Monitoring HPC Clusters: Is It Time for an Extreme Makeover?

By Ian Lumb | November 13, 2014

From the components that comprise a node (e.g., CPU cores plus accelerators and coprocessors, memory, disks), to the operating system and HPC software stack (e.g., client system for executing workload), device-specific metrics abound. As individual nodes communicate via interconnect fabrics (e.g., Ethernet, InfiniBand), metrics mushroom to reveal cluster-wide perspectives.

Everyone involved in running HPC clusters needs metrics. Hybrid-architecture, distributed systems, however, are a challenge to monitor, especially as systems scale out. Given the importance and inherent difficulties, it is not surprising that monitoring is once again topical in the HPC community. In fact, recent discussions are calling for an extreme makeover, by expressing the need to modernize monitoring for HPC clusters.

Read More >

Bright Cluster Manager for Apache Hadoop: The First 6 Months of a Radiant Relationship

By Ian Lumb | September 26, 2014

Last week we released support for Apache Hadoop 2.5.0. I thought it would be a good time to revisit all the updates we’ve made over the last 6 months - in case you missed them.

Latest Updates

Our latest updates include:

  • Apache Hadoop 2.5.1  Typo? Not exactly. We introduced support for Apache Hadoop 2.5.0 last week. Since Version 2.5.1 (released on September 13, 2014) is a minor release that builds upon the stable 2.4.1 release, you can now expect support via YUM updates to Bright. Note that vanilla Apache Hadoop includes HDFS (the Hadoop Distributed File System), ZooKeeper (the coordination service) as well as YARN (the workload manager).
  • Cloudera CDH 5.1.0  We introduced support for CDH 5.1.0 in mid-July.
  • Hortonworks HDP 2.1  We introduced support for Hortonworks Data Platform (HDP) 2.1 in mid-May.

Bottom line: Bright Cluster Manager maintains support for all major distributions of Apache Hadoop. Because we’re obsessed with keeping you current, you don’t have to be. We free you up to focus on the analysis of Big Data - and wasn’t that why you got interested in Hadoop in the first place?

Read More >

CUDA 6.5: Something for Nothing

By Ian Lumb | September 17, 2014

Who says you can’t get something for nothing? With CUDA 6.5, you can! The something is improved performance. The nothing is no code change required.

NVIDIA made improved performance a focal point of the recent CUDA 6.5 release. For example, if you make use of CUDA-enabled FFTs and/or sparse-matrix routines, all you’ll need to do is recompile your application with this latest release of the toolkit. You’ll see performance improvements without having to change a single character in your source code. Even the size of object files can be optimized through use of the nvprune utility.

Read More >

Thousands of servers and not enough staff to manage them. What’s a poor IT manager to do?

By Matthijs van Leeuwen | July 15, 2014

Today’s data center managers have a lot on their plate. They’re responsible for more servers and storage than ever, and have to react quickly to frequent changes in demand for service. Administration and management solutions play a crucial role in helping sysadmins face these challenges.

Gartner refers to the four biggest trends that currently dominate the world of IT as The Nexus of Forces : mobile, social, cloud, and big data. Each of them holds out a great deal of promise for business owners. Mobile applications and social media provide companies with direct access to customers. Cloud services ensure that the business can always access its data and applications efficiently. Big data provides new insights into a company’s business processes and customer buying behavior. Companies that want to stay on the cutting edge of their industry would do well to immerse themselves in these technologies, and that means their data center must also be up to the task.

Read More >

Martijn de Vries to speak at ISC'14 in Leipzig, Germany

By Lionel Gibbons | June 17, 2014

 

For those attending the International Supercomputing Conference (ISC'14) in Leipzig, Germany next week, don't miss Bright Computing CTO Martijn de Vries' presentation at the Exhibitor Forum.

Martijn will present a session entitled Bright Computing: A Unified Management Solution for HPC and Hadoop ClustersHis session is scheduled for Wednesday, June 25th, at 11:00 AM.

Read More >

The Top 4 Reasons You Should Try Cloud-Based GPUs for HPC

By Ian Lumb | April 08, 2014

Have you tried cloud-based GPUs for HPC? If not, here are the top four reasons you might want to.

Read More >

Use Intel’s 20-Step Process to Choose a Management Solution for Your Cluster

By Ian Lumb | February 06, 2014

20 criteria to identify your solution for cluster management? Absolutely!

Read More >
COMMENTS