4 |Gromacs| programs may be influenced by the use of
5 environment variables. First of all, the variables set in
6 the ``GMXRC`` file are essential for running and
7 compiling |Gromacs|. Some other useful environment variables are
8 listed in the following sections. Most environment variables function
9 by being set in your shell to any non-NULL value. Specific
10 requirements are described below if other values need to be set. You
11 should consult the documentation for your shell for instructions on
12 how to set environment variables in the current shell, or in configuration
13 files for future shells. Note that requirements for exporting
14 environment variables to jobs run under batch control systems vary and
15 you should consult your local documentation for details.
20 Print constraint virial and force virial energy terms.
23 |Gromacs| automatically backs up old
24 copies of files when trying to write a new file of the same
25 name, and this variable controls the maximum number of
26 backups that will be made, default 99. If set to 0 it fails to
27 run if any output file already exists. And if set to -1 it
28 overwrites any output file without making a backup.
31 if this is explicitly set, no cool quotes
32 will be printed at the end of a program.
35 prevent dumping of step files during
36 (for example) blowing up during failure of constraint
40 dump all configurations to a :ref:`pdb`
41 file that have an interaction energy less than the value set
42 in this environment variable.
45 ``GMX_VIEW_XVG``, ``GMX_VIEW_EPS`` and ``GMX_VIEW_PDB``, commands used to
46 automatically view :ref:`xvg`, :ref:`xpm`, :ref:`eps`
47 and :ref:`pdb` file types, respectively; they default to ``xv``, ``xmgrace``,
48 ``ghostview`` and ``rasmol``. Set to empty to disable
49 automatic viewing of a particular file type. The command will
50 be forked off and run in the background at the same priority
51 as the |Gromacs| tool (which might not be what you want).
52 Be careful not to use a command which blocks the terminal
53 (e.g. ``vi``), since multiple instances might be run.
55 ``GMX_VIRIAL_TEMPERATURE``
56 print virial temperature energy term
59 the size of the buffer for file I/O. When set
60 to 0, all file I/O will be unbuffered and therefore very slow.
61 This can be handy for debugging purposes, because it ensures
62 that all files are always totally up-to-date.
65 set display color for logo in :ref:`gmx view`.
67 ``GMX_PRINT_LONGFORMAT``
68 use long float format when printing
72 Applies for computational electrophysiology setups
73 only (see reference manual). The initial structure gets dumped to
74 :ref:`pdb` file, which allows to check whether multimeric channels have
75 the correct PBC representation.
77 ``GMX_TRAJECTORY_IO_VERBOSITY``
78 Defaults to 1, which prints frame count e.g. when reading trajectory
79 files. Set to 0 for quiet operation.
83 ``GMX_PRINT_DEBUG_LINES``
84 when set, print debugging info on line numbers.
87 number of steps that elapse between dumping
88 the current DD to a PDB file (default 0). This only takes effect
89 during domain decomposition, so it should typically be
90 0 (never), 1 (every DD phase) or a multiple of :mdp:`nstlist`.
92 ``GMX_DD_NST_DUMP_GRID``
93 number of steps that elapse between dumping
94 the current DD grid to a PDB file (default 0). This only takes effect
95 during domain decomposition, so it should typically be
96 0 (never), 1 (every DD phase) or a multiple of :mdp:`nstlist`.
99 general debugging trigger for every domain
100 decomposition (default 0, meaning off). Currently only checks
101 global-local atom index mapping for consistency.
104 over-ride the number of DD pulses used
105 (default 0, meaning no over-ride). Normally 1 or 2.
107 There are a number of extra environment variables like these
108 that are used in debugging - check the code!
110 Performance and Run Control
111 ---------------------------
112 ``GMX_DO_GALACTIC_DYNAMICS``
113 planetary simulations are made possible (just for fun) by setting
114 this environment variable, which allows setting :mdp:`epsilon-r` to -1 in the :ref:`mdp`
115 file. Normally, :mdp:`epsilon-r` must be greater than zero to prevent a fatal error.
