The OP needs lower-priced spot instances, which are intermittently available and designed exactly for this workflow.
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.
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.
I 100% agree. And f2py is the perfect way to combine the advantages of Fortran + Python. It beats SWIG any day.
I write lots of numerical codes for solving engineering problems, and most of the time I use Fortran, even though I have written some in C as well. Ultimately, the easier memory management and advanced array support is worth using a language considered by many to be strange and unusual. Of course, this is only for numerical analysis. Most other things I write in Python.
Likely the most serious math book you will find in a retail, consumer bookstore. An excellent read and essential to truly understanding modern physics.
Oh, I'm sure there's loads of "special recipes" that people have come up with to solve a particular problem. But, for any single problem, there's probably a dozen recipes that equally solve it. The reason why no one patents them is because they're so easily substitutable and they don't make $$$. Hint: that's the most important part... if it doesn't make $$$, people don't patent it... and it's not really worth reverse engineering that particular recipe (unless you have to solve the problem, and then you can come up with just about any chemical that does the job).
Chemical compounds can be very easily reverse engineered (for example, small molecule drugs). You take a sample of it and put it in a mass spectrometer. The molecule gets fragmented and the identity of each fragment is determined by a mass-to-charge ratio and other observations. Then, from a very very very (very!) large database of identities of fragments, the original molecule can be determined. For really small molecules (many drugs), you don't even need to fragment it. This all takes about 20 minutes.
The barrier to entry for reverse engineering drugs is not determining the identity of the drug. It's a legal barrier (ie. its patent). Once the 17 years are up, a company can produce a generic version in about 6 months (compared to the 5-10 years of development time).
From the same Wikipedia article, they manufacture Psuedoephedrine using yeast fermentation. Biological catalysis is typically chirally specific, so it is quite possible that only the D- form is produced. They would need to find a new manufacturing method that is as cheap.
Action Quake (a QII mod) was the (immediate?) predecessor to Counterstrike. It had the mp5, m3 rifle, sniper, and shotgun with the added realism that made Counterstrike so popular. One of the CS founders was a dev on it (I thought it was Gooseman...?). In Action Quake, if you ran then you're aim accuracy was reduced (more disperson) and if you crouched then you're aim accuracy was improved.
Did Gooseman start another QII mod (Navy Seals) before moving to HL?
SCO is like that child who is constantly shouting, obnoxiously, "Look at me! Look at me!" as they do something stupid. It might be funny for the first... hour? But when real work needs to get done and that child is pulling the same shit, it's time for a Time Out.
So hurrah for SCO's bankruptcy. Now the rest of us can finally get some proverbial litigation silence.
If the Wiimote actually works as expected, then I think both the "hardcore" gamers AND the casual gamers will snatch this console up. The hardcore gamers will get it to experience a truly innovative gaming style (and for only $250, why not??) and the casual gamers will get it because a) it's cheaper than the others (again, only $250!), b) it's got Mario Brothers/Zelda/etc, and c) the Wiimote looks cool.
Cellular automatons can "look" like some physical process, but that doesn't mean the two have any casual relationship whatsoever. I think Wolfram forgot that Correlation != Causation.
Or, more likely, he absolutely knows that the work is crap and so he publishes it in a book rather than submitting it to peer review in a respectable mathematical journal.
And, before I get a nasty reply, let me make this clear:
Science is about PROVING or DISPROVING a hypothesis. (Or, at least, making the attempt to do so.) Does Wolfram do this? Absolutely not. The title of his book makes sense, though. It is a new kind of science...the bad/wrong kind with zero consequences or illumination.
My gut feeling after finishing Ep1 was that it was too short for $20. You can argue for/against this all you want, but I'm sure many customers felt exactly the same way.
I think Ep1 is worth ~$15-12 and no more. $3/hour of gameplay is high enough for a single player game. Now...if Valve created a multiplayer game that was more interesting than deathmatch then I would feel better about spending the $$. At least with multiplayer games, there's a lot more replayability and you can seriously get hours and hours out of it before it becomes stale.
Ultimately, it comes down to how many people buy Ep2. If it's the same length and the same price, I think I'll just wait to buy it...until the price drops or a combo pack comes out. I hope Valve prices it more reasonably nonetheless because I did like Ep1, but not enough to pay the $20.
What are the chances that a related package using Python instead of Java will be released/developed some time in the near future?
I don't know the technical obstacles to doing this (and please elaborate on this if you want), but I think it'd be a useful addition. I'm sure there's lots of people who use Python, but not Java (for a variety of reasons).
Unless you're the guy at the top. In that case, YOU are the one handing down policies and YOU decide the color/shape of both the tent and the sandbox. It's good to be the boss.
But, of course, more $$ usually means more stress. But some people like the stress while others cave in from it.
Slashdot Mirror
Hmmm, low-digit Slashdot accounts just mean you've been a nerd for a very long time ...
Same sig since 2003....
The OP needs lower-priced spot instances, which are intermittently available and designed exactly for this workflow.
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.
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.
That just means he's ignorant AND old.
I 100% agree.
And f2py is the perfect way to combine the advantages of Fortran + Python.
It beats SWIG any day.
I write lots of numerical codes for solving engineering problems, and most of the time I use Fortran, even though I have written some in C as well. Ultimately, the easier memory management and advanced array support is worth using a language considered by many to be strange and unusual. Of course, this is only for numerical analysis. Most other things I write in Python.
