Armonk, N.Y.'s IBM kicked off the festivities Friday with the latest iteration of its Deep Blue supercomputer (yes, very same that beat chess champion Garry Kasparov), the Unix-based eServer p655, which it claimed is capable of reaching half a trillion operations per second (teraflops, for short) in peak processing power.
Such blazing operating speed is prized today to meet the demands of scientific and technical computing, as well as applications that require very large databases or massively parallel processing including digital media and life sciences. IBM's supercomputers, in particular, are used regularly by life sciences organizations to explore genomic research, in automobile design, and in financial markets.
The p655 packs up to 128 POWER4 processors in a single rack and is available in four or eight processor building blocks. IBM is aggressively attacking supercomputing products from HP, specifically the HP rx5670 Itanium2 system. Big Blue said one eServer p655 rack with 128 POWER4 processors occupies as little as one-sixth the floor space of an HP Itanium2 with the same number of processors. Freeing up floor space is considered key for any IT enterprise.
The eServer p655 system can be clustered using eServer cluster 1600, combined using a high-performance switch. Systems can also be defined using logical portioning. IBM will offer its clustering software for administration from a single control workstation.
The starting price tag on the p655 is $73,485. It will ship later in 2002.
This week and next promises a wealth of supercomputing news, as firms prepare for the Super Computing 2002 (SC2002) (November 18-21) show and 3rd International Forum on Grid Computing(November 18). Indeed, Sun plans to announce a significant supercomputing push next week, but there have already been a number of recent announcements in the high-performance computing (HPC) space.
On Monday, Mountain View, Calif,'s SGI unveiled its Origin 3900 server, which can scale to 512 processors and 1 terabyte (TB) of memory in racks that combine up to 128 processors and 256GB of memory per rack.
Thursday was Cray's turn, when it unveiled the most powerful supercomputer to date, the Cray X1, which is available with up to 52.4 trillion calculations per second (teraflops) of peak power and 65.5 terabytes of memory. The Seattle firm boasted that it that it is the first company to accept the government's challenge to provide sustained (not merely "theoretical peak") petaflop computing speed -- 1,000 trillion calculations per second -- for applications by 2010.
Indeed, the supercomputing space is enjoying somewhat of a renaissance, according to some analysts.
"The HPC industry is entering a new chapter in which economic conditions and fierce competition are forcing businesses to change the way they utilize their computing resources," said Debra Goldfarb, group vice president, worldwide systems and servers, IDC. "Efficiently managing and using data in this highly competitive world is central to companies processing information and turning it into superior decisions."