Introduction
This article explores making use of the GPU for general purpose processing from .NET.
Background
Graphics Processing Units (GPUs) are being increasingly used to perform non-graphics work. The world's fastest super computer - the Tianhe-IA - makes use of rather a lot of GPUs. The reason for using GPUs is the massively parallel architecture they provide. Whereas even the top of the range Intel and AMD processors offer six or eight cores, a GPU can have hundreds of cores. Furthermore GPUs have various types of memory that allow efficient addressing schemes. Depending on the algorithm this can all give a massive performance increase, and speed-ups of 100x or more are not uncommon or even that complicated to achieve. This is not just something for super computers though. Even normal PCs can take advantage of GPUs. I'm typing on a fairly cheap Acer laptop ($800) and it has an Intel i5 processor and an NVIDIA GT540M GPU. This little thing hardly runs warm and can give my fairly standard workstation with its two NVIDIA GTX 460s a good run for its money. The amazing thing about this work station is that in ideal conditions it can do 1 teraflops (less than 100GFLOPs are down to the Intel i7 CPU) . If you look through the history of super computers, this means I've got something that matches the performance of the best computer in the world in 1996 (ASCI Red). 1996 is not that long ago. In 2026, can we have a Tianhe-IA under our desks?
The key point above is that whether an application can speed up or not is down to the algorithm. Not everything will benefit and sometimes you need to be creative. Whether you have the time or budget to do this must be weighed up. Anyway whatever lowers the hurdle to taking advantage of the supercomputer in your PC or laptop can only be a good thing. In the world of General Purpose GPU (GPGPU) CUDA from NVIDIA is currently the most user friendly. This is a variety of C. Compiling requires use of the NVIDIA NVCC compiler which then makes use of the Microsoft Visual C++ compiler. It's not a tough language to learn but it does raise some interesting issues. Applications tend to become first and foremost CUDA applications. To extend a 'normal' application to offload to the GPU needs a different approach and typically the CUDA Driver API is used. You compile modules with the NVCC compiler and load them into your application.
This is all very well but it leaves a rather clunky approach. You have two separate code bases. This may not be a big deal if you have not enjoyed Visual Studio and .NET. Using CUDA from .NET is another story. Currently NVIDIA direct .NET people to CUDA.NET which is a nice, if rather thin, wrapper around the CUDA API. However GPU code must still be written and compiled separately with NVCC and work appears to have stopped on CUDA.NET. The latest changes that came in with CUDA 3.2 and now 4.0 mean that a number of things are broken (e.g. CUDA 3.2 introduced 64-bit pointers and v2 versions of much of the API).
Being a die hard .NET developer, it was time to rectify matters and the result is Cudafy.NET. Cudafy is the unofficial verb used to describe porting CPU code to CUDA GPU code. Cudafy.NET allows you to program the GPU completely from within your .NET application with a minimum of messy, clunky business. Now on with the show.
The Code
This project will get you set-up and running with Cudafy.NET. A number of simple routines will be run on the GPU from a standard .NET application. You'll need to do a few things first if you have not already got them. First be sure you have a relatively recent NVIDIA Graphics card, one that supports CUDA. If you don't, then it's not the end of the world since Cudafy supports GPGPU emulation. Emulation is good for debugging but depending on how many threads you are trying to run in parallel can be painfully slow. If you have a normal PC, you can pick up an NVIDIA PCI Express CUDA GPU for very little money. Since the applications scale automatically, your app will run on all CUDA GPUs from the minimal ones in some net books up to the dedicated high end Tesla varieties.
Read more: Codeproject
Posts
0 comments:
Post a Comment