| blob | ac78d8f0dc7c8c2c3687507aaaad2f8bd084475f |
1 Environment Variables
2 =====================
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.
17 Output Control
18 --------------
19 ``GMX_CONSTRAINTVIR``
20 Print constraint virial and force virial energy terms.
22 ``GMX_MAXBACKUP``
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.
30 ``GMX_NO_QUOTES``
31 if this is explicitly set, no cool quotes
32 will be printed at the end of a program.
34 ``GMX_SUPPRESS_DUMP``
35 prevent dumping of step files during
36 (for example) blowing up during failure of constraint
37 algorithms.
39 ``GMX_TPI_DUMP``
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.
44 ``GMX_VIEW_XPM``
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
58 ``GMX_LOG_BUFFER``
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.
64 ``GMX_LOGO_COLOR``
65 set display color for logo in :ref:`gmx view`.
67 ``GMX_PRINT_LONGFORMAT``
68 use long float format when printing
69 decimal values.
71 ``GMX_COMPELDUMP``
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 Debugging
78 ---------
79 ``GMX_PRINT_DEBUG_LINES``
80 when set, print debugging info on line numbers.
82 ``GMX_DD_NST_DUMP``
83 number of steps that elapse between dumping
84 the current DD to a PDB file (default 0). This only takes effect
85 during domain decomposition, so it should typically be
86 0 (never), 1 (every DD phase) or a multiple of :mdp:`nstlist`.
88 ``GMX_DD_NST_DUMP_GRID``
89 number of steps that elapse between dumping
90 the current DD grid to a PDB file (default 0). This only takes effect
91 during domain decomposition, so it should typically be
92 0 (never), 1 (every DD phase) or a multiple of :mdp:`nstlist`.
94 ``GMX_DD_DEBUG``
95 general debugging trigger for every domain
96 decomposition (default 0, meaning off). Currently only checks
97 global-local atom index mapping for consistency.
99 ``GMX_DD_NPULSE``
100 over-ride the number of DD pulses used
101 (default 0, meaning no over-ride). Normally 1 or 2.
103 There are a number of extra environment variables like these
104 that are used in debugging - check the code!
106 Performance and Run Control
107 ---------------------------
108 ``GMX_DO_GALACTIC_DYNAMICS``
109 planetary simulations are made possible (just for fun) by setting
110 this environment variable, which allows setting :mdp:`epsilon-r` to -1 in the :ref:`mdp`
111 file. Normally, :mdp:`epsilon-r` must be greater than zero to prevent a fatal error.
112 See webpage_ for example input files for a planetary simulation.
114 ``GMX_ALLOW_CPT_MISMATCH``
115 when set, runs will not exit if the
116 ensemble set in the :ref:`tpr` file does not match that of the
117 :ref:`cpt` file.
119 ``GMX_CUDA_NB_EWALD_TWINCUT``
120 force the use of twin-range cutoff kernel even if :mdp:`rvdw` equals
121 :mdp:`rcoulomb` after PP-PME load balancing. The switch to twin-range kernels is automated,
122 so this variable should be used only for benchmarking.
124 ``GMX_CUDA_NB_ANA_EWALD``
125 force the use of analytical Ewald kernels. Should be used only for benchmarking.
127 ``GMX_CUDA_NB_TAB_EWALD``
128 force the use of tabulated Ewald kernels. Should be used only for benchmarking.
130 ``GMX_CUDA_STREAMSYNC``
131 force the use of cudaStreamSynchronize on ECC-enabled GPUs, which leads
132 to performance loss due to a known CUDA driver bug present in API v5.0 NVIDIA drivers (pre-30x.xx).
133 Cannot be set simultaneously with ``GMX_NO_CUDA_STREAMSYNC``.
135 ``GMX_DISABLE_CUDALAUNCH``
136 disable the use of the lower-latency cudaLaunchKernel API even when supported (CUDA >=v7.0).
137 Should only be used for benchmarking purposes.
139 ``GMX_DISABLE_CUDA_TIMING``
140 Disables GPU timing of CUDA tasks; synonymous with ``GMX_DISABLE_GPU_TIMING``.
142 ``GMX_CYCLE_ALL``
143 times all code during runs. Incompatible with threads.
145 ``GMX_CYCLE_BARRIER``
146 calls MPI_Barrier before each cycle start/stop call.
