Problems continuing DMC calculations

[braulio “braulio gabriel” gabriel]

I would like to know how to continue a DMC calculation after running a certain number of simulation blocks. I made an initial calculation using the input "Cu.dmc.in-wfj.xml" (attached). After this calculation finished I tried to continue using the generated file "Cu.s001.cont.xml". I made a modification to this file generating "Cu.dmc_cont.xml" (attached) to continue the DMC calculations. However the calculations continued but without considering the blocks previously run. I wanted to continue the calculations accumulating more DMC simulation blocks.

[Paul R. C. Kent]

There is not a clear example of this in the manual unfortunately -- https://qmcpack.readthedocs.io/en/develop/methods.html?highlight=continuing#listing-43 is close. We do have several restart tests with example input. See e.g. https://github.com/QMCPACK/qmcpack/tree/develop/tests/io and , say, compare the restart XML file with the regular one in https://github.com/QMCPACK/qmcpack/tree/develop/tests/io/restart_batch .

Note that QMCPACK will always generate a new section (a set of numbered output files) and not add output to an existing set. VMC will do an exact restart, but in DMC an exact restart is not currently possible because the DMC trial energy is recomputed. If this is critical let us know; the main thing is that the equilibrated walkers are properly continued. Use qmca to merge the output scalar.dat and dmc.dat of the different sections when analyzing them. If you truly need an exact restart please make your vote clear. It is one of those "nice to have" things, but other features currently appear to be more important.

-- Paul

[braulio “braulio gabriel” gabriel]

Thanks for the reply.