-
I have tried to compile the `devel` version of QPhiX on JURECA (Haswell) and Marconi A2 (KNL). Since I use the same compilation script, it should still work. I get the following error message on both …
-
Having to wrap for variable soalen is also really annoying. I thought that you had simply added code to the tests which instantiated all currently useful soalen and appropriate switch statements (or o…
-
Currently there seems to be an issue if Python libraries are in a nonstandard place. An example is NERSC, where the python3 environment is an ANACONDA environment created by 'conda'.
In this instance…
-
Allow control of frequency of group barriers at configure time. This could then become an autotuning parameter.
-
Ideally one has `M^\dagger M + \rho^2` where `\rho` is the preconditioning mass. The solvers in QPhiX apply `M` and then `M^\dagger` in two steps. This `\rho` should not live within the solver but rat…
-
# Martin
When one calls `make -j`, it will start a job server and then other `make` processes can inherit this and coordinate the number of processes running. By explicitly giving the option `-j` o…
-
I have been meaning to put this for a while. We should reduce the verbosity from the Geometry Constructor. In an HMC run it just prints too much...
-
Whenever I change a single test case, I need around 100 seconds of compilation time. Changing something in the library (like `geometry.h`) will result in over five minutes of recompilation.
I wonde…
-
It seems that the output directories are not created. If they don't exist, the codegen does not fail, it just does not creates anything. Then subsequent parts of the build fail because the files do no…
-
QPhiX currently needs QDP++ to build its tests. These tests are not needed for production use, so that should be an optional dependency.
In order to use QPhiX in tmLQCD, this dependency should be c…