IBM on Monday unveiled high-capacity servers that are the first to be based on its new, multi-core POWER7 chip. The company said the new product line is designed “to manage the most demanding emerging applications,” including high-capacity smart electrical grids and real-time analytics for financial markets.
The servers are optimized for processing huge workloads of simultaneous transactions, data handling, and analysis. IBM said they offer “dramatic improvements” in price versus performance, energy savings, and server virtualization.
Energy Efficiency
The new systems are built around the newest generation of IBM’s POWER chip, with eight processing cores, and are designed to manage millions of transactions in real time while utilizing a variety of approaches to remain highly energy-efficient. With Intelligent Energy technology, for instance, parts of the system can be turned on or off, and clock speed can be lowered or raised, on one server or a group. The company said these systems can deliver twice the performance per watt as Sun SPARC servers and eight times that of comparable Hewlett-Packard Itanium-based servers.
POWER7 systems are optimized for massive parallel processing, high-end throughput, and analytics, and the systems can be used to manage others, such as IBM mainframes and x86-based servers.
Laura DiDio, a research fellow at Information Technology Intelligence Corp., said these “turbocharged systems should really jumpstart multi-threading computing.” She added that each of the eight cores can execute up to four tasks each, so “it’s as if this were a virtual 32-core processor on a chip.”
This capacity is four times the maximum number of cores in POWER6 systems and eight times the processing threads.
Smart Energy Grids
DiDio noted that these systems will enable a huge performance boost in the growing market of smart energy grids, allowing utility companies to go from a million meter reads a day to 85 million. In its announcement, IBM noted…
Via NewsFactor
Yacine.org: IBM’s New POWER7 Servers Save Energy with Big Loads http://bit.ly/clGdcS
This comment was originally posted on Twitter