This is a set of regression tests for NAMD, starting with tests for alchemical perturbation, with and without the IDWS scheme.

Created using the layout and scripts of the Colvars regression tests. Credits to Giacomo Fiorin for much of the scripting.

cd Tests
./run_tests.sh [NAMD binary | namd2] [list of test directories to run | all tests]

Add option -g to generate reference outputs (overwriting current ones). The reference outputs include the log file, the energy lines from the log, and the alchemical output if any.

• Water: simulation of a small water box
• MD: plain MD without alchemical transformation
• alch: making a water molecule alchemically disappear
• ann: annihilation (alchDecouple off)
• dec: decoupling (alchDecouple on) - for a single water molecule the result should be the same.
• IDWS: Interleaved Double-Wide Sampling (specifying alchLambdaIDWS)
• noPME: not enabling PME, as is done in all other tests
• alchOutFreq: using alchOutFreq 12 as opposed to alchOutFreq 4 in the other tests
• CUDASOAint: enables the GPU-resident mode of namd3 (fails on builds that do not support that feature)
• GlobalMasterSync: enables a Tcl script that outputs messages every timestep to troubleshoot synchronization issues with the GlobalMaster mechanism used by TclForces and Colvars

cd Tests
mkdir T_<My_Test>
cp T_Water_MD/test.namd T_<My_Test>
# Edit T_<My_Test>/test.namd to cutomize test
./run_tests -g <reference NAMD2 binary> T_<My_Test>
# Check that reference results in T_<My_Test>/AutoDiff are sensible