Cray builds new XC-40 and CS-400 supercomputers using Haswell processors
Share on Twitter.
Get the most reliable SMTP service for your business. You wished you got it sooner!
September 30, 2014
Cray is building new XC-40 and CS-400 supercomputers using Haswell processors and DataWarp burst
buffer technology, in an effort to greatly increase the data processing speeds.
To be sure, the XC-40 clocks in at twice the speed of the existing XC-30, thanks to Intel Xeon E5-2600 v3
Haswell processors, scaling past a million cores.
The complex architecture implements two processor engines per computing node, with four compute
nodes per blade.
The appliance's blades stack in 8 pairs and each cabinet can be populated with up to three
chassis, meaning 384 sockets per cabinet.
This configuration can deliver up to 6,144 cores and enables 226 teraflops of performance per
Cray says-- "Future processor upgrades will boost clock frequency and boost the number of
embedded cores, accelerating overall system performance by a wide margin."
DataWarp is an application IO accelerator using flash memory directly connected to the XC-40 compute
nodes-- ie, PCIe flash caching. It gets data from storage and feeds it fast to the capable Haswells. Cray
says this means it meets "the worst case data I/O surge needs".
DataWarp PCIe-connect I/O blades with SSDs are inserted into XC40 banks of compute blades
and they are all connected via the Aries HPC interconnect.
Cray adds-- "High bandwidth can be delivered with virtually no impact on other I/O executing in
the system, ensuring QoS (quality of service) and sustained bandwidth to specific applications."
Gary Grider, the High Performance Computing Division Leader at Los Alamos National Lab says-- "The
new Cray XC-40 Trinity system will provide the first multi-petabyte, multi-terabyte-per second
burst handling capability ever.”
Both DataDirect Networks and EMC have been working on burst buffer technology at Los Alamos and
it looks like Cray has leap-frogged them.
The new XC-40 and CS-400 systems will be available with NVIDIA Tesla GPU accelerators and
Intel Xeon Phi coprocessors to provide even more processing power.
Other XC-40 features include-- Intel's Aries ASIC-based system interconnect for compute and I/O
nodes on the XC40 base blades using a PCIe Gen3 host interface. They also include Aries' Dragonfly
network topology, with which all processors are linked to all other processors, with no more
than five hops between any two.
The new system also provides a better cooling system. And you can count on a HPC-optimized programming
Overall, the CS-400 clustered supercomputers use industry-standard server building blocks that come with
air or liquid-cooling on CS-400-AC and LC systems respectively.
They scale past 11,000 computing nodes and 40 peak petaflops. Cray says the XC-400s can be
tailored to multiple purposes, from an all-purpose massively parallel HPC cluster, to one suited
for shared memory parallel tasks, to a cluster optimized for hybrid computing and data-intensive
Both systems are integrated with Cray’s HPC software stack and include tools compatible
with most open source and commercial compilers, schedulers and libraries.
Cray's Advanced Cluster Engine is a management software suite providing network, server, cluster
and storage management capabilities.
Recently, Cray has won a contract to provide an XC-40 supercomputer to the Swiss National Supercomputing
Centre in Lugano, Switzerland.
The Cray XC-40 and CS-400 systems are available now, likely priced in the tens of million of dollars. We'll
update you if we get clarity on Cray's pricing.
Source: Cray Inc.
Get the most dependable SMTP server for your company. You will congratulate yourself!
Share on Twitter.