148 ``GMX_DD_ORDER_ZYX``
149 build domain decomposition cells in the order
150 (z, y, x) rather than the default (x, y, z).
152 ``GMX_DD_USE_SENDRECV2``
153 during constraint and vsite communication, use a pair
154 of ``MPI_Sendrecv`` calls instead of two simultaneous non-blocking calls
155 (default 0, meaning off). Might be faster on some MPI implementations.
157 ``GMX_DLB_BASED_ON_FLOPS``
158 do domain-decomposition dynamic load balancing based on flop count rather than
159 measured time elapsed (default 0, meaning off).
160 This makes the load balancing reproducible, which can be useful for debugging purposes.
161 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.
163 ``GMX_DLB_MAX_BOX_SCALING``
164 maximum percentage box scaling permitted per domain-decomposition
165 load-balancing step (default 10)
167 ``GMX_DD_RECORD_LOAD``
168 record DD load statistics for reporting at end of the run (default 1, meaning on)
170 ``GMX_DETAILED_PERF_STATS``
171 when set, print slightly more detailed performance information
172 to the :ref:`log` file. The resulting output is the way performance summary is reported in versions
173 4.5.x and thus may be useful for anyone using scripts to parse :ref:`log` files or standard output.
175 ``GMX_DISABLE_SIMD_KERNELS``
176 disables architecture-specific SIMD-optimized (SSE2, SSE4.1, AVX, etc.)
177 non-bonded kernels thus forcing the use of plain C kernels.
179 ``GMX_DISABLE_GPU_TIMING``
180 timing of asynchronously executed GPU operations can have a
181 non-negligible overhead with short step times. Disabling timing can improve performance in these cases.
183 ``GMX_DISABLE_GPU_DETECTION``
184 when set, disables GPU detection even if :ref:`gmx mdrun` was compiled
185 with GPU support.
187 ``GMX_GPU_APPLICATION_CLOCKS``
188 setting this variable to a value of "0", "ON", or "DISABLE" (case insensitive)
189 allows disabling the CUDA GPU allication clock support.
191 ``GMX_DISRE_ENSEMBLE_SIZE``
192 the number of systems for distance restraint ensemble
193 averaging. Takes an integer value.
195 ``GMX_EMULATE_GPU``
196 emulate GPU runs by using algorithmically equivalent CPU reference code instead of
197 GPU-accelerated functions. As the CPU code is slow, it is intended to be used only for debugging purposes.
199 ``GMX_ENX_NO_FATAL``
200 disable exiting upon encountering a corrupted frame in an :ref:`edr`
201 file, allowing the use of all frames up until the corruption.
203 ``GMX_FORCE_UPDATE``
204 update forces when invoking ``mdrun -rerun``.
206 ``GMX_GPU_ID``
207 set in the same way as ``mdrun -gpu_id``, ``GMX_GPU_ID``
208 allows the user to specify different GPU id-s, which can be useful for selecting different
209 devices on different compute nodes in a cluster. Cannot be used in conjunction with ``mdrun -gpu_id``.
211 ``GMX_IGNORE_FSYNC_FAILURE_ENV``
212 allow :ref:`gmx mdrun` to continue even if
213 a file is missing.
215 ``GMX_LJCOMB_TOL``
216 when set to a floating-point value, overrides the default tolerance of
217 1e-5 for force-field floating-point parameters.
219 ``GMX_MAXCONSTRWARN``
220 if set to -1, :ref:`gmx mdrun` will
221 not exit if it produces too many LINCS warnings.
223 ``GMX_NB_GENERIC``
224 use the generic C kernel. Should be set if using
225 the group-based cutoff scheme and also sets ``GMX_NO_SOLV_OPT`` to be true,
226 thus disabling solvent optimizations as well.
228 ``GMX_NB_MIN_CI``
229 neighbor list balancing parameter used when running on GPU. Sets the
230 target minimum number pair-lists in order to improve multi-processor load-balance for better
231 performance with small simulation systems. Must be set to a non-negative integer,
232 the 0 value disables list splitting.
233 The default value is optimized for supported GPUs (NVIDIA Fermi to Maxwell),
234 therefore changing it is not necessary for normal usage, but it can be useful on future architectures.
236 ``GMX_NBLISTCG``
237 use neighbor list and kernels based on charge groups.
239 ``GMX_NBNXN_CYCLE``
240 when set, print detailed neighbor search cycle counting.
