This project has moved. For the latest updates, please go here.

Harnessing the GPU to boost performance

Feb 10 at 2:05 AM
Edited Feb 10 at 2:08 AM
First let me say glancing around your code was inspiring. I know adding an option to hardness the GPU is a huge undertaking. However the GPUs abundant RAM and parallel processing are screaming to be together. In my humble opinion harnessing a GPU would make BrightstarDB a glaring standout. Below is a few resources and questions if you have a moment. Contrary to the only GPU triplestore solution (Blazegraph), creating a single local GPU solution might be more digestible/robust (ie. NVIDIA titian GPU).

Resources:
-Blazegraph released an article with amazing claims of their enterprise GPU solution.

-C# has some great managed libraries that utilize the NVIDIA GPU CUDA platform.

Questions:
-Just curious if you would ever consider the BrightstarDB/GPU pairing option?

-Is it naive to hope that the lowest level BrighstarDB APIs could be conditionally exchanged to use the GPU resources?

-Are you still actively designing BrighstarDB?