Go ahead. Argue. I dare you.
v same sig since 2002. v
I'll 4th that.
Wow, nigh universal consensus on Slashdot. Next, it'll be snowing in April in the Mid-Atlantic! (erp, nevermind)
The Road to Reality : A Complete Guide to the Laws of the Universe
by Roger Penrose
http://www.amazon.com/Road-Reality-Complete-Guide-Universe/dp/0679454438
Likely the most serious math book you will find in a retail, consumer bookstore. An excellent read and essential to truly understanding modern physics.
Sojourn rocked.
Did you know the creator of EQ was a guild leader on Sojourn?
Sojourn was better.
Oh, I'm sure there's loads of "special recipes" that people have come up with to solve a particular problem. But, for any single problem, there's probably a dozen recipes that equally solve it. The reason why no one patents them is because they're so easily substitutable and they don't make $$$. Hint: that's the most important part ... if it doesn't make $$$, people don't patent it ... and it's not really worth reverse engineering that particular recipe (unless you have to solve the problem, and then you can come up with just about any chemical that does the job).
Chemical compounds can be very easily reverse engineered (for example, small molecule drugs). You take a sample of it and put it in a mass spectrometer. The molecule gets fragmented and the identity of each fragment is determined by a mass-to-charge ratio and other observations. Then, from a very very very (very!) large database of identities of fragments, the original molecule can be determined. For really small molecules (many drugs), you don't even need to fragment it. This all takes about 20 minutes.
The barrier to entry for reverse engineering drugs is not determining the identity of the drug. It's a legal barrier (ie. its patent). Once the 17 years are up, a company can produce a generic version in about 6 months (compared to the 5-10 years of development time).
From the same Wikipedia article, they manufacture Psuedoephedrine using yeast fermentation. Biological catalysis is typically chirally specific, so it is quite possible that only the D- form is produced. They would need to find a new manufacturing method that is as cheap.
Strut Strut ;)
Ah, gotcha.
AHL felt like a completely different game than AQII.
Man, I had forgotten about the hand cannon. That was a blast. (literally and figuratively!)
Action Quake (a QII mod) was the (immediate?) predecessor to Counterstrike. It had the mp5, m3 rifle, sniper, and shotgun with the added realism that made Counterstrike so popular. One of the CS founders was a dev on it (I thought it was Gooseman...?). In Action Quake, if you ran then you're aim accuracy was reduced (more disperson) and if you crouched then you're aim accuracy was improved.
Did Gooseman start another QII mod (Navy Seals) before moving to HL?
SCO is like that child who is constantly shouting, obnoxiously, "Look at me! Look at me!" as they do something stupid. It might be funny for the first ... hour? But when real work needs to get done and that child is pulling the same shit, it's time for a Time Out.
So hurrah for SCO's bankruptcy. Now the rest of us can finally get some proverbial litigation silence.
Espresso Beans, Chocolate, Dying Awake.
He Insists, Thesis, I'm Remiss.
Hear, Hear!
If the Wiimote actually works as expected, then I think both the "hardcore" gamers AND the casual gamers will snatch this console up. The hardcore gamers will get it to experience a truly innovative gaming style (and for only $250, why not??) and the casual gamers will get it because a) it's cheaper than the others (again, only $250!), b) it's got Mario Brothers/Zelda/etc, and c) the Wiimote looks cool.
The Wii will Win. 'Nough said.
Too bad Wolfram's book isn't science either.
Cellular automatons can "look" like some physical process, but that doesn't mean the two have any casual relationship whatsoever. I think Wolfram forgot that Correlation != Causation.
Or, more likely, he absolutely knows that the work is crap and so he publishes it in a book rather than submitting it to peer review in a respectable mathematical journal.
And, before I get a nasty reply, let me make this clear:
Science is about PROVING or DISPROVING a hypothesis. (Or, at least, making the attempt to do so.) Does Wolfram do this? Absolutely not. The title of his book makes sense, though. It is a new kind of science...the bad/wrong kind with zero consequences or illumination.
My gut feeling after finishing Ep1 was that it was too short for $20. You can argue for/against this all you want, but I'm sure many customers felt exactly the same way.
...until the price drops or a combo pack comes out. I hope Valve prices it more reasonably nonetheless because I did like Ep1, but not enough to pay the $20.
I think Ep1 is worth ~$15-12 and no more. $3/hour of gameplay is high enough for a single player game. Now...if Valve created a multiplayer game that was more interesting than deathmatch then I would feel better about spending the $$. At least with multiplayer games, there's a lot more replayability and you can seriously get hours and hours out of it before it becomes stale.
Ultimately, it comes down to how many people buy Ep2. If it's the same length and the same price, I think I'll just wait to buy it
What are the chances that a related package using Python instead of Java will be released/developed some time in the near future?
I don't know the technical obstacles to doing this (and please elaborate on this if you want), but I think it'd be a useful addition. I'm sure there's lots of people who use Python, but not Java (for a variety of reasons).
Thanks!
Salis
Unless you're the guy at the top. In that case, YOU are the one handing down policies and YOU decide the color/shape of both the tent and the sandbox. It's good to be the boss.
But, of course, more $$ usually means more stress. But some people like the stress while others cave in from it.