Hi Victor, I've written Penumbra (http://github.com/ztellman/penumbra), which is a wrapper for OpenGL that allows for some limited general purpose computation. I've also written Calx (http://github.com/ztellman/calx) which is a wrapper for OpenCL, and it's designed for general purpose computation on the GPU.
The thing is, though, that when people say "general purpose GPU", they don't mean "my webserver runs ten times faster now!". GPUs use the space usually reserved for L1-3 caches for more computing power. Each compute kernel has a very limited amount of memory that's fast to access (generally as fast as registers), but otherwise you have to reach out to main memory, which is hundreds of times slower. So basically, if you want to write GPU targeted stuff, use Calx. It's a pretty nice abstraction over the OpenCL API; certainly it's better than writing it in C. It's also not much slower than writing it in C, because all you're doing is enqueueing computations, writes, and reads. Penumbra has a Clojure DSL that can run on the GPU, but the further away you get from graphics operations, the leakier the abstraction gets. The day where you can write plain Clojure and run it on the GPU is a long way off. Zach On Jul 19, 6:31 am, Timothy Baldridge <tbaldri...@gmail.com> wrote: > > Most architectures are made to execute a lot of times the same instruction > > on a different block of data. > > You can have if...then...else... in your GPU programs, but every processor > > will execute both branchs (with a tag saying wether > > or not it is *really* executing it or not) > > This is kindof true...although it's not as bad as it sounds. IIRC the > current break-point is 16 threads will be executing the same > instruction in modern NVidia GPUs. This is called the "Warp size". > Modern GPUs can execute hundreds of threads at one time, so it is > possible to have some fairly branch heavy code. It may run slower than > non-branch code, but can often still out perform CPUs. > > I would look at something like > scriptjure:http://github.com/arohner/scriptjurecombined with Penumbra. This > way > you could code in Clojure, have the code translated to CUDA or OpenCL > code, and then executed on the GPU and have the results returned as a > sequence. > > Timothy -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
