Loading doc/src/Build_settings.rst +59 −61 Original line number Original line Diff line number Diff line Loading @@ -341,8 +341,9 @@ If this option is enabled, large files can be read or written with gzip compression by several LAMMPS commands, including gzip compression by several LAMMPS commands, including :doc:`read_data <read_data>`, :doc:`rerun <rerun>`, and :doc:`dump <dump>`. :doc:`read_data <read_data>`, :doc:`rerun <rerun>`, and :doc:`dump <dump>`. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash Loading @@ -350,28 +351,24 @@ CMake build # default is yes if CMake can find gzip, else no # default is yes if CMake can find gzip, else no -D GZIP_EXECUTABLE=path # path to gzip executable if CMake cannot find it -D GZIP_EXECUTABLE=path # path to gzip executable if CMake cannot find it Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make LMP_INC = -DLAMMPS_GZIP LMP_INC = -DLAMMPS_GZIP CMake and make info This option requires that your operating system fully supports the "popen()" ^^^^^^^^^^^^^^^^^^^ function in the standard runtime library and that a ``gzip`` executable can be This option requires that your machine supports the "popen()" function in the standard runtime library and that a gzip executable can be found by LAMMPS during a run. found by LAMMPS during a run. .. note:: .. note:: On some clusters with high-speed networks, using the fork() On some clusters with high-speed networks, using the "fork()" library library call (required by popen()) can interfere with the fast call (required by "popen()") can interfere with the fast communication communication library and lead to simulations using compressed output library and lead to simulations using compressed output or input to or input to hang or crash. For selected operations, compressed file hang or crash. For selected operations, compressed file I/O is also I/O is also available using a compression library instead, which is available using a compression library instead, which is what the what the :ref:`COMPRESS package <PKG-COMPRESS>` enables. :ref:`COMPRESS package <PKG-COMPRESS>` enables. ---------- ---------- Loading @@ -380,37 +377,38 @@ found by LAMMPS during a run. Memory allocation alignment Memory allocation alignment --------------------------------------- --------------------------------------- This setting enables the use of the posix_memalign() call instead of This setting enables the use of the "posix_memalign()" call instead of malloc() when LAMMPS allocates large chunks or memory. This can make "malloc()" when LAMMPS allocates large chunks or memory. Vector vector instructions on CPUs more efficient, if dynamically allocated instructions on CPUs may become more efficient, if dynamically allocated memory is aligned on larger-than-default byte boundaries. memory is aligned on larger-than-default byte boundaries. On most On most current systems, the malloc() implementation returns current operating systems, the "malloc()" implementation returns pointers that are aligned to 16-byte boundaries. Using SSE vector pointers that are aligned to 16-byte boundaries. Using SSE vector instructions efficiently, however, requires memory blocks being instructions efficiently, however, requires memory blocks being aligned aligned on 64-byte boundaries. on 64-byte boundaries. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_MEMALIGN=value # 0, 8, 16, 32, 64 (default) -D LAMMPS_MEMALIGN=value # 0, 8, 16, 32, 64 (default) Use a ``LAMMPS_MEMALIGN`` value of 0 to disable using posix_memalign() Use a ``LAMMPS_MEMALIGN`` value of 0 to disable using and revert to using the malloc() C-library function instead. When "posix_memalign()" and revert to using the "malloc()" C-library compiling LAMMPS for Windows systems, malloc() will always be used function instead. When compiling LAMMPS for Windows systems, and this setting ignored. "malloc()" will always be used and this setting is ignored. Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make LMP_INC = -DLAMMPS_MEMALIGN=value # 8, 16, 32, 64 LMP_INC = -DLAMMPS_MEMALIGN=value # 8, 16, 32, 64 Do not set ``-DLAMMPS_MEMALIGN``, if you want to have memory allocated Do not set ``-DLAMMPS_MEMALIGN``, if you want to have memory with the malloc() function call instead. ``-DLAMMPS_MEMALIGN`` **cannot** allocated with the "malloc()" function call be used on Windows, as it does use different function calls for instead. ``-DLAMMPS_MEMALIGN`` **cannot** be used on Windows, as Windows different function calls with different semantics for allocating aligned memory, that are not compatible with how LAMMPS allocating aligned memory, that are not compatible with how LAMMPS manages its dynamical memory. manages its dynamical memory. Loading @@ -419,22 +417,22 @@ manages its dynamical memory. .. _longlong: .. _longlong: Workaround for long long integers Workaround for long long integers ------------------------------------------------ --------------------------------- If your system or MPI version does not recognize "long long" data If your system or MPI version does not recognize "long long" data types, the following setting will be needed. It converts "long long" types, the following setting will be needed. It converts "long long" to a "long" data type, which should be the desired 8-byte integer on to a "long" data type, which should be the desired 8-byte integer on those systems: those systems: CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_LONGLONG_TO_LONG=value # yes or no (default) -D LAMMPS_LONGLONG_TO_LONG=value # yes or no (default) Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make Loading @@ -453,15 +451,15 @@ Instead, the call stack is unwound and control returns to the caller, e.g. to Python. Of course, the calling code has to be set up to e.g. to Python. Of course, the calling code has to be set up to *catch* exceptions thrown from within LAMMPS. *catch* exceptions thrown from within LAMMPS. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_EXCEPTIONS=value # yes or no (default) -D LAMMPS_EXCEPTIONS=value # yes or no (default) Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make Loading Loading
