Archive notice: This article was originally published on June 18, 2008. Links and embedded videos are preserved as part of the historical record.
NVIDIA is presenting its new Tesla 10-series at the International Supercomputing Conference (ISC), held in Dresden from June 17 to 20. The computing platform is binary-compatible and supports the C programming language. This enables developers to solve their computing problems in a standard environment that can be scaled easily without recoding. The Tesla 10-series is ideally suited to high-performance computing (HPC) in areas such as video encoding, medicine, and scientific research.
The new Tesla product family includes the Tesla S1070 1U computing system and the Tesla C1060 computing processor. Compared with their predecessors, they offer:
- Twice the performance, with up to 4 teraflops per 1U system
- Double precision with IEEE 754 support
- Twice the memory, with 16GB per 1U system
-
Up to three times greater energy efficiency.
Combined with NVIDIA’s CUDA C development software for parallel computing, the new Tesla products extend the use of GPUs to any compute-intensive application. To date, more than 70 million CUDA-capable GPUs have been sold, and more than 60,000 downloads of the C compiler have been recorded on NVIDIA’s CUDA community website, CUDA Zone. This makes NVIDIA’s CUDA and Tesla technologies the most widely used GPU platforms in the HPC community. NVIDIA’s CUDA tools are used in areas such as financial analysis, astrophysics, and seismic imaging. Developers in these fields can parallelize their software with NVIDIA products and use the parallel computing architecture of GPUs to distribute computing tasks automatically across several hundred processing cores.
NVIDIA’s Tesla S1070 and Tesla C1060 will be available from August 2008.
