faq:cuda_support
This is an old revision of the document!
Which parts of CP2K are CUDA-accelerated?
- FFTs, when compiled with
-D__PW_CUDA - Anything that uses
dbcsr_multiply, when compiled with-D__DBCSR_CUDA. This includes in particular the linear scaling DFT code. - Anything that uses
dgemm(notpdgem), when linked against Cray's accelerated LibSci. Currently, this is most useful for theMP2method.
faq/cuda_support.1393322577.txt.gz · Last modified: (external edit)