doc/src/Build_settings.rst +59 −61 Original line number Original line Diff line number Diff line Loading @@ -341,8 +341,9 @@ If this option is enabled, large files can be read or written with gzip compression by several LAMMPS commands, including gzip compression by several LAMMPS commands, including :doc:`read_data <read_data>`, :doc:`rerun <rerun>`, and :doc:`dump <dump>`. :doc:`read_data <read_data>`, :doc:`rerun <rerun>`, and :doc:`dump <dump>`. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash Loading @@ -350,28 +351,24 @@ CMake build # default is yes if CMake can find gzip, else no # default is yes if CMake can find gzip, else no -D GZIP_EXECUTABLE=path # path to gzip executable if CMake cannot find it -D GZIP_EXECUTABLE=path # path to gzip executable if CMake cannot find it Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make LMP_INC = -DLAMMPS_GZIP LMP_INC = -DLAMMPS_GZIP CMake and make info This option requires that your operating system fully supports the "popen()" ^^^^^^^^^^^^^^^^^^^ function in the standard runtime library and that a ``gzip`` executable can be This option requires that your machine supports the "popen()" function in the standard runtime library and that a gzip executable can be found by LAMMPS during a run. found by LAMMPS during a run. .. note:: .. note:: On some clusters with high-speed networks, using the fork() On some clusters with high-speed networks, using the "fork()" library library call (required by popen()) can interfere with the fast call (required by "popen()") can interfere with the fast communication communication library and lead to simulations using compressed output library and lead to simulations using compressed output or input to or input to hang or crash. For selected operations, compressed file hang or crash. For selected operations, compressed file I/O is also I/O is also available using a compression library instead, which is available using a compression library instead, which is what the what the :ref:`COMPRESS package <PKG-COMPRESS>` enables. :ref:`COMPRESS package <PKG-COMPRESS>` enables. ---------- ---------- Loading @@ -380,37 +377,38 @@ found by LAMMPS during a run. Memory allocation alignment Memory allocation alignment --------------------------------------- --------------------------------------- This setting enables the use of the posix_memalign() call instead of This setting enables the use of the "posix_memalign()" call instead of malloc() when LAMMPS allocates large chunks or memory. This can make "malloc()" when LAMMPS allocates large chunks or memory. Vector vector instructions on CPUs more efficient, if dynamically allocated instructions on CPUs may become more efficient, if dynamically allocated memory is aligned on larger-than-default byte boundaries. memory is aligned on larger-than-default byte boundaries. On most On most current systems, the malloc() implementation returns current operating systems, the "malloc()" implementation returns pointers that are aligned to 16-byte boundaries. Using SSE vector pointers that are aligned to 16-byte boundaries. Using SSE vector instructions efficiently, however, requires memory blocks being instructions efficiently, however, requires memory blocks being aligned aligned on 64-byte boundaries. on 64-byte boundaries. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_MEMALIGN=value # 0, 8, 16, 32, 64 (default) -D LAMMPS_MEMALIGN=value # 0, 8, 16, 32, 64 (default) Use a ``LAMMPS_MEMALIGN`` value of 0 to disable using posix_memalign() Use a ``LAMMPS_MEMALIGN`` value of 0 to disable using and revert to using the malloc() C-library function instead. When "posix_memalign()" and revert to using the "malloc()" C-library compiling LAMMPS for Windows systems, malloc() will always be used function instead. When compiling LAMMPS for Windows systems, and this setting ignored. "malloc()" will always be used and this setting is ignored. Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make LMP_INC = -DLAMMPS_MEMALIGN=value # 8, 16, 32, 64 LMP_INC = -DLAMMPS_MEMALIGN=value # 8, 16, 32, 64 Do not set ``-DLAMMPS_MEMALIGN``, if you want to have memory allocated Do not set ``-DLAMMPS_MEMALIGN``, if you want to have memory with the malloc() function call instead. ``-DLAMMPS_MEMALIGN`` **cannot** allocated with the "malloc()" function call be used on Windows, as it does use different function calls for instead. ``-DLAMMPS_MEMALIGN`` **cannot** be used on Windows, as Windows different function calls with different semantics for allocating aligned memory, that are not compatible with how LAMMPS allocating aligned memory, that are not compatible with how LAMMPS manages its dynamical memory. manages its dynamical memory. Loading @@ -419,22 +417,22 @@ manages its dynamical memory. .. _longlong: .. _longlong: Workaround for long long integers Workaround for long long integers ------------------------------------------------ --------------------------------- If your system or MPI version does not recognize "long long" data If your system or MPI version does not recognize "long long" data types, the following setting will be needed. It converts "long long" types, the following setting will be needed. It converts "long long" to a "long" data type, which should be the desired 8-byte integer on to a "long" data type, which should be the desired 8-byte integer on those systems: those systems: CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_LONGLONG_TO_LONG=value # yes or no (default) -D LAMMPS_LONGLONG_TO_LONG=value # yes or no (default) Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make Loading @@ -453,15 +451,15 @@ Instead, the call stack is unwound and control returns to the caller, e.g. to Python. Of course, the calling code has to be set up to e.g. to Python. Of course, the calling code has to be set up to *catch* exceptions thrown from within LAMMPS. *catch* exceptions thrown from within LAMMPS. CMake build .. tabs:: ^^^^^^^^^^^ .. tab:: CMake build .. code-block:: bash .. code-block:: bash -D LAMMPS_EXCEPTIONS=value # yes or no (default) -D LAMMPS_EXCEPTIONS=value # yes or no (default) Traditional make .. tab:: Traditional make ^^^^^^^^^^^^^^^^ .. code-block:: make .. code-block:: make Loading
