August 19, 2012
If you wondered where I’ve been the last year, I’ve been on Google+ and somehow didn’t feel like posting anything useful in the last couple of months.
However, I’ve got quite a few cool projects running:
- Got me a pair of Infiniband Adapters. Plan: Brew up a software which will receive data via IP on multiple IB-nodes and write the data to distributed shared memory via RDMA/IB. Then let one or more IB-nodes read from that (ring)-buffer and aggregate data so that it can be shoved into an RDBMS.
- Brew up an AWS image for easy BOINC-crunching while preserving the workunits on a headnode. Why? Because images on the AWS spot-market are cheap. But spot-machine do not retain any data. So I might be putting it into S3 along with some headnode directing which machine (machines id keeps changing!) can work on which workunits.
- Become more confident in the language Erlang. Got quite a few projects which could benefit from easy protocol prototyping with ASN.1 in Erlang. Stay tuned.
- Got me a Spartan-3E FPGA board. Not sure what to do with it, but it’s awesome :)
- Was playing around with GNU Radio and my RTL DVB-receiver recently.
- Doing some serious Openstreetmap work recently.
- Mrs. Janssen and be got fond of geocaching. Lot’s of outdoor stuff. Sweet!
- Been to this year’s Linuxbierwanderung. Was awesome!
I ain’t dead yet.
August 7, 2006
Via HPC Wire:
Dauger Research‘s PoochMPI is now working together with Wolfram Research‘s Mathematica, every student’s favourite equation-solver. The Mac-software will enable Mathematica kernels to communicate directly via the network in a cluster way, instead of using a traditional grid-approach where a master hands out work-units to slave-nodes, as HPC Wire points out.
Dauger build the software together with Advanced Cluster Systems, who already build the Math Grid Toolkit for the Mathematica/Mac.
The software will be available in six weeks.
August 1, 2006
Scientists and engineers from Cornell Uni, NCAR and Uni of Cincinnati report about their experience with Microsoft’s Windows Compute Cluster Server 2003.
To take the site’s words, “Customers report that Microsoft Windows Compute Cluster Server 2003 makes it easier and more cost effective to develop, deploy and use HPC systems, and they report that a Windows-based HPC platform integrates smoothly with existing IT resources.”
The whole article is a bit “Hooray, no more geeky Linux stuff!” but in the end they come to reasonable conclusions:
- MS focuses on small- and mid-sized clusters and not so much on large Tera-FLOPS HPC-systems.
- It focuses on easy integration and management of cluster-nodes into the existing infrastructure; submitting jobs to the cluster is easier.
- They want to make HPC-systems accessible easier for unexperienced end-users or research-groups, where expertise is non-existent.
- Not stated, but existent between the lines in the whole article: You can integrate existing Desktops into the Windows’ HPC-infrastructure.
And I nominate Mr. Matt Wortman from the University of Cincinnati for having the sexiest prediction in that article: He predicts that the carrier of the HPC-applications will be held on small USB- or Ethernet-connected devices which are enhanced by specialised hardware:
“Application developers will provide plug-and-play devices that integrate into your infrastructure via USB or Ethernet. These devices will be simple and capable of a small variety of very high-speed calculations. For example, a standalone bioinformatics server will store and analyze sequence data, or a drug discovery appliance will screen chemical compounds. These simple “unitaskers” will be made and supported for integration into your existing Windows-based environment.”
My conclusion: I can’t fight the feeling that this article is a bit sexed up by MS, although I can’t prove it. There are sentences which just don’t look sane to me:
“Q: What business needs are you solving with high-performance computing?
Wortman: For us, the key business need was reducing costs by eliminating complexity. We did that by eliminating Linux support costs.”
You decide yourself. The conclusions sound plausible though.
microsoft, windows, hpc, Windows Compute Cluster Server 2003