS if found. This prevents picking up the wrong (32bit) library on 64bit SGIs. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1495 17b73243-c579-4c4c-a9d2-2d5706c11dac
* bug fix in the USE_MACRO applicationlanfer2000-03-23
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1490 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Ed Evans changes to add timelevels properlyallen2000-03-23
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1486 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Make some effort to catch NaNs in comparison files, this needs to be greatlyallen2000-03-23
| | | | | | | improved git-svn-id: http://svn.cactuscode.org/flesh/trunk@1485 17b73243-c579-4c4c-a9d2-2d5706c11dac
* A few changes to an initial script to generate latex from the param.cclallen2000-03-21
| | | | | | | files. This is not a working script in any sense. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1482 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixing order of libraries for pgf90.goodale2000-03-16
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1472 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Checks for zlib in old 1.0.x versions of HDF5.tradke2000-03-14
| | | | | | | | | | Fixes (second part of) Erik's BR Compiler/274. Thomas git-svn-id: http://svn.cactuscode.org/flesh/trunk@1471 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Removing T3E libraries since they don't improve performanceallen2000-03-13
| | | | git-svn-id: http://svn.cactuscode.org/flesh/trunk@1466 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Addition of g2c/f2c library when using g77 and no f90 available.goodale2000-03-13
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1460 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixing the default fortran name script which is generated when there isgoodale2000-03-13
| | | | | | | | | no fortran whatsoever. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1459 17b73243-c579-4c4c-a9d2-2d5706c11dac
* revised design for the USE_MACRO perl routine, tested on standard and ↵lanfer2000-03-10
| | | | | | development arrangement git-svn-id: http://svn.cactuscode.org/flesh/trunk@1456 17b73243-c579-4c4c-a9d2-2d5706c11dac
* New option for gmakeallen2000-03-10
| | | | | | | | | | | | | | | | | | | | PROMPT=no using this at either configuration or compile time will run gmake until an executable is produced without giving further prompts. The default is PROMPT=yes This means that you can do gmake <config> THORNLIST=MyFavList PROMPT=no to get an executable. This is experimental at the moment, since I might change it so that PROMPT=no still requires you to do a separate configuation and compile step, and add a new option to run them together. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1454 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Fixed cvsdiff to do what it was supposed to.allen2000-03-09
| | | | | | | Note that cvsdiff tells you the difference between the file you have checked out and possibly edited, and the latest version in the repository ... it doesn't tell you what "cvs diff" would tell you (the changes you have made to the file you checked out). I will change the name if I can think of a better one, to make this difference clearer. git-svn-id: http://svn.cactuscode.org/flesh/trunk@1451 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added /usr/local/hdf5/serial/ to the search path for HDF5 installations.tradke2000-03-08
| | | | | | | | | | | | | | | | | This is where they installed HDF5 on berte.zib.de. As soon as parallel HDF5 is working the path should be changed to point to /usr/local/hdf5/parallel/. This fixes BR IO:212. The script now also checks if libhdf5.a was compiled with libz.a - if so, it does a search on this lib and adds it to LIBS and LIBDIRS. You can also set LIBZ_DIR manually. Thomas git-svn-id: http://svn.cactuscode.org/flesh/trunk@1448 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Now compiles the flesh out of the box 8-)goodale2000-03-07
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1441 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Addedgoodale2000-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | C_POSTPROCESSING CXX_POSTPROCESSING F77_POSTPROCESSING F_POSTPROCESSING F90_POSTPROCESSING as things a known-architecture can set. This allows object files to be moved and things like that. Also added the ability of the architecture files to write to a file calles make.arch.defn which is included at the bottom of make.code.defn. Thus, for example, an architecture requiring mpicc and its ilk for compiling with MPI can replace the compilers at this point, but still allow the rest of the configuration to use the normal compiler. Use with caution. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1438 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Added yet another thing which can be set by the known-architecture stuff:goodale2000-03-06
| | | | | | | | | | | | | | | | | | | | | | | | CXX_WORKING_NAME this defaults to the working name being the filename without the directory info, but can be used to give the file generated after preprocessing a different name. E.g. if your machine requires C++ files to end with .C you should be able to do CXX_WORKING_NAME='$(notdir $<).C to solve the problem. At some point we're going to have to go through all these obscure flags and decide on nice names and what functionality they really should have. For instance the above could also have been done by introducing CXX_SUFFIX which would of course be easier for the end-user, but might in some circumstances be less useful. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1420 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Check to see if determination of cache size was succesful. Should solvegoodale2000-03-06
| | | | | | | | | | proablems were the resulting #define was of the form *1024 thus leading to compilation errors. Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1419 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Dependencies now work (at least for C).goodale2000-03-06
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1418 17b73243-c579-4c4c-a9d2-2d5706c11dac
* Removing some pesky errors. Really ugly.goodale2000-03-06
| | | | | | | Tom git-svn-id: http://svn.cactuscode.org/flesh/trunk@1417 17b73243-c579-4c4c-a9d2-2d5706c11dac