SQL Anywhere only executes on the licensed CPUs, not on any graphics cards (GPUs). If you have code you want to run on the NVidia Tesla, you may be able to compile a DLL with |
Many of the more typical database applications are not usually very compute intensive. Of course in a research or design setting (for a couple of obvious examples) there could be a different and possibly more interesting story. Do you have a particular application domain or focus where this could be a real benefit? There are some characteristics of the current generations of massively parallel GPUs that could make them less ammendable to database intensitive applicaitons but a good example of what you are looking for could help shape the responses and suggestions others might be able provide. |
Should you want just the information on using GPUs in the database then there is an interesting project called PGStorm in the PostgreSQL camp: https://wiki.postgresql.org/wiki/PGStrom http://on-demand.gputechconf.com/gtc/2015/presentation/S5276-Kohei-KaiGai.pdf Also just found this quite recent article: http://www.theplatform.net/2015/06/26/gpu-accelerators-radically-boost-sql-queries/ |
maybe change this to a feature request ;-)
In case there would be any general advantage for SQL Anywhere in using a GPU (which I don't know) - given it's focus on platform and hardware independency, OpenCL would possibly be a preferred choice over Nvidia's CUDA here...