14 DAY TRIAL // It's a known fact that GPU architecture is in some ways more advanced than CPU related one. And it's also known the fact that - in order to release all the computational power residing in your GPU - you would need something like a GPGPU (general purpose GPU) and by that I mean a GPU with architecture capable of interpreting codes written in a specific language. Ati made the first move by optimizing its architecture in order to run Folding@Home project. At the moment Nvidia stood still.
Then the launch of G80 GPUs arrived. These chaps use a unified shader architecture incorporating 128 independent stream processors which actually run at a higher speed than the core itself (1.35GHz vs. 575MHz) making 8800 GTX somewhat similar to NetBurst CPUs who had the ALU units working at twice the speed of the CPU.
Nvidia is full of surprises. The latest came in form of a C compiler for G80 GPUs integrated into their latest software suit named CUDA (Compute Unified Device Architecture). What this actually does is transform the C code into instructions that the G80 GPU can execute. And, if you take into account the massive power this GPU has, you'll get the idea.
"CUDA gives us a whole new level of computing capability and enables closer access to the hardware. CUDA makes it possible for Acceleware's electromagnetic simulation and geophysical processing products to continue to double in speed each year, and, with our OEM partners like SPEAG, will enable us to address the needs of new markets such as biomedical imaging and reservoir modeling. The latest advancements from NVIDIA are helping to quickly push the boundaries of product development and commercial science," said Ryan Schneider, CTO of Acceleware Corp.
"CUDA opens the door for new ways to analyse and interpret seismic data, allowing for interaction with multi-terabyte prestack surveys. With this compelling new GPU architecture from NVIDIA, we can accelerate some of the most computationally intensive algorithms in oil and gas exploration - far beyond the performance CPUs are capable of delivering," stated Alex Krueger, president of Headwave, Inc.
CUDA SDK is available for download on Nvidia developer web page. Enjoy!