Summit is the next leap in leadership-class computing systems for open science. With Summit we will be able to address, with greater complexity and higher fidelity, questions concerning who we are, our place on earth, and in our universe.
Summit will deliver more than five times the computational performance of Titan’s 18,688 nodes, using only approximately 3,400 nodes when it arrives in 2017. Like Titan, Summit will have a hybrid architecture, and each node will contain multiple IBM POWER9 CPUs and NVIDIA Volta GPUs all connected together with NVIDIA’s high-speed NVLink. Each node will have over half a terabyte of coherent memory (high bandwidth memory + DDR4) addressable by all CPUs and GPUs plus 800GB of non-volatile RAM that can be used as a burst buffer or as extended memory. To provide a high rate of I/O throughput, the nodes will be connected in a non-blocking fat-tree using a dual-rail Mellanox EDR InfiniBand interconnect.
Upon completion, Summit will allow researchers in all fields of science unprecedented access to solving some of the world’s most pressing challenges.
|Application Performance||Baseline||5-10x Titan|
|Number of Nodes||18,688||~4,600|
|Node performance||1.4 TF||> 40 TF|
|Memory per Node||32 GB DDR3 + 6 GB GDDR5||512 GB DDR4 + HBM|
|NV memory per Node||0||1600 GB|
|Total System Memory||710 TB||>10 PB DDR4 + HBM + Non-volatile|
(node injection bandwidth)
|Gemini (6.4 GB/s)||Dual Rail EDR-IB (23 GB/s)|
|Interconnect Topology||3d Torus||Non-blocking Fat Tree|
|Processors||1 AMD Opteron™
1 NVIDIA Kepler™
|2 IBM POWER9™
6 NVIDIA Volta™
|File System||32 PB, 1 TB/s, Lustre©||250 PB, 2.5 TB/s, GPFS™|
|Peak power consumption||9 MW||15 MW|
The OLCF has selected the next set of partnership projects into its Center for Accelerated Application Readiness (CAAR) program, a collaborative effort of application development teams and staff from the OLCF Scientific Computing group to prepare for Summit. These projects were chosen based on a computational and scientific review conducted by the OLCF in consultation with the ALCF, NERSC, IBM and NVIDIA. The application teams represent a broad range of computational algorithms and programming approaches in a diverse range of scientific disciplines including astrophysics, biophysics, chemistry, climate modeling, combustion engineering, materials science, nuclear physics, plasma physics and seismology.
Summit will provide revolutionary performance by way of evolutionary changes to the current Titan hybrid architecture, making Summit an ideal follow-on system to Titan. By developing and refactoring applications to improve performance portability on accelerated architectures, Titan users will be better positioned to take advantage of Summit, as well as other next-generation leadership computing resources, and beyond. Users can create applications that explore performance portability and exploit untapped parallelism by:
The OLCF has established best practices for leadership-class computing for more than a decade. The OLCF created the Center for Accelerated Application Readiness, or CAAR, to help prepare codes for future generation systems. For Summit, CAAR will have a competition and select eight partnership teams to prepare their important scientific applications for highly effective use on Summit. The partnership teams, consisting of the core developers of the application and OLCF staff, will receive support from the IBM/NVIDIA Center of Excellence at ORNL and have access to multiple computational resources. For more information about CAAR, please visit CAAR: Call for Proposals.
Coinciding with the Summit launch, the OLCF will offer various tutorials, webinars, and workshops to help make sure users are ready on day one. During this three-year period of planning and development, please be aware that the OLCF’s goal is to smoothly transition from Titan to Summit by providing an abundance of guidance to users during the process.
Prior to the Summit launch, OLCF user assistance staff and vendor partners will host workshops to ensure users get the most out of Summit as well as the facility’s data and storage resources.
While users can still join us in person for Summit workshops, we will also provide them as webinars so users can plug in from anywhere in the world and refer to these resources at any time.
Users preparing applications for Summit will be able to access coding tutorials prior to the Summit launch that provide “how-to” examples of programming operations for an accelerated architecture.
Extensive details on the Summit system, including software installations, file systems, programming environment, and more will be available online as a Summit User Guide before Summit is operational.
U.S. sets sights on 300 petaflop supercomputerRead more
U.S. government spending $425 mln to build fastest supercomputersRead more
ORNL to partner with IBM on next big supercomputer: ‘Summit’Read more
ORNL to Acquire New SupercomputerRead more
Oak Ridge National Laboratory signs contract with IBM for next-generation of supercomputerRead more
Next-generation supercomputer coming to LabRead more
Mellanox EDR 100Gb/s InfiniBand Solutions Chosen for Leading Supercomputer Project (“CORAL”)Read more
Joint effort nabs next wave of US supercomputersRead more
IBM Signs $325 Million Supercomputing Deal With Dept. Of EnergyRead more
IBM and Nvidia win $425M to build two monstrous supercomputers for the Department of EnergyRead more
Alexander: World’s fastest supercomputer will again be at ORNLRead more