242 ``GMX_NBNXN_EWALD_ANALYTICAL``
243 force the use of analytical Ewald non-bonded kernels,
244 mutually exclusive of ``GMX_NBNXN_EWALD_TABLE``.
246 ``GMX_NBNXN_EWALD_TABLE``
247 force the use of tabulated Ewald non-bonded kernels,
248 mutually exclusive of ``GMX_NBNXN_EWALD_ANALYTICAL``.
250 ``GMX_NBNXN_SIMD_2XNN``
251 force the use of 2x(N+N) SIMD CPU non-bonded kernels,
252 mutually exclusive of ``GMX_NBNXN_SIMD_4XN``.
254 ``GMX_NBNXN_SIMD_4XN``
255 force the use of 4xN SIMD CPU non-bonded kernels,
256 mutually exclusive of ``GMX_NBNXN_SIMD_2XNN``.
258 ``GMX_NO_ALLVSALL``
259 disables optimized all-vs-all kernels.
261 ``GMX_NO_CART_REORDER``
262 used in initializing domain decomposition communicators. Rank reordering
263 is default, but can be switched off with this environment variable.
265 ``GMX_NO_LJ_COMB_RULE``
266 force the use of LJ paremeter lookup instead of using combination rules
267 in the non-bonded kernels.
269 ``GMX_NO_CUDA_STREAMSYNC``
270 the opposite of ``GMX_CUDA_STREAMSYNC``. Disables the use of the
271 standard cudaStreamSynchronize-based GPU waiting to improve performance when using CUDA driver API
272 ealier than v5.0 with ECC-enabled GPUs.
274 ``GMX_NO_INT``, ``GMX_NO_TERM``, ``GMX_NO_USR1``
275 disable signal handlers for SIGINT,
276 SIGTERM, and SIGUSR1, respectively.
278 ``GMX_NO_NODECOMM``
279 do not use separate inter- and intra-node communicators.
281 ``GMX_NO_NONBONDED``
282 skip non-bonded calculations; can be used to estimate the possible
283 performance gain from adding a GPU accelerator to the current hardware setup -- assuming that this is
284 fast enough to complete the non-bonded calculations while the CPU does bonded force and PME computation.
285 Freezing the particles will be required to stop the system blowing up.
287 ``GMX_NO_PULLVIR``
288 when set, do not add virial contribution to COM pull forces.
290 ``GMX_NOPREDICT``
291 shell positions are not predicted.
293 ``GMX_NO_SOLV_OPT``
294 turns off solvent optimizations; automatic if ``GMX_NB_GENERIC``
295 is enabled.
297 ``GMX_NSCELL_NCG``
298 the ideal number of charge groups per neighbor searching grid cell is hard-coded
299 to a value of 10. Setting this environment variable to any other integer value overrides this hard-coded
300 value.
302 ``GMX_PME_NTHREADS``
303 set the number of OpenMP or PME threads (overrides the number guessed by
304 :ref:`gmx mdrun`.
306 ``GMX_PME_P3M``
307 use P3M-optimized influence function instead of smooth PME B-spline interpolation.
309 ``GMX_PME_THREAD_DIVISION``
310 PME thread division in the format "x y z" for all three dimensions. The
311 sum of the threads in each dimension must equal the total number of PME threads (set in
312 `GMX_PME_NTHREADS`).
314 ``GMX_PMEONEDD``
315 if the number of domain decomposition cells is set to 1 for both x and y,
316 decompose PME in one dimension.
318 ``GMX_REQUIRE_SHELL_INIT``
319 require that shell positions are initiated.
321 ``GMX_REQUIRE_TABLES``
322 require the use of tabulated Coulombic
323 and van der Waals interactions.
325 ``GMX_SCSIGMA_MIN``
326 the minimum value for soft-core sigma. **Note** that this value is set
327 using the :mdp:`sc-sigma` keyword in the :ref:`mdp` file, but this environment variable can be used
328 to reproduce pre-4.5 behavior with respect to this parameter.
330 ``GMX_TPIC_MASSES``
331 should contain multiple masses used for test particle insertion into a cavity.
332 The center of mass of the last atoms is used for insertion into the cavity.
334 ``GMX_USE_GRAPH``
335 use graph for bonded interactions.
337 ``GMX_VERLET_BUFFER_RES``
338 resolution of buffer size in Verlet cutoff scheme. The default value is
339 0.001, but can be overridden with this environment variable.
