SIMD < SIMT < SMT: parallelism in NVIDIA GPUs edit / delete

How parallel constructs in CUDA et al. map to hardware.

to ag0801 gpgpu gpu parallel ... on 03 December 2011

The ryg blog edit / delete

"A trip through the graphics pipeline."

to ce0813 gpu graphics hardware ... on 15 August 2011

Software | Andreas Klöckner's web page edit / delete

Lots of software related to scientific computing, and in particular doing high-performance stuff with code generation and GPUs. There's also a particle simulation engine and various other fun toys here.

to c++ gpu maths python scientific software ... on 31 August 2010

Welcome — Theano v0.1 documentation edit / delete

Python maths library that uses a number of tricks to speed up operations: CAS techniques to rewrite equations, runtime code generation, GPU bindings...

to cas compiler gpu maths numeric python ... on 31 August 2010

Welcome to PyOpenCL’s documentation! — PyOpenCL v0.92beta documentation edit / delete

OpenCL bindings for Python. (The same site also offers PyCUDA.)

to gpu opencl python software ... on 30 August 2010

Browser bookmarks: tasty+ | tasty= Log in | Export | Atom