DCA
DCA copied to clipboard
CompFUSE
DCA++
If there is a system installe hdf5 but you would like to run a different one you must set both -DHDF5_ROOT and -DHDF5_DIR on Centos7 to avoid picking up the...
`SpaceTransform2D::execute(func::function& f_input)` and `SpaceTransform2D::execute(func::function& f_input, func::function& f_output)` seem to use different conventions for where to put the normalization constant. See e.g. https://github.com/CompFUSE/DCA/blob/2583dba4fca20ce9181d345c3b62fbe3ce6cd82c/include/dca/math/function_transform/special_transforms/space_transform_2D.hpp#L95 @gbalduzz Can you clarify and document this in...
As commented by @biddisco, PR #43 should be implemented in a way that automatically decides the number of available processes.
GTest is tripping the following warnings in build. We should resolve these, they create noise in that CI channel that prevents really using the build warning feedback. ``` warning: converting...
The documentation of the parameters in and out of the kernels and the more opaque C-like mechanics need to be added. I'll add a checklist for kernels soon.
In light of coming work with our collaborators on G4 this code should be meticulously documented and the indice and memory layout documented. We should discuss how to best do...
There is a fair case of copy/paste repetition in the tp_accumulator_kernels.cu Do we have others in kernels as well?
If @biddisco managed to integrate a performance metric in the testing suite it might be a good idea to develop a performance test of the whole DCA algorithm. My question...
All copies between Vector and Matrix objects are handled by cudaMemcpy with cudaMemcpyDefault. This kills our chances to execute CPU code in parallel with GPU memory transfers and kernel execution....
In many places (see here for example https://github.com/CompFUSE/DCA/blob/12c6c1279139994e8eb1656806f4fbb235f49296/cmake/dca_cuda.cmake#L10-L12 ) vars are set and made CACHE vars. This is annoying because if you compile WITH cuda, then change the option later...
