Yesterday AMD announced its new FireStream 9250 general purpose GPU, which is…
…specifically designed to accelerate critical algorithms in high-performance computing (HPC), mainstream and consumer applications. Leveraging the GPU design expertise of AMD’s Graphics Product Group, AMD FireStream 9250 breaks the one teraflop barrier for single precision performance. It occupies a single PCI slot, for unmatched density and with power consumption of less than 150 watts, the AMD FireStream 9250 delivers an unprecedented rate of performance per watt efficiency with up to eight gigaflops per watt.
The 9250 replaces the 9170 announced last year. Perhaps more important than the 32-bit performance for HPC customers (the 1 TFLOPS level was also achieved by the NVIDIA T10P also announced Monday), the 9250 can get up to 200 GFLOPS of 64-bit floating point performance. This is twice the 9170, and roughly twice what NVIDIA’s new kit achieves. The 9250 only has 1 GB of GDDR3 RAM however — small compared to the 4 GB in NVIDIA’s newest product — which may restrict its usefulness in some applications.
An interesting bit of tid included in this announcement is AMD’s support for the newly announced Open Compute
Library (OpenCL), and effort championed by Apple to standardize the programming interface to GPUs.
In keeping with its open systems philosophy, AMD has also joined the Khronos Compute Working Group. This working group’s goals include developing industry standards for data parallel programming and working with proposed specifications like OpenCL. The OpenCL specification can help provide developers with an easy path to development across multiple platforms.
OpenCL will be part of the technology rolled into Snow Leopard, the next release of Apple’s Mac OS X operating system. In an interview in the NY Times Bits blog Steve Jobs said, “Basically it lets you use graphics processors to do computation. It’s way beyond what NVIDIA or anyone else has, and it’s really simple.”
Read more about the FireStream 9250, OpenCL, and what AMD is doing in my feature story at HPCwire today.