116 See webpage_ for example input files for a planetary simulation.
118 ``GMX_ALLOW_CPT_MISMATCH``
119 when set, runs will not exit if the
120 ensemble set in the :ref:`tpr` file does not match that of the
123 ``GMX_CUDA_NB_EWALD_TWINCUT``
124 force the use of twin-range cutoff kernel even if :mdp:`rvdw` equals
125 :mdp:`rcoulomb` after PP-PME load balancing. The switch to twin-range kernels is automated,
126 so this variable should be used only for benchmarking.
128 ``GMX_CUDA_NB_ANA_EWALD``
129 force the use of analytical Ewald kernels. Should be used only for benchmarking.
131 ``GMX_CUDA_NB_TAB_EWALD``
132 force the use of tabulated Ewald kernels. Should be used only for benchmarking.
134 ``GMX_CUDA_STREAMSYNC``
135 force the use of cudaStreamSynchronize on ECC-enabled GPUs, which leads
136 to performance loss due to a known CUDA driver bug present in API v5.0 NVIDIA drivers (pre-30x.xx).
137 Cannot be set simultaneously with ``GMX_NO_CUDA_STREAMSYNC``.
139 ``GMX_DISABLE_CUDALAUNCH``
140 disable the use of the lower-latency cudaLaunchKernel API even when supported (CUDA >=v7.0).
141 Should only be used for benchmarking purposes.
143 ``GMX_DISABLE_CUDA_TIMING``
144 Disables GPU timing of CUDA tasks; synonymous with ``GMX_DISABLE_GPU_TIMING``.
147 times all code during runs. Incompatible with threads.
149 ``GMX_CYCLE_BARRIER``
150 calls MPI_Barrier before each cycle start/stop call.
153 build domain decomposition cells in the order
154 (z, y, x) rather than the default (x, y, z).
156 ``GMX_DD_USE_SENDRECV2``
157 during constraint and vsite communication, use a pair
158 of ``MPI_Sendrecv`` calls instead of two simultaneous non-blocking calls
159 (default 0, meaning off). Might be faster on some MPI implementations.
161 ``GMX_DLB_BASED_ON_FLOPS``
162 do domain-decomposition dynamic load balancing based on flop count rather than
163 measured time elapsed (default 0, meaning off).
164 This makes the load balancing reproducible, which can be useful for debugging purposes.
165 A value of 1 uses the flops; a value > 1 adds (value - 1)*5% of noise to the flops to increase the imbalance and the scaling.
167 ``GMX_DLB_MAX_BOX_SCALING``
168 maximum percentage box scaling permitted per domain-decomposition
169 load-balancing step (default 10)
171 ``GMX_DD_RECORD_LOAD``
172 record DD load statistics for reporting at end of the run (default 1, meaning on)
174 ``GMX_DETAILED_PERF_STATS``
175 when set, print slightly more detailed performance information
176 to the :ref:`log` file. The resulting output is the way performance summary is reported in versions
177 4.5.x and thus may be useful for anyone using scripts to parse :ref:`log` files or standard output.
179 ``GMX_DISABLE_SIMD_KERNELS``
180 disables architecture-specific SIMD-optimized (SSE2, SSE4.1, AVX, etc.)
181 non-bonded kernels thus forcing the use of plain C kernels.
183 ``GMX_DISABLE_GPU_TIMING``
184 timing of asynchronously executed GPU operations can have a
185 non-negligible overhead with short step times. Disabling timing can improve performance in these cases.
187 ``GMX_DISABLE_GPU_DETECTION``
188 when set, disables GPU detection even if :ref:`gmx mdrun` was compiled
191 ``GMX_GPU_APPLICATION_CLOCKS``
192 setting this variable to a value of "0", "ON", or "DISABLE" (case insensitive)
193 allows disabling the CUDA GPU allication clock support.
195 ``GMX_DISRE_ENSEMBLE_SIZE``
196 the number of systems for distance restraint ensemble
197 averaging. Takes an integer value.
200 emulate GPU runs by using algorithmically equivalent CPU reference code instead of
201 GPU-accelerated functions. As the CPU code is slow, it is intended to be used only for debugging purposes.
204 disable exiting upon encountering a corrupted frame in an :ref:`edr`
205 file, allowing the use of all frames up until the corruption.
208 update forces when invoking ``mdrun -rerun``.
