BRUSSELS--HP ProLiant Generation 8 servers will include both new Intel Xeon E5-2600 series CPUs and Nvidia Tesla GPUs, it was revealed on Wednesday (March 7).
The HP ProLiant SL250 Gen8 CPU-GPU hybrid servers will combine Nvidia Tesla M2090 GPUs, with the newly announced Intel server parts based on the firm’s Sandy Bridge microarchitecture.
With the E5-2600 CPUs enabling an average of 12 percent higher performance for scalar processing, sequential bottlenecks can be eased and then tackled by the GPUs, which makes parallel computing significantly faster, something Nvidia said would result in “dramatically higher GPU utilization,” and “overall increased application performance.”
Benchmarked using LAMMPS, a molecular simulation application, the SL250 Gen8 server purportedly sped up workloads by over 20X, with three additional Tesla M2090 GPUs in place as accelerators.
With four CPUs and six GPUs in a 2U server configuration, the SL250 Gen8 can apparently deliver up to 4.6 teraflops of peak performance, thanks to its massively parallel accelerators.
Nvidia’s director of Tesla GPU computing Sumit Gupta said the announcement was a sign that the HPC industry was “rapidly embracing hybrid computing architectures,” for their superior performance and energy efficiency.”
Gupta also said that Intel’s higher performance Sandy Bridge CPUs had helped Nvidia maximize acceleration.
“Higher-performance Sandy Bridge CPUs unleash the power of NVIDIA GPUs to an even greater degree, enabling them to maximize acceleration for a range of commercial and scientific HPC applications, include AMBER, GROMACS, LAMMPS, Simulia Abaqus, seismic processing, and others.”
Tesla GPUs power three of the world’s top five supercomputers, today.
Previous：
Next ：