China’s ambition to become a major power in the supercomputing arena has become plainly obvious with the introduction of a supercomputing system nicknamed "Nebulae", which has earned the title of second-fastest supercomputer in the world with a Linpack performance of 1.271 PFlop/s. The Nebulae system itself is a hybrid design comprised of a Dawning TC3600 Blade system with Intel X5650 processors and NVidia Tesla C2050 GPUs. Despite being ranked #2 on the Top500, Nebulae is currently praised as being the fastest system worldwide in theoretical peak performance which is rated at 2.98 PFlop/s, the highest ever seen in the Top500. For a quick reference at just how fast this system can crunch numbers, a single minute of calculations from the Nebulae system would take your home computer over three weeks to complete.
Currently the United States still dominates the list, holding the #1 spot with its Jaguar Supercomputer housed at the Oak Ridge National Laboratory in Tennessee which has a peak performance of 1,750 trillion calculations per second. By comparison, the Jaguar System is over 33% faster than the Chinese contender but pales in comparison in theoretical yield which only reaches 2.3 petaflops. In addition to the Nebulae system, China has a total of 24 high performance systems on the Top500 with the Tianhe-1 supercomputer ranking in at number seven.
China is without a doubt rapidly becoming a major player in high performance computing and is seeking to solidify its holdings in the supercomputing world. Currently it is rumored that Dawning, the company responsible for the Nebulae machine, is currently developing an even faster machine for the National Supercomputer Center in Tianjin, China. The main purpose behind this machine will be to model industrial research such as aircraft design, aerospace fundamentals, and petroleum exploration. In a stark contrast, many of the US machines which are owned by the government are used to monitor nuclear weapon stockpiles.
Looking for other Technology Rental information? Visit the Tech Travel Site Map for a variety of computer and technology rental ideas.
Thursday, July 15, 2010
Friday, July 2, 2010
Oracle Unveils New High-Performance Sun Fire Clusters
Holding true to CEO Larry Ellison's promise that Oracle would primarily focus on the high-performance server market, last Monday Oracle expanded its line of x86 Sun servers with new rackmount, blade and network clustering servers. The newly released Sun Fire x86 Clustered Systems are intended for massive server configurations and will deliver a smaller footprint than previous generations of Sun File Server hardware. The servers themselves will come equipped with Xeon 5600 and 7500 processors, the latter is aimed at mission-critical systems which must always run and never go down. In doing so Oracle is competing against Intels Itanium processor line and even Sun's own SPARC line which has raised a few eyebrows.
The Sun Fire x86 Clustered Systems are designed for customers that run a mix of Oracle and non-Oracle enterprise workloads across a variety of systems. The cluster servers themselves consist of five rackmounted and two bladed servers that can hold two to eight processors and have been tightly integrated with Oracle software, middleware and management applications. Oracle has assured their customers that all Oracle software and middleware has been certified to run on these systems which have been optimized for Oracle Solaris, Oracle Enterprise Linux, and Oracle VM which supports Red Hat, Suse Linux, and even the KVM hypervisor.
In addition to optimizing their software and operating systems, Oracle has integrated Sun's Ops Center with its own Enterprise Manager providing a dynamic work flow for a single "lights-out" point of management. Blades, servers, storage, networking, virtualization and even powering the systems on and off can be handled all from a single web browser. Oracle has even included an Integration Assistant so you can configure and boot the systems straight out of the box within minutes. Even with no OS installed, the systems can reach out across the Internet to Oracle servers to check for firmware and BIOS updates and then download and install them.
"We claim we can manage a full blade ecosystem without requiring any network skills, because network virtualization is done in the silicon and through Oracle middleware technology," - Dimitris Dovas, director of product management for Sun hardware at Oracle
Oracle claims that this hardware will be able to deliver up to a 45 percent improvement in energy efficiency over previous generation systems and also the ability to run 70 percent more workload. They also claim that their new hardware can deliver the same performance in one-tenth the space with one-tenth the devices and one-fifth the number of network cables. The simpler cabling on the Sun Fire systems come through oracles Sun Blade 6000 10 GbE switches which are embedded within the blade itself or inserted through the top of the rack to connect the clusters. This improvement allows the server clusters to communicate to a network without having to connect though a networking cable and a switch.
Looking for other Technology Rental information? Visit the Tech Travel Site Map for a variety of computer and technology rental ideas.
Subscribe to:
Posts (Atom)