Appro: Bringing HPC to the Mainstream Enterprise
"Scalable Units enable multiple clusters to be built, based on the same architecture for small, medium and large-sized HPC deployments across a broad range of vertical markets," said Maria McLaughlin, Director of Marketing for Appro International (Milpitas, Calif.).
She makes the point that the HPC market, which used to be aimed solely at large government and university research facilities, has become far more mainstream. As well as large commercial users, she said small and midsize companies have jumped on the bandwagon.
"Appro is focusing its product design to address the HPC cluster market and key customer requirements, including system management, high availability and price/performance for HPC applications," said IDC Earl Joseph. "Appro has shown the ability to win highly sought-after, large-scale HPC deals positioning the company to benefit from strong market growth that IDC projects through 2012."
Since the last Server Snapshot more than two years ago, Appro has made a whole lot of changes. Gone are the Xtreme Server, Hyper Server and HyperBlade lines, replaced with the categories covered below. Like other so many other businesses, the company has gone green. The Appro GreenBlade system is an example of its new emphasis.
|Appro's green goods|
The GreenBlade was released in April 2009. It targets small to midsize workgroups and departmental HPC businesses. It can hold up to 10 blade servers, offering a total of 20 CPUs and 80 processing cores in a 5U form factor.
"The GreenBlade system is a modular and energy-efficient HPC cluster building block solution," said McLaughlin. "It addresses customers' need for flexibility of system configurations with different types of high speed connection, as well as reliability, density, power efficiency, storage and remote management at an affordable price."
There are two GreenBlade compute nodes. The Appro gB122H is a 2-socket six-core AMD Opteron 2000 series processor machine with up to 64 GB RAM and 1 TB storage capacity per node. The gB222X uses a quad-core Intel Xeon 5500 series processor, up to 48 GB RAM and 1TB storage capacity per node.
Appro has also released "Ready-To-Go-Clusters" based on the GreenBlade system. They support up to 28 blade systems and 224 processing cores in a 24U half-size standard rack cabinet.
"The 'Ready-To-Go-Cluster' targets small businesses with HPC industry specific application requirements," said McLaughlin.
For medium to large deployments, Appro offers the HyperGreen Cluster, also based on the GreenBlade system.
"The HyperGreen Cluster features up to 80 GreenBlade Nodes, totaling 640 processing cores per 42U standard rack cabinet, doubling the density compared with traditional 1U rack-mounted servers," said McLaughlin.
Fully configured, this Cluster comes with 160 Intel Xeon 5500 series processors, 640 cores per rack cabinet and 80TB storage capacity. How green is it? The HyperGreen Cluster reduces power consumption by up to 20 percent per node compared to traditional 1U servers. Users include organizations, such as Rice University's Ken Kennedy Institute for Information Technology in Houston, Texas.
"They are using Appro's HyperGreen Cluster to handle large HPC workloads for both high-memory bandwidth and a fast interconnection network that is critical for diverse data-intensive parallel applications," said McLaughlin.
In addition to its green lines, Appro offers the HyperPower Cluster and the Xtreme-X Supercomputer. The HyperPower Cluster is a hybrid solution based on Intel's 5500 series processors, and Appro's high density 1U servers that contain NVIDIA Tesla Graphics Processing Unit (GPU) cards combining the latest in CPU and GPU computing technologies.
"The Appro HyperPower Cluster supports up to 304 CPU cores and 18,240 GPU cores," said McLaughlin. "By using fewer systems than standard CPU-only clusters, this product delivers more computing power in an ultra-dense architecture at a lower cost."
This Xtreme-X Supercomputer, on the other hand, was designed for medium- to large-scale enterprise and HPC deployments that require high-performance and high-availability computing. It includes the Appro Cluster engine, which is a complete management system. This monster is a 128-socket/512 cores per rack cabinet that can be configured with 64-bit Opteron or quad-core Intel 5500 series processors.
|Name||GreenBlade Series||HyperGreen Clusters||HyperPower Clusters||Xtreme-X Supercomputer|
|Description||The GreenBlade System is a building block package offering performance density and energy efficiency. It can hold up to 10 blade servers offering a total of 20 CPUs/80 cores in a 5U form factor.||HyperGreen clusters are based on the GreenBlade system; it offers performance scalability, density and energy efficiency. It features up to 80 GreenBlade nodes, totaling 640 processing cores per 42U standard rack cabinet.||The Appro HyperPower Cluster is a Hybrid solution based on the latest CPUs/GPUs computing technologies. It supports 304 CPU/18,240 GPU cores in a 42U standard rack configuration.||The Appro Xtreme-X scales out to up to a 1000TF of computing power offering a reliable and scalable architecture that unites high performance capacity computing with fault-tolerant capability computing.|
|Target Deployment||Small to midsize HPC deployments||Medium to large HPC deployments||Small, medium and large HPC deployments||Midsize to large HPC deployments|
|Processor Type||Intel Xeon 5500 Series and AMD Opteron 2000||Intel Xeon 5500 Series||Intel Xeon 5500 Series and NVIDIA Tesla GPU Cards||Supports quad-core AMD Opteron and Intel Xeon processors|
|Price Range||Contact Appro||Contact Appro||Contact Appro||Contact Appro|