There’s an interesting little paper published in Login on, er, well, “Benchmarking Amazon EC2 for High-performance Scientific Computing” (PDF). Basically it shows that whilst there’s promise, there’s also a massive gulf between what you can currently do on EC2 and on a regular HPC cluster in terms of MPI jobs.
(Hat tip)
Oh dear, it looks like Denis Ritchie’s web pages, and all associated documentation, has gone 404 at Bell Labs and, worse still, it’s been removed from archive.org due to the Bell Labs robots.txt file.
Now I really wish I’d grabbed a copy of the B documents he had there! 🙁 If anyone has a copy or an alternative location if it’s moved elsewhere I’d love to hear from you..
Update: Thanks so much to Ian for providing copies and pointing out that the original webpages and documents have re-appeared.
A press release from the University of British Columbia (found via a tweet on the Planck CMB telecopes Twitter feed) talks about their use of the KDE program KST, a real-time large-dataset viewing and plotting tool, chosen because of the amount of data that would be generated:
But the cameras will produce a large amount of scientific data to process–with the LFI instrument alone producing more than 100 Gigabytes a year. Traditional data plotting and analysis packages like MATLAB and IDL wouldn’t cut it.
Both UBC and the University of Toronto have been involved with the development of the KST project, and the Canadian Space Agency has contributed funding to it.
I so hope this finally fixes my Kopete TLS problems! Need to wait for the Kubuntu packages to appear..
Update: the Ubuntu Jaunty packages are currently in the Kubuntu experimental PPA. Downloading now..
Update 2: Upgrading ain’t for the faint hearted, I’ve had to remove packages, used dpkg --force-overwrite -i /var/cache/apt/archives/$foo.deb a couple of times and generally had a fun hour or so trying to upgrade. All done now so this is being edited with the new Konqueror:
Version 4.2.85 (KDE 4.2.85 (KDE 4.3 Beta1))
Using KDE 4.2.85 (KDE 4.2.85 (KDE 4.3 Beta1))
We’ve got Kepler now starting it’s science mission after finishing commissioning, Atlantis at the Hubble Space Telescope for the final repair mission and in just over 1 hour the ESA will launch a single rocket with both the Herschel infra-red space telescope and the Planck CMB telescope. That’s two very expensive satellites sat on top of a rocket with a rather chequered launch history!
Best of luck folks (which is also what the Mars Exploration Rover “Spirit” needs given it’s gotten itself stuck on Mars!)..
Update: So far so good, first and second stages ran and separated OK, third stage running at the moment prior to separation of Herschel and Planck.
Update 2: Third stage burn complete, in ballistic phase.
Update 3: Herschel separated from third stage!
Update 4: The cylinder protecting Planck has been jettisoned.
Update 5: Planck has separated from the third stage!
That’s it for me for tonight, it’s almost midnight here…
Update 6: ESA has established communications with both telescopes and currently they’re both looking good for their journey to L2.
A few weeks ago Martin Sevior and Tom Ffield of the University of Melbourne did a talk at VPAC called “Belle Monte-Carlo production on the Amazon EC2 cloud” based on a paper they’d presented at the International Conference of Computing in High Energy and Nuclear Physics. The presentation is now available on the VPAC website.
It’s all about testing the cloud computing model via Amazon EC2 for Monte Carlo production for the SuperBelle experiment at the KEK collider in Japan. My favourite comment is that for a real full production run on Amazon EC2 to be useful it would need to be able to return data from S3 to the KEK collider at 600MB/s (~4.7Gb/s) sustained.
I don’t know what Amazon would say to that – well, apart from maybe “no”. 🙂
NB: This is the talk I mentioned in the comments on Joe Landman’s blog post called “Cloudy Issues“.
So going to SGI’s website now redirects you to Rackable, which is now badged as SGI. Then if you click on the Silicon Graphics brand you’ll end up back on a page on the original SGI website! Confused ? You will be.. 🙂
SGI!? Don’t you mean Rackable Systems!? No, I mean SGI. As you recall, Rackable successfully completed a bid process to acquire the assets of Silicon Graphics/SGI. The new organization has been named, you guessed it, “SGI.â€
They themselves explain it like this:
First, our corporate name will be Silicon Graphics International, but from a marketing and branding perspective, we will be referred to as SGI®. The Rackable name will become the brand for the SGI x86 cluster compute products. Rackable will join our other industry-recognized brands – such as ICE Cube, Altix, InfiniteStorage, CloudRack, MicroSlice, Origin, and VUE – to comprise the new SGI.
My brain hurts.. It also looks like they’re going to continue trying to sell the Itanic based Altix systems, which surprises me.
Wessex Archaeology (a registered charity in the UK) have the Time Team excavation reports from 2005-2009 on their website as part of their Time Team site.
They’ve also got a number of blogs (and an aggregator for them), an awesome HD LIDAR flyover of the Stonehenge World Heritage Site, a Twitter feed and a Flickr stream including a Time Team set. Nice one folks! 🙂
Having just watched the first showing of the Time Team Friars Wash dig on the ABC I went searching for any post excavation reports and managed to find one on Scribd written by Wessex Archaeology (where W.A. seem to be putting up a number of reports) which has some interesting follow ups to what’s seen in the program, for instance:
You can download the PDF of the report if you login to Scribd (they support OpenID, but not OpenID redirections like WP-Yaddis can do to WordPress.com, etc).
Joseph Reeves has an interesting little story about how a badly designed e-commerce system results in both an obvious security flaw and a missed opportunity for better customer relations over in Oxford. This all started when the bus driver refused to accept the email displayed on a phone, but wanted a paper copy, even though he had no way of verifying either as being legit! Quite how the driver thinks that printing out a copy of an email you have makes it more legitimate than the electronic version is left as an exercise for the bemused.
Exploiting the hole in the Oxford Bus Company’s awful system is easy; print a genuine looking email that contains the details of any bus journey you want to take. The bus driver only wants to see an email and your ID, they have no access to any passenger lists; should anyone with a passenger list board the bus (some kind of ticket inspector, I guess), you can remind them that the customer is always right; furthermore, act mortified at the fact that they blame a failing of the booking processing system to register your journey as some sort of criminal action on your behalf.
Josephs suggestion of how this should work is straightforward:
Fixing this bus ticket problem would be very simple – the Oxford Bus Company just needs to generate a unique ID number that it includes in emails to customers and to provide drivers with access to a passenger database. Buses are already fitted with Internet connections to be used by passengers on the journey, so all that needs to be provided is a very simple device to the driver.
Leading to possible improvements in service, like:
A passenger boards the bus, hands over their ID and says “my number is 546672”, the driver taps this into the machine and replies “ah yes, hello Mr Reeves, I’ll let you know when we’re at Heathrow Central bus terminal”.
It may be that the Oxford Bus Company has done a risk assessment and believes that the loss due to the fraudulent use of the service is lower than the cost in equipment and time taken for the driver to validate a ticket, but I doubt it. In contrast the Melbourne SkyBus service (from memory) lets you buy online, you then have to print out a ticket which has a barcode on it which the driver scans to validate it. Of course that means that you can’t carry it around electronically (well, not easily on a phone I guess) but does have the advantage that the barcode scanners are pretty quick, much faster than having to have someone type it in (and maybe get it wrong).