Transtec has released GPU computing solutions for particularly intensive computing operations such as process simulations or rendering and modelling complex data items in research. Computing cards are used in the Transtec 1000W Cuda supercomputer PC Tesla 1060 while the 1000R Cuda supercomputer features more powerful Tesla 1070 systems. Both systems are compatible with Linux and Windows.
The Transtec 1000R Cuda supercomputer is based on Nvidia's Tesla S1070 system. Its GPU platform makes it possible to accommodate four Tesla graphics cards with one Teraflop of computing power each plus a total of 16 Gigabytes of graphic storage in a single height unit (1 HU). Up to 960 computing units (cores) can be assigned for complex and data-intensive processes. Data transfer rate has increased to 408 Gigabytes per second. The front-end, which has a connection via a fast PCI-Express 2.0 port, can be fitted with up to four Intel Xeon processors offering a maximum of 128 Gigabytes of main memory. Up to four SATA hard drives (hot-swappable) can be integrated into the system.
The Transtec 1000W Cuda supercomputer PC is suitable as an individual supercomputing workstation. Its tower, measuring almost 43cm high by 53cm deep, can accommodate up to two Teraflops of computing power depending on the respective configuration. The basis for this is offered by a CPU with an Nvidia Quadro NVS 290 graphics card plus two Nvidia Tesla C1060 computing cards. These Tesla cards offer more than four Gigabytes of storage each. Add to this, 8 Gigabytes of RAM. The CPU is an Intel Core 2 duo and a 160 Gigabyte SATA-2 hard disk comes as standard.
