IBM & the US Govt’s Oak Ridge National Laboratory (ORNL) have unveiled ‘Summit’, a supercomputer, which they claim is the world’s “most powerful & smartest scientific supercomputer”.
The US Department of Energy’s Summit supercomputer enables scientists to simulate complex physical systems & make predictions critical to advancing research & development, says the ORNL on its Website.
Summit’s “smart” architecture merges GPU acceleration & “dense local memory” will support expanding applications in data science & Artificial Intelligence (AI).
Here’s what this supercomputer can do:
- A 200-petaflop machine, Summit can perform 200 quadrillion (peta-) floating point operations per second (flops). If every person on Earth completed one calculation per second, it would take 305 days to do what Summit can do in 1 second.
- For some AI applications, researchers can use less precise calculations than flops, potentially quadrupling Summit’s performance to exascale levels, or more than a billion billion calculations per second.
Summit is connected by 185 miles of fiber optic cables, its file system can store 250 petabytes of data, or the equivalent of 74 years of high-definition video. At over 340 tons, Summit’s cabinets, file system, & overhead infrastructure weigh more than a large commercial aircraft. It occupies 5,600 sq. ft. of floor space, which is the equivalent of 2 tennis courts. More than 4,000 gallons of water pump through Summit’s cooling system every minute, carrying away about 13 megawatts of heat.
Specifications & Features
- Processor: IBM POWER9
- GPUs: NVIDIA Volta
- Nodes: ~4600
- Node Performance: >40TF
- Memory/node: 512GB DDR4 + HBM
- NV Memory/node: 1600GB
- Total System Memory: >10PB DDR4 + HBM + Non-volatile
- Interconnect Topology: Non-blocking InfiniBand Fat Tree
- Peak Power Consumption: 15MW
Image Credit: Oak Ridge National Laboratory