OpenCL and GPU Computing Latest News


I received / found some interesting news about OpenCL this week. I will talk more in detail about OpenCL next week because it will be in the spotlight…

1 – OpenCL demo with 1,000,000 particles

You can find more information about this demo HERE.

2 – Direct Compute is Less Advanced than OpenCL

I found this article where Neil Trevett, the president of the Khronos Group and a vice president at Nvidia Corp, explains why OpenCL is more advanced than Direct Compute:

OpenCL is a standalone, complete compute solution you can use for protein folding and particle analysis never touching the pixel, and you have the option of inter-opping it very closely with OpenGL, so you can use it for image processing and feeding into and feeding out of the OpenCL pipeline. The approach that DirectX 11 Compute takes is ‘super shaders’, which are like general-purpose C shaders. But those shaders exist within the context of the DX graphics pipeline, so it’s intended to soup up your graphics applications but you’d probably find it more difficult to write, you know, a general-purpose animation package.

3 – Port of Apple OpenCL demos to Windows

OpenCL Displacement demo
Displacement demo.

oscarbg — I think it’s the nickname… I couldn’t find an About section on his blog 🙁 — has ported to Windows some of the OpenCL demos you can find on the Apple OpenCL SDK. You can find more information and download link HERE.

4 – Supercomputer with 2560 ATI Radeon HD 4870 X2 Cards

Supercomputer with 2560 ATI Radeon HD 4870 X2 Cards

Working on a budget of 87.88 million USD, Chinese scientists created a supercomputer consisted out of 24,576 Intel Core2-based cores [6144 Harpertown CPUs – 3072 Xeon E5540s and 3072 Xeon E5450s] and 5120 AMD RV770 GPUs [2560 ATI Radeon HD 4870 X2 2GB cards]. Together, 6144 Intel CPUs and 2560 AMD GPUs reach a theoretical speed of 1.206 PFLOPS. More information HERE and HERE.

5 – OpenCL Development Kit for Linux

The OpenCL Development Kit for Linux on Power is an IBM implementation of the OpenCL Specification, Version 1.0. This implementation is for Power hardware running the Linux operating system and has been tested on the IBM BladeCenter QS22 systems running Fedora 9 and on the IBM BladeCenter JS23 systems running Red Hat Enterprise Linux 5.3.
More information HERE.

6 – Some OpenCL Video

7 – First Direct Compute Benchmark

Premier benchmark DirectCompute en ligne : nos résultats @ PC INpact
First DirectCompute Benchmark Released @ NGOHQ

Direct Compute Benchmark

8 – Debug your CUDA kernel with Cudapad


What is cudapad? Cudapad is a software that helps developments develop and test small kernals for NVidia’s© Cuda© language. Sometimes in your IDE you’ll want a quick way build or test a piece of Cuda code and CudaPad lets you do it. It shows the ptx code, cubin code, register count, error and more on the fly. There is no need to manually compile your code.
More information HERE (via Vizworld).