AMD launches new Opteron server chips
X86 microprocessor vendor AMD launched its latest server chips, the six- and eight-core Opteron 4200, aka “Valencia,” and the four-, eight-, 12- and 16-core 6200, code-named “Interlagos.” The chips are generally available, and will be adopted by key server OEM partners, including HP, Dell and IBM. List price for the AMD 4200 and 6200 Opterons ranges from $125 and $1,019 per chip.
Compared with its competitor, the Intel Xeon, the new Opteron chips can deliver up to 84% more performance, 73% more memory bandwidth and half the power per core, AMD said.
Like its predecessors, the new Opterons still offer up to four memory channels, using 1,600 MHz memory speeds. They support up to 12 memory DIMMs per CPU, for a maximum of 384 GB, assuming 32 GB DIMMs.
The new AMD Opteron chips also introduce new power management technologies. They now support the CFIX standard for up to 46% less power consumption by an idling chip. In addition, AMD’s new Thermal Design Power PowerCap feature allows administrators to set a chip’s maximum power draw and AMD CoreSelect allows administrators to turn off cores as needed.
AMD also announced future plans to deliver the AMD Opteron 3000 series for the ultra-dense low-power single-processor Web serving and hosting market. The four- and eight-core “Zurich” processor based on AMD’s Bulldozer architecture will ship in the first half of 2012.
Get a mainframe on demand
Maintec Technologies is offering a new “mainframe on demand” service for companies with short- or long-term mainframe infrastructure needs; for example, technology training and development firms, or production companies with a temporary need for increased capacity. Maintec’s mainframe infrastructure consists of an IBM System z9 BC. The new service will be available on a pay-as-you-go basis.
Egenera automates disaster recovery
Network and I/O virtualization provider Egenera has improved the disaster recovery capabilities of its PAN Manager software. PAN Manager 7.1 now has the ability to automate disaster recovery failover and recovery between dissimilar systems, reducing costs associated with redundant hardware, and minimizing software costs by repurposing software licenses. A new “one-click” restore feature automates the restore of a multi-tier environment, including server, storage and network configurations, according to the administrator’s prioritization scheme. In addition, PAN Manager 7.1 can manage twice the number of blades than the previous version, depending on the blade system being managed.
HPC cluster manager bursts to the cloud
Bright Computing, maker of cluster management software for Linux-based high performance computing clusters, has introduced version 6.0 of its Bright Cluster Manager with support for cloud bursting. Bright Cluster Manager 6.0 supports two scenarios for cloud bursting: creating new clusters in the cloud, or adding cloud-based nodes to an existing cluster. In the latter case, BCM manages the cloud-based nodes as if they are part of the local cluster, providing provisioning, monitoring, scheduling and other management functions, the company said. Initially, BCM provides cloud bursting to Amazon EC2, and the company is in conversations with customers about support for other public clouds. BCM 6.0 will be generally available in January.
Provisioning system separates OS, apps from cloud provider
ScaleXtreme, a cloud provisioning system provider, introduced the concept of Dynamic Server Assembly (DSA), which attempts to disassociate operating system and application packages to be provisioned from the underlying format imposed by a particular cloud provider. Using DSA, administrators can model, provision and launch servers in a cloud-agnostic fashion, and change cloud providers down the road in response to availability, performance or cost needs. DSA functionality is available to users of the free ScaleXtreme Xpress and paid ScaleXtreme Xpert provisioning services.
ADVA FSP 3000 qualified for IBM zBX
Optical networking vendor ADVA has qualified its FSP 3000 low-latency wave division multiplexing solution with the IBM zEnterprise BladeCenter Extension, or zBX. This paves the way for increased availability and data connectivity for IBM Power and x86 blades housed in the zBX chassis, thanks to the FSP 3000’s ability to multiplex, transport and protect high-speed data, storage, voice and video applications over fiber optic networks.
Dig deeper on x86 commodity rackmount servers