Linux Gains Two New Virtualization Solutions 
By Tyler Bletsch - Saturday, July 21, 2007, 10:38 AM - Systems, Software news
By way of slashdot, we find that the Linux kernel (2.6.23 and up) now sports three virtualization techniques out-of-box: KVM, Xen (just merged), and Lguest (also recently merged).

Lguest in particular looks interesting, as it doesn't require virtualization hardware support (like KVM), but is as simple as a single modprobe (as opposed to the Xen behemoth). Performance isn't too great right now, though (-30%).
  |  permalink  |  related link
Power Plays: How power consumption will shape the future of computing 
By Tyler Bletsch - Sunday, July 1, 2007, 08:52 AM - Power-aware computing
This Ars Technica article spotlights the development of power-aware technologies at the chip, system, network, and data center levels. It analyzes recent developments in terms of granularity, i.e. the frequency of reaction. Overall, a well written article.
  |  permalink  |  related link
Development and Optimization Techniques for Multi-Core Processors 
By Min Yeol Lim - Wednesday, June 27, 2007, 09:18 PM - Power-aware computing
This article talks about the performance issues on multi-core system. It basically recommends to use the parallelism, such as OpenMP, in order to take full advantage of it. It also gives the common issues limiting the performance. I think it's a very good article to summarize idea.

http://www.devx.com/go-parallel/Article/34428
  |  permalink
Summary of Power-aware computing research in OpenMP applications 
By Min Yeol Lim - Tuesday, June 19, 2007, 03:04 PM - Power-aware computing
Here are the paper list proposing research using the OpenMP library in Power-aware computing area.

1. Chun Liu, et. al, "Exploiting Barriers to Optimize Power Consumption of CMPs", IPDPS 2005.
This work is to use slack time among processors. By figuring out stall time at the end of each iteration, it reduces the frequency to save power without performance degradation. The evaluation in th paper is done only with simulator, not real experiment. SpecOMP is used to verify the idea.

2. Matthew Curtis-Maury, et. al, "Online Power-Performance Adaptation of Multithreaded Programs using Hardware Event-Based Prediction", ICS 2006.
This paper designed and implemented a framework that can adaptively regulate the concurrency level during program execution. So, the processors/threads configuration is changed based to achieve near-optimal energy efficiency. It build power/performance models and uses the hardware counters. For evaluation, 4 hyperthreaded Intel processors are used.

3. Jian Li, et. al, "Dynamic Power-Performance Adaptation of Parallel Computation on Chip Multiprocessors", HPCA 2006.
This paper proposes a heuristic method to determine # of processors and frequency level on one CMP node. All evaluation are performed on simulator. It does not expand the approach to multiple CMP nodes.

  |  permalink
Hybrid OpenMP / MPI programming  
By Min Yeol Lim - Tuesday, June 12, 2007, 01:46 PM - HPC/Scientific computing
This presentation file explains why the hybrid of MPI/OpenMP programming is required.
It comes with the examples and strategies.
Also, it talks about when the hybrid mode performs better.

http://www.nersc.gov/nusers/services/tr ... hybrid.ppt

Also, here are the paper list focusing on the performance in hybrid MPI/OpenMP applications.

1. Felix Wolf, et. al, "Automatic performance anlysis of hybrid MPI/OpenMP applications", Journal of Systems Architecture 2003.

2. Laksono Adhianto, et. al, "Performance Modeling of Communication and Computation in Hybrid MPI/OpenMP applications", ICPADS 2006.

3. Edmond Chow, et. al, "Assessing Performance of Hybrid MPI/OpenMP Programs on SMP Clusters"
  |  permalink
ext3cow - A file system that preserves the past 
By Tyler Bletsch - Wednesday, May 2, 2007, 12:31 PM - Storage technology, Software news
The ext3cow file system has been released for the 2.6 Linux kernel. This extension to ext3 adds "copy on write" (cow) functionality, which in turn allows the user to view the file system as it existed in the past. Unlike fixed checkpointing, this means that a user could dial in any time whatsoever, and see a consistent filesystem from that moment. It would be interesting to see what the performance characteristics of the system are.

If you're interested, you can get it from ext3cow.com; it consists of a kernel patch based on kernel 2.6.20.3, as well as some userland tools.
  |  permalink  |  related link
Using the S3 sleep state, even on a server 
By Tyler Bletsch - Monday, April 23, 2007, 10:01 AM - Power-aware computing
By way of slashdot, this Exoid article details how to make sure Windows uses the S3 sleep state, even for an always-on server. Also linked is this handy reference to the possible ACPI sleep states.

Perhaps something like this could be developed into a vehicle for low-utilization servers?
  |  permalink  |  related link
A great OS/systems web site 
By vince - Friday, April 20, 2007, 08:11 AM - General interest, Systems
Clarkson University Prof Jeanna Matthews has one of the best round ups of system papers. It comes as no surprise that she received her training at Berkeley. Kudos to her.

  |  permalink  |  related link
The Parallel Workloads Archive 
By Tyler Bletsch - Thursday, April 19, 2007, 10:32 AM - HPC/Scientific computing
The Parallel Workloads Archive is an archive of real workload submission and execution data from real clusters from 1993 to the present. It supports a unified log file format (the Standard Workload Format, SWF) so that data from diverse sources can be analyzed together. With over 3 million jobs from 21 clusters, there's plenty of information to be had. It's maintained by Dror Feitelson of the Hebrew University of Jerusalem.

You can either visit the original Parallel Workloads Archive or our mirror of the site.
  |  permalink  |  related link