341 ``HWLOC_XMLFILE``
342 Not strictly a |Gromacs| environment variable, but on large machines
343 the hwloc detection can take a few seconds if you have lots of MPI processes.
344 If you run the hwloc command `lstopo out.xml` and set this environment
345 variable to point to the location of this file, the hwloc library will use
346 the cached information instead, which can be faster.
348 ``MPIRUN``
349 the ``mpirun`` command used by :ref:`gmx tune_pme`.
351 ``MDRUN``
352 the :ref:`gmx mdrun` command used by :ref:`gmx tune_pme`.
354 ``GMX_DISABLE DYNAMICPRUNING``
355 disables dynamic pair-list pruning. Note that :ref:`gmx mdrun` will
356 still tune nstlist to the optimal value with dynamic pruning. Thus
357 for good performance the -nstlist option should be used.
359 ``GMX_NSTLIST_DYNAMICPRUNING``
360 overrides the dynamic pair-list pruning interval chosen heuristically
361 by mdrun. Values should be between the pruning frequency value
362 (1 for CPU and 2 for GPU) and :mdp:`nstlist - 1`.
364 ``GMX_USE_TREEREDUCE``
365 use tree reduction for nbnxn force reduction. Potentially faster for large number of
366 OpenMP threads (if memory locality is important).
368 .. _opencl-management:
370 OpenCL management
371 -----------------
372 Currently, several environment variables exist that help customize some aspects
373 of the OpenCL_ version of |Gromacs|. They are mostly related to the runtime
374 compilation of OpenCL kernels, but they are also used in device selection.
376 ``GMX_OCL_NOGENCACHE``
377 If set, disable caching for OpenCL kernel builds. Caching is
378 normally useful so that future runs can re-use the compiled
379 kernels from previous runs. Currently, caching is always
380 disabled, until we solve concurrency issues.
382 ``GMX_OCL_GENCACHE``
383 Enable OpenCL binary caching. Only intended to be used for
384 development and (expert) testing as neither concurrency
385 nor cache invalidation is implemented safely!
387 ``GMX_OCL_NOFASTGEN``
388 If set, generate and compile all algorithm flavors, otherwise
389 only the flavor required for the simulation is generated and
390 compiled.
392 ``GMX_OCL_DISABLE_FASTMATH``
393 Prevents the use of ``-cl-fast-relaxed-math`` compiler option.
395 ``GMX_OCL_DUMP_LOG``
396 If defined, the OpenCL build log is always written to the
397 mdrun log file. Otherwise, the build log is written to the
398 log file only when an error occurs.
400 ``GMX_OCL_VERBOSE``
401 If defined, it enables verbose mode for OpenCL kernel build.
402 Currently available only for NVIDIA GPUs. See ``GMX_OCL_DUMP_LOG``
403 for details about how to obtain the OpenCL build log.
405 ``GMX_OCL_DUMP_INTERM_FILES``
407 If defined, intermediate language code corresponding to the
408 OpenCL build process is saved to file. Caching has to be
409 turned off in order for this option to take effect (see
410 ``GMX_OCL_NOGENCACHE``).
412 - NVIDIA GPUs: PTX code is saved in the current directory
413 with the name ``device_name.ptx``
414 - AMD GPUs: ``.IL/.ISA`` files will be created for each OpenCL
415 kernel built. For details about where these files are
416 created check AMD documentation for ``-save-temps`` compiler
417 option.
419 ``GMX_OCL_DEBUG``
420 Use in conjunction with ``OCL_FORCE_CPU`` or with an AMD device.
421 It adds the debug flag to the compiler options (-g).
423 ``GMX_OCL_NOOPT``
424 Disable optimisations. Adds the option ``cl-opt-disable`` to the
425 compiler options.
427 ``GMX_OCL_FORCE_CPU``
428 Force the selection of a CPU device instead of a GPU. This
429 exists only for debugging purposes. Do not expect |Gromacs| to
430 function properly with this option on, it is solely for the
431 simplicity of stepping in a kernel and see what is happening.
433 ``GMX_OCL_DISABLE_I_PREFETCH``
434 Disables i-atom data (type or LJ parameter) prefetch allowig
435 testing.
437 ``GMX_OCL_ENABLE_I_PREFETCH``
438 Enables i-atom data (type or LJ parameter) prefetch allowig
439 testing on platforms where this behavior is not default.
