Hello,

./TestCudaEwald

gives
Error: invalid argument launching kernel kFindAtomRangeForGrid

Jerome

Hi !

have you tried

cmake -i <SRC>

On my side, I had to correct and set up a lot of CUDA variables

hth,
Jerome

<t>Hello,<br/> <br/> I added the following option to cmake<br/> <br/> -D OPENMM_BUILD_CUDA_LIB=ON<br/> -D OPENMM_BUILD_FREE_ENERGY_CUDA_LIB=ON<br/> -D OPENMM_BUILD_FREE_ENERGY_PLUGIN=ON<br/> <br/> I thought the cmake was able to detect CUDA:<br/> as a matter of fact, CUDA seems poorly detectedby cma...

Hello,

I have just built with success OpenMM on my Linux box (amd64) with CUDA 3.0 support.
All test passed but the CUDA Ewald test (test 34).

Is it a known bug ?

Thanks in advance,
Jerome

Hi,

finally I find the way to do so:
sorry for the noise.

Jerome

<t>Hello,<br/> <br/> I am trying to build OpenMM on my Linux box (amd64) from source with CUDA support (version 3.0):<br/> at the very end of the build process I get the message:<br/> <br/> cannot find -lOpenMMCuda<br/> <br/> What is going on ?<br/> How can I fix it ?<br/> <br/> <br/> Please note th...