Ye Luo
Test this please
It needs to be an option not on by default. nvcc checks host compiler for good reason. Host compiler may not have compatible header files. More likely in C++. I...
changed my view and implemented in #3957
This input parameter is painful to control. Just like walkers, we don't control walkers per crowd but only `walker_per_rank`.
`walkers_per_rank=1792`, 7 threads (7 crowds) in VMC [opt-NiO-fcc-S4.batch1.txt](https://github.com/QMCPACK/qmcpack/files/8101139/opt-NiO-fcc-S4.batch1.txt) ``` Reading configurations from h5FileRoot Using Nonlocal PP in Opt: no VMC Eavg = -1.4151e+03 VMC Evar = 4.2034e+02 Total weights =...
A second thought on this. If we directly borrow the crowds from VMC, we neither need to care about input nor concern about additional memory usage by WFOpt.
In the current implementation, WFOpt batched driver owns it own set of walker objects (particle set, TWF, Ham) which all cost memory. They are not shared with the VMC owned...
In CostFunctionCrowdData ``` // List of objects for use in flex_* calls UPtrVector wf_ptr_list_; UPtrVector p_ptr_list_; UPtrVector h_ptr_list_; UPtrVector h0_ptr_list_; ```
> Trying to reuse memory between Crowd and CostFunctionCrowdData seems complicated - either they need to be combined as classes, which seems complicated. Or the walkers need to get moved...
In addition, you also need to delete all the walker object in CostFunctionCrowdData before creating the VMC object to keep the memory minimal.