Remember System X? The 1100 node, G5-based super built out at Virginia Tech in 2003? Me too…it was number three on The List and I had just recently started using Macs so I thought it was pretty cool.
Those intrepid adventurers are building a new cluster, this one 29 TFLOPS, with 324 Mac Pro towers for research. From an article at Ars Technica
The new system will not replace System X, according to CHECS Director Dr. Srinidhi Varadarajan. “Unlike System X, which is a production machine for computational science and engineering researchers, the new system is intended to be pure computer science research system,” he told Ars. The system will be used to study two main areas: power-aware software systems that can adjust performance automatically to maximize efficiency, and distributed shared memory systems that can run existing threaded code on high-performance clusters “as if they were scaled up versions of multi-core desktops.”
Stats, and some refreshing honesty
System X uses 1100 Xserves with dual-2.3GHz, 64-bit G5 processors—2,200 cores total—connected with a standard data rate InfiniBand switching network with 20 Gbps bandwidth. In comparison, the new, presently-unnamed system is comprised of 324 dual-quad core 2.8GHz Xeon processors for a total of 2,592 cores. The system is also connected by a quad-data rate InfiniBand network, which provides 80 Gbps total bandwidth, a 300 percent increase. Thought the hardware should give some blazing performance, it probably won’t reach 29 Tflops when benchmarked. “I would be vastly surprised if it sustained 29 teraflops on Linpack,” said Varadarajan. “The new cluster will provide us with a large scale development and performance tuning platform for both these activities,” however.