http://bitcortex.com/2009/05/25/in-defense-of-efficient-computing/ Artificial Intelligence, Neuroscience, Quantitative Finance and the unedited thoughts of a soon-to-be robot Fri, 30 Sep 2011 19:27:59 +0000 hourly 1 http://wordpress.org/?v=3.3.1 http://bitcortex.com/2009/05/25/in-defense-of-efficient-computing/#comment-39 Ken Dakin Mon, 12 Sep 2011 17:58:38 +0000 /2009/05/25/in-defense-of-efficient-computing/#comment-39 Rod, I completely agree with your article and always wonder where people are coming from when they happily waste resources just because they can get away with it. Having faster processors is good because we can do lots more work for the same cost (or get faster answers to time critical questions) - not because we can acheive the same result by allowing ourselves to get lazy. I agree that for one-off programs that are run a couple of times it doesn't matter too much - but one-off programs have a tendency to outlive their creators and get modified to create more one-off programs that may also persist. This way lies madness (non-maintainable code that is both inefficient and inconsistent). For production programs we should construct solid frameworks that are as efficient as possible (and as re-useable as possible unless this conflicts with performance to a great extent). Most of the programs and sub-routines I produced over 40+ years were written with efficiency constantly in mind and yet they also had close to zero bugs and needed few changes if any over their very long lifetimes of 20 years or more. I produced instruction set simulators for test/debugging purposes where zero tolerance was the only option (otherwise the products would have been worse than useless and given false positives). I would like to say that I very much enjoy your blog and I am as impressed, as you clearly are, by the awe inspiring JFK speech which always gives me goosebumps each time I see it. My own wish is that world leaders get together soon and jointly commit to setting up a base on Mars by 2030 say. This might also kick start the world economy for years to come and inspire a new generation like nothing else could. My audacious vision is that we should send self assembly robots to carve out an underground base first using only materials to hand (e.g. to fashion most of the 'boring components' out of natural rock for instance - at ever increasing scales) and follow this with a manned expedition to secure and populate the prepared space. Rod,
I completely agree with your article and always wonder where people are coming from when they happily waste resources just because they can get away with it. Having faster processors is good because we can do lots more work for the same cost (or get faster answers to time critical questions) – not because we can acheive the same result by allowing ourselves to get lazy. I agree that for one-off programs that are run a couple of times it doesn’t matter too much – but one-off programs have a tendency to outlive their creators and get modified to create more one-off programs that may also persist. This way lies madness (non-maintainable code that is both inefficient and inconsistent).
For production programs we should construct solid frameworks that are as efficient as possible (and as re-useable as possible unless this conflicts with performance to a great extent). Most of the programs and sub-routines I produced over 40+ years were written with efficiency constantly in mind and yet they also had close to zero bugs and needed few changes if any over their very long lifetimes of 20 years or more. I produced instruction set simulators for test/debugging purposes where zero tolerance was the only option (otherwise the products would have been worse than useless and given false positives).
I would like to say that I very much enjoy your blog and I am as impressed, as you clearly are, by the awe inspiring JFK speech which always gives me goosebumps each time I see it. My own wish is that world leaders get together soon and jointly commit to setting up a base on Mars by 2030 say. This might also kick start the world economy for years to come and inspire a new generation like nothing else could. My audacious vision is that we should send self assembly robots to carve out an underground base first using only materials to hand (e.g. to fashion most of the ‘boring components’ out of natural rock for instance – at ever increasing scales) and follow this with a manned expedition to secure and populate the prepared space.

]]>