211 set in the same way as ``mdrun -gpu_id``, ``GMX_GPU_ID``
212 allows the user to specify different GPU id-s, which can be useful for selecting different
213 devices on different compute nodes in a cluster. Cannot be used in conjunction with ``mdrun -gpu_id``.
215 ``GMX_IGNORE_FSYNC_FAILURE_ENV``
216 allow :ref:`gmx mdrun` to continue even if
220 when set to a floating-point value, overrides the default tolerance of
221 1e-5 for force-field floating-point parameters.
223 ``GMX_MAXCONSTRWARN``
224 if set to -1, :ref:`gmx mdrun` will
225 not exit if it produces too many LINCS warnings.
228 use the generic C kernel. Should be set if using
229 the group-based cutoff scheme and also sets ``GMX_NO_SOLV_OPT`` to be true,
230 thus disabling solvent optimizations as well.
233 neighbor list balancing parameter used when running on GPU. Sets the
234 target minimum number pair-lists in order to improve multi-processor load-balance for better
235 performance with small simulation systems. Must be set to a non-negative integer,
236 the 0 value disables list splitting.
237 The default value is optimized for supported GPUs (NVIDIA Fermi to Maxwell),
238 therefore changing it is not necessary for normal usage, but it can be useful on future architectures.
241 use neighbor list and kernels based on charge groups.
244 when set, print detailed neighbor search cycle counting.
246 ``GMX_NBNXN_EWALD_ANALYTICAL``
247 force the use of analytical Ewald non-bonded kernels,
248 mutually exclusive of ``GMX_NBNXN_EWALD_TABLE``.
250 ``GMX_NBNXN_EWALD_TABLE``
251 force the use of tabulated Ewald non-bonded kernels,
252 mutually exclusive of ``GMX_NBNXN_EWALD_ANALYTICAL``.
254 ``GMX_NBNXN_SIMD_2XNN``
255 force the use of 2x(N+N) SIMD CPU non-bonded kernels,
256 mutually exclusive of ``GMX_NBNXN_SIMD_4XN``.
258 ``GMX_NBNXN_SIMD_4XN``
259 force the use of 4xN SIMD CPU non-bonded kernels,
260 mutually exclusive of ``GMX_NBNXN_SIMD_2XNN``.
263 disables optimized all-vs-all kernels.
265 ``GMX_NO_CART_REORDER``
266 used in initializing domain decomposition communicators. Rank reordering
267 is default, but can be switched off with this environment variable.
269 ``GMX_NO_LJ_COMB_RULE``
270 force the use of LJ paremeter lookup instead of using combination rules
271 in the non-bonded kernels.
273 ``GMX_NO_CUDA_STREAMSYNC``
274 the opposite of ``GMX_CUDA_STREAMSYNC``. Disables the use of the
275 standard cudaStreamSynchronize-based GPU waiting to improve performance when using CUDA driver API
276 ealier than v5.0 with ECC-enabled GPUs.
278 ``GMX_NO_INT``, ``GMX_NO_TERM``, ``GMX_NO_USR1``
279 disable signal handlers for SIGINT,
280 SIGTERM, and SIGUSR1, respectively.
283 do not use separate inter- and intra-node communicators.
286 skip non-bonded calculations; can be used to estimate the possible
287 performance gain from adding a GPU accelerator to the current hardware setup -- assuming that this is
288 fast enough to complete the non-bonded calculations while the CPU does bonded force and PME computation.
289 Freezing the particles will be required to stop the system blowing up.
292 when set, do not add virial contribution to COM pull forces.
295 shell positions are not predicted.
298 turns off solvent optimizations; automatic if ``GMX_NB_GENERIC``
302 the ideal number of charge groups per neighbor searching grid cell is hard-coded
303 to a value of 10. Setting this environment variable to any other integer value overrides this hard-coded
307 set the number of OpenMP or PME threads (overrides the number guessed by
311 use P3M-optimized influence function instead of smooth PME B-spline interpolation.
313 ``GMX_PME_THREAD_DIVISION``
314 PME thread division in the format "x y z" for all three dimensions. The
315 sum of the threads in each dimension must equal the total number of PME threads (set in
319 if the number of domain decomposition cells is set to 1 for both x and y,
320 decompose PME in one dimension.
