NVIDIA Tesla GPU powered supercomputer performs better than CPU-only sys
Date: 12/06/2013
Oak Ridge National Laboratory of U.S. has said in its release the Titan supercomputer runs molecular dynamics application LAMMPS 7x faster compared CPU-only system. Oak Ridge National Lab has completed rigorous acceptance testing of this one of the world's most powerful supercomputing systems for open science. Titan features 18,688 NVIDIA Tesla graphic processing units (GPUs) with 299,008 AMD Opteron CPU cores.
"The real measure of a system like Titan is how it handles working scientific applications and critical scientific problems," said Buddy Bland, project director at the Oak Ridge Leadership Computing Facility. "The purpose of Titan's incredible power is to advance science, and the system has already shown its abilities on a range of important applications and has validated ORNL's decision to rely on GPU accelerators."
The combination of GPU and CPU is said to offer low power processing. GPUs preferred for their parallel processing efficiency. Nearly Basket ball court size Titan super computer said to have costed close US$97 Million.
The Tesla K20 is having 2,496 stream processors with 5GB of memory with processing performance of 3.52 teraflops of single-precision and 1.17 teraflops of double-precision peak performance. The Tesla K20X is even more powerful with 2,688 stream processors and 6GB of GDDR5 memory delivering 3.95 teraflops single-precision and 1.31 teraflops double-precision peak floating point performance. Another good example of graphic processor for supercomputer is FirePro SM10000 dual-GPU workstation graphic processor from AMD delivering 1.48 teraflops of peak double-precision performance.
