Here's how to utilize lower-priced spot instances for scientific computing:

1. Set up one long-running, low-cost instance (a small is fine) that creates a distributed queue using Amazon's SQS, and adds jobs to the queue corresponding to each "unit" of the relevant computational problem of interest. New jobs can be added using a command line interface, or through a web interface.

2. Create a user start-up Bash script for the spot instances that runs your main program -- I prefer using Python and boto for simplicity. The main program should connect to the SQS queue, and begin an "infinite" while loop. Inside the loop, the next job off the queue is pulled, containing the input parameters that define the "unit" of the computational problem of interest. These input parameters are fed to the main algorithm, and the resulting output is uploaded to Amazon S3. The loop continues.

3. Any time the queue is empty or the spot instance remains idle for ~5 minutes, the spot instance then auto-terminates using EC2's command line interface.

4. Finally, just write a simple Python script to pull all the results off S3, combine & analyze them, and export to another useful format.

You'll also need to set up your spot instance price threshold, and make sure the queue has jobs to run. That's it, it's fairly simple.

Man, here I was hoping helm's deep would reinvigorate LOTRO. And yes, I had even pre-ordered it.

It should have been used on the computer before submitting this article.

Disappointing,

myke

Emergent OOism -- that everything is an object, including the variable types -- can provide continual surprises of what is possible, even to veteran programmers in other languages. As you were developing and using Python, Guido, what was your favorite surprise? What was now easily possible using Python that would have been very difficult with another language (at the time, or even nowadays)?

Mine: a dictionary of lambda functions for parsing text, and writing a custom MapReduce capability for AWS in 372 lines.

Hi Timothy,

Unfortunately, you didn't provide a lot of information in your post as to what the problems are.

As people have pointed out, there are a ton of USB to Serial solutions out there so having the modern hardware with the ability to communicate over RS-232 is generally not a problem (although, depending on the connections used, you might want to invest in a RS-232 breakout box and read up on RS-232 handshaking as many of the older devices do use hardware handshaking). I have a few hand wired 9 pin to 25 pin connectors with the CTS-RTS and DSR-DTR pins shorted together as they can simplify your life immeasurably.

In my experience, the biggest problem is retaining floppies & CDs with the original software on them (assuming that the developers are no longer supporting the product/are out of business). If the company is still in business, usually they're pretty good at providing updated software for their products. If they're not in business, then look to see if they were bought out by anybody. Chances are you'll find that the purchaser is still supporting the product, although it may be under another name.

Personally, the biggest issue that I see when I have encountered this type of situation is that the original programs are on floppies. If this is the case, you will need to find somebody with a Windows/95 machine that they're keeping together with spit, bailing wire, gaffer's tape and good intentions - you should be able to copy the program onto a USB key and then burn it on a CD/DVD for more permanent storage.

Once you have the program in a media that you can work with, you may have problems with the installation. You will probably have to create a virtual machine on your PC AND there may be 16 bit programs that you have to convert to 32 bit - here's a great resource that's saved me a couple of times: http://www.reactos.org/forum/viewtopic.php?f=22&t=10988

Finally, Google is your friend. Chances are the answers are out there for your particular equipment.

Good luck!

myke

Other than that?

I imagine it has some scientific value for showing how well a highly trained and experienced individual goes through a stressful situation, but other than that?

myke

Wouldn't the best answer for an individual be based on their driving habits and history?

If you tended to stay at the speed limit (or reasonably above according to traffic), were a defensive driver and were reasonably confident that you wouldn't cause an accident, wouldn't you want tracking on to show that it's the other guys fault?

Depending on your hubris level, the next step is a dashboard camera because clearly you are never going to cause an accident - right?

myke

That sounds *really* interesting - anybody know where the rest of the population can get Windows 3.1 licenses?

I doubt that Microsoft would be willing to release it into the wild - so where can we find them?

myke

Jörg,

With the recent controversy regarding assault weapons in the US and the plans to reduce clip sizes, I'm wondering if you would think your, um, devices would be more dangerous if they could shoot multiple objects rather than just one at a time?

myke

Interesting question and one I've thought about for a bit since the original question, from the other side came out.

I would say, bring in the team lead as an arbitrator. It's their job to direct the work and (hopefully) develop team members - this guy sounds like he needs a bit of development and level setting.

myke

Very insightful.

Somebody with points please mod the parent up.

myke

When I was at RIM, we used a broken quick sort method that the candidate was asked to fix. We didn't time how long it took the candidate to implement the fix, but it generally required the candidate to do some research as to what was wrong. One of the team leads created a simple app that tested the performance (ie speed) of the fix.

What was really interesting to me was the number of candidates who refused to do the test (50-60%) because they said it was "beneath them". The big problem was, RIM's HR (OD) that insisted we interview the candidatest that refused to do the test because we were losing potentially half the candidates that were responding to the job applications (this was when RIM was The Place To Be).

The best candidates were the ones that did the test and asked if we had any more. These candidates also tended to produce code that ran sort the fastest.

myke