Interesting. As an active (but young) researcher in distributed-memory computing, I wouldn't say that the funding has been turned off. I would instead argue that many-core/GPU/MIC and MapReduce approaches took over as the parallel computing topics du jour.