This account is pending registration confirmation. Please click on the link within the confirmation email previously sent you to complete registration. Need a new registration confirmation email? Click here
BILLERICA, Mass., Sept. 6, 2012 (GLOBE NEWSWIRE) -- CSP Inc. (Nasdaq:CSPI) MultiComputer Division, a leading supplier of commercially developed rugged embedded computers, introduces the 3300GTX NVIDIA GPGPU coprocessor blade for its TeraXP™ product line. Leveraging open standards technologies, the 6U OpenVPX™ blade delivers high performance parallel stream processing capabilities and high bandwidth FDR (56 Gb/s) InfiniBand connectivity to radar, sonar, and surveillance signal processing applications.
The 3300GTX currently hosts a GPGPU, with the NVIDIA Fermi architecture, optimized for floating-point math and memory bandwidth to provide the performance required for processing large volumes of data in scientific and technical applications. Designed for rapid technology refresh, the 3300GTX implements the industry standard MXM 3.0 Graphics Processing Unit (GPU) to easily accommodate an upgrade to newer GPGPU's with the Kepler architecture keeping pace with the ever increasing performance needs and lower power consumption requirements of today's signal processing applications.
Beyond raw computational power, the 3300GTX features the latest Mellanox ConnectX®-3 HCA with Virtual Protocol Interconnect (VPI) to support interoperability with an (FDR/QDR/DDR) InfiniBand or 40GbE fabric. By employing a switched fabric interconnect, the 3300GTX provides the system builder with high throughput, low latency, quality of service, failover capabilities, and scalability. The 3300GTX board failover capabilities are implemented via the HCA supporting a 56 Gbp/s QSFP transceiver on the front panel and a VITA 66.1 optical interconnect on the backplane. This addition of fiber-optic backplane connectivity brings benefits in improved density, up to 24 fibers per connector, and offers room for future growth.
It is simple to realize the performance advantages of the 3300GTX GPGPU by porting existing applications or developing new code using the supported OpenCL, OpenACC, and CUDA toolkits.
The 3300GTX blades are designed to support air-cooled and conduction-cooled thermal management for deployment in harsh environments where Size, Weight, and Power (SWaP) are critical. The new blades will be available for delivery Q1 2013.