322 ``GMX_REQUIRE_SHELL_INIT``
323 require that shell positions are initiated.
325 ``GMX_REQUIRE_TABLES``
326 require the use of tabulated Coulombic
327 and van der Waals interactions.
330 the minimum value for soft-core sigma. **Note** that this value is set
331 using the :mdp:`sc-sigma` keyword in the :ref:`mdp` file, but this environment variable can be used
332 to reproduce pre-4.5 behavior with respect to this parameter.
335 should contain multiple masses used for test particle insertion into a cavity.
336 The center of mass of the last atoms is used for insertion into the cavity.
339 use graph for bonded interactions.
341 ``GMX_VERLET_BUFFER_RES``
342 resolution of buffer size in Verlet cutoff scheme. The default value is
343 0.001, but can be overridden with this environment variable.
346 Not strictly a |Gromacs| environment variable, but on large machines
347 the hwloc detection can take a few seconds if you have lots of MPI processes.
348 If you run the hwloc command `lstopo out.xml` and set this environment
349 variable to point to the location of this file, the hwloc library will use
350 the cached information instead, which can be faster.
353 the ``mpirun`` command used by :ref:`gmx tune_pme`.
356 the :ref:`gmx mdrun` command used by :ref:`gmx tune_pme`.
358 ``GMX_DISABLE_DYNAMICPRUNING``
359 disables dynamic pair-list pruning. Note that :ref:`gmx mdrun` will
360 still tune nstlist to the optimal value picked assuming dynamic pruning. Thus
361 for good performance the -nstlist option should be used.
363 ``GMX_NSTLIST_DYNAMICPRUNING``
364 overrides the dynamic pair-list pruning interval chosen heuristically
365 by mdrun. Values should be between the pruning frequency value
366 (1 for CPU and 2 for GPU) and :mdp:`nstlist` ``- 1``.
368 ``GMX_USE_TREEREDUCE``
369 use tree reduction for nbnxn force reduction. Potentially faster for large number of
370 OpenMP threads (if memory locality is important).
372 .. _opencl-management:
376 Currently, several environment variables exist that help customize some aspects
377 of the OpenCL_ version of |Gromacs|. They are mostly related to the runtime
378 compilation of OpenCL kernels, but they are also used in device selection.
380 ``GMX_OCL_NOGENCACHE``
381 If set, disable caching for OpenCL kernel builds. Caching is
382 normally useful so that future runs can re-use the compiled
383 kernels from previous runs. Currently, caching is always
384 disabled, until we solve concurrency issues.
387 Enable OpenCL binary caching. Only intended to be used for
388 development and (expert) testing as neither concurrency
389 nor cache invalidation is implemented safely!
391 ``GMX_OCL_NOFASTGEN``
392 If set, generate and compile all algorithm flavors, otherwise
393 only the flavor required for the simulation is generated and
396 ``GMX_OCL_DISABLE_FASTMATH``
397 Prevents the use of ``-cl-fast-relaxed-math`` compiler option.
400 If defined, the OpenCL build log is always written to the
401 mdrun log file. Otherwise, the build log is written to the
402 log file only when an error occurs.
405 If defined, it enables verbose mode for OpenCL kernel build.
406 Currently available only for NVIDIA GPUs. See ``GMX_OCL_DUMP_LOG``
407 for details about how to obtain the OpenCL build log.
409 ``GMX_OCL_DUMP_INTERM_FILES``
411 If defined, intermediate language code corresponding to the
412 OpenCL build process is saved to file. Caching has to be
413 turned off in order for this option to take effect (see
414 ``GMX_OCL_NOGENCACHE``).
416 - NVIDIA GPUs: PTX code is saved in the current directory
417 with the name ``device_name.ptx``
418 - AMD GPUs: ``.IL/.ISA`` files will be created for each OpenCL
419 kernel built. For details about where these files are
420 created check AMD documentation for ``-save-temps`` compiler
424 Use in conjunction with ``OCL_FORCE_CPU`` or with an AMD device.
425 It adds the debug flag to the compiler options (-g).
428 Disable optimisations. Adds the option ``cl-opt-disable`` to the
431 ``GMX_OCL_FORCE_CPU``
432 Force the selection of a CPU device instead of a GPU. This
433 exists only for debugging purposes. Do not expect |Gromacs| to
434 function properly with this option on, it is solely for the
435 simplicity of stepping in a kernel and see what is happening.