441 ``GMX_OCL_NB_ANA_EWALD``
442 Forces the use of analytical Ewald kernels. Equivalent of
443 CUDA environment variable ``GMX_CUDA_NB_ANA_EWALD``
445 ``GMX_OCL_NB_TAB_EWALD``
446 Forces the use of tabulated Ewald kernel. Equivalent
447 of CUDA environment variable ``GMX_OCL_NB_TAB_EWALD``
449 ``GMX_OCL_NB_EWALD_TWINCUT``
450 Forces the use of twin-range cutoff kernel. Equivalent of
451 CUDA environment variable ``GMX_CUDA_NB_EWALD_TWINCUT``
453 ``GMX_DISABLE_OCL_TIMING``
454 Disables timing for OpenCL operations
456 ``GMX_OCL_FILE_PATH``
457 Use this parameter to force |Gromacs| to load the OpenCL
458 kernels from a custom location. Use it only if you want to
459 override |Gromacs| default behavior, or if you want to test
460 your own kernels.
462 ``GMX_OCL_DISABLE_COMPATIBILITY_CHECK``
463 Disables the hardware compatibility check. Useful for developers
464 and allows testing the OpenCL kernels on non-supported platforms
465 (like Intel iGPUs) without source code modification.
467 Analysis and Core Functions
468 ---------------------------
469 ``GMX_QM_ACCURACY``
470 accuracy in Gaussian L510 (MC-SCF) component program.
472 ``GMX_QM_ORCA_BASENAME``
473 prefix of :ref:`tpr` files, used in Orca calculations
474 for input and output file names.
476 ``GMX_QM_CPMCSCF``
477 when set to a nonzero value, Gaussian QM calculations will
478 iteratively solve the CP-MCSCF equations.
480 ``GMX_QM_MODIFIED_LINKS_DIR``
481 location of modified links in Gaussian.
483 ``DSSP``
484 used by :ref:`gmx do_dssp` to point to the ``dssp``
485 executable (not just its path).
487 ``GMX_QM_GAUSS_DIR``
488 directory where Gaussian is installed.
490 ``GMX_QM_GAUSS_EXE``
491 name of the Gaussian executable.
493 ``GMX_DIPOLE_SPACING``
494 spacing used by :ref:`gmx dipoles`.
496 ``GMX_MAXRESRENUM``
497 sets the maximum number of residues to be renumbered by
498 :ref:`gmx grompp`. A value of -1 indicates all residues should be renumbered.
500 ``GMX_FFRTP_TER_RENAME``
501 Some force fields (like AMBER) use specific names for N- and C-
502 terminal residues (NXXX and CXXX) as :ref:`rtp` entries that are normally renamed. Setting
503 this environment variable disables this renaming.
505 ``GMX_PATH_GZIP``
506 ``gunzip`` executable, used by :ref:`gmx wham`.
508 ``GMX_FONT``
509 name of X11 font used by :ref:`gmx view`.
511 ``GMXTIMEUNIT``
512 the time unit used in output files, can be
513 anything in fs, ps, ns, us, ms, s, m or h.
515 ``GMX_QM_GAUSSIAN_MEMORY``
516 memory used for Gaussian QM calculation.
518 ``MULTIPROT``
519 name of the ``multiprot`` executable, used by the
520 contributed program ``do_multiprot``.
522 ``NCPUS``
523 number of CPUs to be used for Gaussian QM calculation
525 ``GMX_ORCA_PATH``
526 directory where Orca is installed.
528 ``GMX_QM_SA_STEP``
529 simulated annealing step size for Gaussian QM calculation.
531 ``GMX_QM_GROUND_STATE``
532 defines state for Gaussian surface hopping calculation.
534 ``GMX_TOTAL``
535 name of the ``total`` executable used by the contributed
536 ``do_shift`` program.
538 ``GMX_ENER_VERBOSE``
539 make :ref:`gmx energy` and :ref:`gmx eneconv`
540 loud and noisy.
542 ``VMD_PLUGIN_PATH``
543 where to find VMD plug-ins. Needed to be
544 able to read file formats recognized only by a VMD plug-in.
546 ``VMDDIR``
547 base path of VMD installation.
549 ``GMX_USE_XMGR``
550 sets viewer to ``xmgr`` (deprecated) instead of ``xmgrace``.