piece on Sun's Constellation

Print Friendly, PDF & Email

More from c|net’s on Sun’s Constellation announcement

The linchpin in the system is the switch, the piece of hardware that conducts traffic between the servers, memory and data storage. Code-named Magnum, the switch comes with 3,456 ports, a larger-than-normal number that frees up data pathways inside these powerful computers.

I had a preview of this switch back in the spring, and I can tell you that it’s impressive.

It’s really just hard core engineering: lots of connections in a very small space. The smarts are in the design, not in the components. Still, “anyone” can add a little value at the top. Focusing hard on the un-glamorous part of the underpinnings of modern supercomputers is how the advancements that really push things forward get made.

By way of comparison the largest InfiniBand switch on the market today has 288 ports; Mangum, with 3,456 ports will go a long way toward flattening the HPC interconnect, pulling the nodes closer together and pushing down latency.

The density of ports, and the large number of them, creates a cascading effect in performance and pricing, [Andy Bechtolsheim, chief architect and senior vice president of the systems group at Sun] asserted. By deploying Magnum, which sports a “fat tree” style architecture where servers branch out from the trunk of switches, customers will need to install far fewer switches when building large computers, he said. Fewer networking boxes mean about one-sixth the number of cables.

The system that Sun has put together gives it a machine in the same league as Blue Gene

A fully built-out Constellation system, with contemporary components, could hit a peak of 2 petaflops, or 2 quadrillion operations per second.

Though the recently announced Blue Gene/P keeps Big Blue out in front with a 3 PFLOPS max config.

Sun’s also worked on density, with a fully populated 42U Constellation rack piling in 768 cores (with quad-core nodes).

Constellation blades can accommodate Sun’s UltraSparc chips, AMD processors and Intel chips. AMD, however, currently provides better performance on floating point calculations than Intel’s chips, according to Bechtolsheim. The TACC system is based around Barcelona. Whether or not the TACC system can make the next Top 500 list revolves around availability of Barcelona, which is due in the third quarter.

By which means that whether TACC’s system makes the list depends upon whether the chips are available from AMD in time for the November list. If chips are available and the machine can be built, it will certainly “make” the Top500.