LinuxHPC.org recently published an article on the new Cray XD1 where they claim to have the lowest latency in HPC systems, at 1.8us for an MPI short message, 4 times faster than the competition. The Cray XD1 is indeed fast, and by developing custom processor boards and interconnects they've achieved an impressive result, however Quadrics' current generation QsNet II interconnect is faster still at 1.6us for the same MPI latency measurement. More importantly this is achieved by connecting standard commodity 1U dual processor Opteron servers.
The ability to deliver the world class, low latency performance, using standard commodity servers has made Quadrics the interconnect of choice where performance matters. Quadrics interconnects are part of some of the world's fastest computers; the Thunder system at Livermore National Laboratory currently number 2 in the top 500; and the ASCI Q at Los Alamos National Lab which is currently number 3.
We thank you for the opportunity to make the record straight.
Video - The Road to PetaFlop Computing
Explore the Scalable Unit concept where multiple clusters of various sizes can be rapidly built and deployed into production. This new architectural approach yields many subtle benefits to dramatically lower total cost of ownership.
White Paper - Optimized HPC Performance
Multi-core processors provide a unique set of challenges and opportunities for the HPC market. Discover MPI strategies for the Next-Generation Quad-Core Processors.
Appro and the Three National Laboratories
[Appro delivers a new breed of highly scalable, dynamic, reliable and effective Linux clusters to create the next generation of supercomputers for the National Laboratories.