rCUDA 3.1.1 Now Released
The new version 3.1.1 of rCUDA is now available. Release highlights:
*Fully updated API to CUDA 4.0 (added support for modules "Peer Device Memory Access" and "Unified Addressing").
*Fixed low level Surface Reference management functions.
*Added complete support for simple precision CUBLAS routines.
rCUDA distribution re-start
The rCUDA Team apologizes for the break on rCUDA distribution users might have experienced.
The delay has been caused by a distribution policy change: now rCUDA is not distributed as Open Source. Instead, it is distributed in binary format.
The process of preparing the different compilation environments and the new license has been an overhead to our continuous improvement tasks on rCUDA, which have not stopped at all during this period. We plan to re-start distributing a new version of rCUDA following the new distribution policy in a few days.
rCUDA in SuperComputing 2011
If you are attending SC11, don't miss rCUDA live demonstration on InfiniBand network at Mellanox's booth.
rCUDA in "The New York Times"
Link to an article in "The New York Times" printed about our presentation in Supercomputing 2011.