Titan (supercomputer): Difference between revisions
m merged paragraphs |
No edit summary |
||
| Line 7: | Line 7: | ||
| Location = [[Oak Ridge National Laboratory]] |
| Location = [[Oak Ridge National Laboratory]] |
||
| Architecture = 18,688 [[AMD Opteron]] 6274 16-core CPUs<br>18,688 [[Nvidia Tesla]] K20 GPUs<br>[[Linux]] Operating system |
| Architecture = 18,688 [[AMD Opteron]] 6274 16-core CPUs<br>18,688 [[Nvidia Tesla]] K20 GPUs<br>[[Linux]] Operating system |
||
| Memory = 710 [[TB]] (598 TB CPU and 112& |
| Memory = 710 [[TB]] (598 TB CPU and 112 TB GPU)<ref name=water/> |
||
| Storage = 13.6 [[Petabyte|PB]]<ref name=water/> |
| Storage = 13.6 [[Petabyte|PB]]<ref name=water/> |
||
| Speed = 20 [[FLOPS|petaFLOPS]] |
| Speed = 20 [[FLOPS|petaFLOPS]] |
||
Revision as of 15:17, 30 October 2012
| Active | Became operational October 29, 2012 |
|---|---|
| Sponsors | US DOE and NOAA (<10%) |
| Operators | Cray Inc. |
| Location | Oak Ridge National Laboratory |
| Architecture | 18,688 AMD Opteron 6274 16-core CPUs 18,688 Nvidia Tesla K20 GPUs Linux Operating system |
| Power | 12.7 MW |
| Space | 404 sqm (4352 sq ft) |
| Memory | 710 TB (598 TB CPU and 112 TB GPU)[1] |
| Storage | 13.6 PB[1] |
| Speed | 20 petaFLOPS |
| Cost | US$97 million |
| Legacy | First GPU based supercomputer to perform over 10 petaFLOPS |
| Website | http://www.olcf.ornl.gov/titan/ |
Titan is a supercomputer developed by Cray Inc. at the Oak Ridge National Laboratory for use in a variety of science projects. Titan is an upgrade of Jaguar, a previous supercomputer at Oak Ridge, to use GPUs in addition to CPUs.[2][3] Titan was announced in October 2011 and became operational in October 2012.[4][5]
According to Cray, the overall development costs were estimated in 2011 to be $97 million.[4] The NOAA contributed a small amount of the funding in return for computation time when completed.[6]The computer performs 20 petaFLOPS, or 20,000 trillion calculations per second although has a theoretical peak performance of 24.8 PetaFLOPS.[2][5]
Hardware
You simply can't get these levels of performance, power- and cost-efficiency with conventional CPU-based architectures. Accelerated computing is the best and most realistic approach to enable exascale performance levels within the next decade.
Titan uses the same building and 200 cabinets covering 404 sqm (4352 sq ft) that Jaguar did, simply replacing the internals.[5][7] Reusing the power and cooling systems already in place for Jaguar saved the lab approximately US $20 million.[6] The project replaced Jaguar's Cray XT5 compute blades with Cray XK7 blades which use AMD Opteron 6274 CPUs and Nvidia Tesla K20 GPUs.[4] There are 18,688 nodes (2 per blade), each containing a 16 core CPU with 32GB of RAM and a GPU with 6GB.[8][9] The total number of processor cores is 299,008 and total amount of RAM is over 710 Terabytes.[7] Titan draws 12.7 MW, 2 MW more than Jaguar did, but it is almost ten times as fast in terms of floating point calculations.[1]
GPUs were selected for their vastly higher parallel processing efficiency over CPUs.[9] Although the GPUs have a slower clock speed than the CPUs, each GPU contains 2,496 CUDA cores at 732 MHz resulting in a faster overall system.[1][7] Consequently the CPUs cores are used to allocate tasks to the GPUs rather than for directly processing the data as in previous supercomputers.[9]
Research projects
In 2009 the Oak Ridge Computing Leadership Facility (OCLF) considered fifty applications for use of the supercomputer but narrowed it down to six successful candidates chosen not only for the importance of the research, but for their ability to fully utilise the computing power of the hybrid system.[7][10] The code of projects had to be modified to suit the GPU processing of Titan but it was required that would still be possible on CPU based systems so that projects were not solely dependent on Titan.[10] Some projects found that the changes increased efficiency of their code on non-GPU machines; Denovo doubled the performance on CPU based machines.[10]
S3D is a project that models fine grain physics surrounding combustion aiming to improve the efficiency of diesel and biofuel engines. In 2009 they produced the first fully resolved simulation of autoigniting hydrocarbon flames relevant to the efficiency of direct injection diesel engines using Jaguar.[10] The WL-LSMS project simulates the interactions between electrons and atoms in magnetic materials at temperatures other than absolute zero. An earlier version of the code was the first to perform at greater than one petaFLOPS on Jaguar.[10] Denovo simulates nuclear reactions with the aim of improving the efficiency and reducing the waste of nuclear reactors.[7] Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) is a molecular dynamics code that simulates particles across a range of scales from atomic to relativistic to improve materials science with potential improvements to semi-conductors, biomolecules and polymers.[11] CAM-SE is a combination of two codes: Community Atmosphere Model, a global atmosphere model, and High Order Method Modeling Environment, a code that solves fluid and thermodynamic equations. CAM-SE will allow greater accuracy in climate simulations.[10] Non-Equilibrium Radiation Diffusion plots non-charged particles with potential applications in laser fusion, fluid dynamics, medical imaging, nuclear reactors, energy storage and combustion.[10]
References
- ^ a b c d Feldman, Michael (October 29, 2012). "Titan Sets High Water Mark for GPU Supercomputing". HPC Wire. Retrieved October 30, 2012.
- ^ a b Feldman, Michael (October 11, 2011). "GPUs Will Morph ORNL's Jaguar Into 20-Petaflop Titan". HPC Wire. Retrieved October 29, 2012.
- ^ Oak Ridge changes Jaguar's spots from CPUs to GPUs by Timothy Prickett Morgan, The Register Oct 11, 2011 [1]
- ^ a b c Poeter, Damon (October 11, 2011). "Cray's Titan Supercomputer for ORNL Could Be World's Fastest". PC Magazine. Retrieved October 29, 2012.
- ^ a b c Tibken, Shara (October 29, 2012). "Titan supercomputer debuts for open scientific research". CNET. Retrieved October 29, 2012.
- ^ a b Munger, Frank (October 29, 2012). "Titan's ready to roll; ORNL supercomputer may become world's No. 1". Knox News. Retrieved October 29, 2012.
- ^ a b c d e f "Introducing Titan". Oak Ridge Leadership Computing Facility. Retrieved October 29, 2012.
- ^ "Cray XK7 Specifications". Cray Inc. Retrieved October 30, 2012.
- ^ a b c "ORNL Debuts Titan Supercomputer" (PDF). Oak Ridge Leadership Computing Facility. Retrieved October 29, 2012.
- ^ a b c d e f g "TITAN: Built for Science" (PDF). Oak Ridge Leadership Computing Facility. Retrieved October 29, 2012.
- ^ Zybin, Sergey. "LAMMPS Molecular Dynamics Simulator". Sandia National Laboratories. Retrieved October 29, 2012.