Supermicro announces TESLA based HPC servers.
9/30/2009 by: Anshel Sag
Today at GTC 09, Supermicro has announced that they will be releasing a slew of nvidia GPU based TESLA server solutions which implement CUDA to improve performance in the HPC segment. They demonstrated the world’s fastest 1U server, which is capable of 2 Teraflops featuring current generation of dual TESLA GPUs.
They will be releasing this in concert with the SuperWorkstation 7046GT-TRF-TC4, which supports four NVIDIA Tesla C1060 GPUs [and the upcoming GT300-based Tesla card] and supports up to three additional PCIe add-on cards for high-bandwidth I/O. The company will also be unveiling a brand new 2U Twin server that supports up to two hot-pluggable GPU nodes with redundant power on Friday during its presentation on advanced GPU server technology.
"To further extend our leadership position in GPU computing system architectures, the new 2U Twin GPU server features two hot-plug GPU computing nodes with multiple x16 non-blocking native Gen2 PCI-Express connectivity and redundant 1400-watt Gold Level (93%+ efficiency) power supplies for maximum availability," said Charles Liang, CEO and president of Supermicro. "Equipped with Supermicro’s patented thermal design, these highly parallel, multi-GPU systems are optimized for a wide range of graphics and computationally intensive applications in fields like medical imaging, oil and gas exploration, quantum chemistry, financial simulation, genomics and astrophysics."
Considered by Supermicro as the fastest 1U server in the world, Supermicro’s 6016GT-TF-TM2 Tesla-based server provides the high density computing capability which can be implemented on a large scale. Optimized for performance and reliability, the 6016GT-TF-TM2 supports dual Intel® Xeon® 5500 series processors and features two NVIDIA Tesla M1060 GPUs via two Gen2 PCI-Express x16 connections.
This implementation of Nvidia’s TESLA GPUs allows for companies to fit even more computing power into an even smaller space. They help save space and help save power and many companies looking for HPC solutions would gladly run servers with this Supermicro 1U solution because they save power and take up less rack space which in the end requires less racks and less power to the servers.
Nvidia, Supermicro, TESLA, GPU, Intel, HPC, AMD, CUDA, 1U, 2U, Teraflop, Power, Performance, GPGPU, cGPU, GT300, Fermi, TFLOPS
© 2009 - 2011 Bright Side Of News*, All rights reserved.