437 ``GMX_OCL_DISABLE_I_PREFETCH``
438 Disables i-atom data (type or LJ parameter) prefetch allowig
441 ``GMX_OCL_ENABLE_I_PREFETCH``
442 Enables i-atom data (type or LJ parameter) prefetch allowig
443 testing on platforms where this behavior is not default.
445 ``GMX_OCL_NB_ANA_EWALD``
446 Forces the use of analytical Ewald kernels. Equivalent of
447 CUDA environment variable ``GMX_CUDA_NB_ANA_EWALD``
449 ``GMX_OCL_NB_TAB_EWALD``
450 Forces the use of tabulated Ewald kernel. Equivalent
451 of CUDA environment variable ``GMX_OCL_NB_TAB_EWALD``
453 ``GMX_OCL_NB_EWALD_TWINCUT``
454 Forces the use of twin-range cutoff kernel. Equivalent of
455 CUDA environment variable ``GMX_CUDA_NB_EWALD_TWINCUT``
457 ``GMX_DISABLE_OCL_TIMING``
458 Disables timing for OpenCL operations
460 ``GMX_OCL_FILE_PATH``
461 Use this parameter to force |Gromacs| to load the OpenCL
462 kernels from a custom location. Use it only if you want to
463 override |Gromacs| default behavior, or if you want to test
466 ``GMX_OCL_DISABLE_COMPATIBILITY_CHECK``
467 Disables the hardware compatibility check. Useful for developers
468 and allows testing the OpenCL kernels on non-supported platforms
469 (like Intel iGPUs) without source code modification.
471 Analysis and Core Functions
472 ---------------------------
474 accuracy in Gaussian L510 (MC-SCF) component program.
476 ``GMX_QM_ORCA_BASENAME``
477 prefix of :ref:`tpr` files, used in Orca calculations
478 for input and output file names.
481 when set to a nonzero value, Gaussian QM calculations will
482 iteratively solve the CP-MCSCF equations.
484 ``GMX_QM_MODIFIED_LINKS_DIR``
485 location of modified links in Gaussian.
488 used by :ref:`gmx do_dssp` to point to the ``dssp``
489 executable (not just its path).
492 directory where Gaussian is installed.
495 name of the Gaussian executable.
497 ``GMX_DIPOLE_SPACING``
498 spacing used by :ref:`gmx dipoles`.
501 sets the maximum number of residues to be renumbered by
502 :ref:`gmx grompp`. A value of -1 indicates all residues should be renumbered.
504 ``GMX_FFRTP_TER_RENAME``
505 Some force fields (like AMBER) use specific names for N- and C-
506 terminal residues (NXXX and CXXX) as :ref:`rtp` entries that are normally renamed. Setting
507 this environment variable disables this renaming.
510 ``gunzip`` executable, used by :ref:`gmx wham`.
513 name of X11 font used by :ref:`gmx view`.
516 the time unit used in output files, can be
517 anything in fs, ps, ns, us, ms, s, m or h.
519 ``GMX_QM_GAUSSIAN_MEMORY``
520 memory used for Gaussian QM calculation.
523 name of the ``multiprot`` executable, used by the
524 contributed program ``do_multiprot``.
527 number of CPUs to be used for Gaussian QM calculation
530 directory where Orca is installed.
533 simulated annealing step size for Gaussian QM calculation.
535 ``GMX_QM_GROUND_STATE``
536 defines state for Gaussian surface hopping calculation.
539 name of the ``total`` executable used by the contributed
540 ``do_shift`` program.
543 make :ref:`gmx energy` and :ref:`gmx eneconv`
547 where to find VMD plug-ins. Needed to be
548 able to read file formats recognized only by a VMD plug-in.
551 base path of VMD installation.
554 sets viewer to ``xmgr`` (deprecated) instead of ``xmgrace``.