No comments yet.
No trackbacks yet.
10 reasons my O(n²) algorithm is better than your O(n) algorithm
about 1 year ago - No comments
OK, so it’s a controversial title, but the point I want to make is that programmers often get confused about the time complexity, i.e. order O() of an algorithm and performance. They will happily declare that their O(n) algorithm is so much better than someone else’s O(n²) algorithm as they rip out some simple, stable, More >
Measuring website browser performance
about 2 years ago - No comments
www.webpagetest.org is a great resource for anyone looking into web site performance.
Shaving seconds off page load times
about 2 years ago - No comments
IIS6 comes with some antiquated settings for Gzip compression. Even after enabling it in the IIS management console nothing happens until you edit the metabase file to enable the files you really want to compress. With the correct settings the main page for my site now loads in 5.3s instead of 6.0s and subsequent loads More >
Optimization Advice
about 2 years ago - No comments
Optimization is the root cause of much unmaintainable code. It is also the root cause of many bugs. My advice to the vast majority of programmers is to stay away from it: if you have a good architecture you should not need to optimize it at this level. Personally I divide optimization into three buckets: More >
It’s all about disk speed
about 3 years ago - No comments
Recently I upgraded to a 10,000 rpm drive in my desktop and a 7,200 rpm drive in my laptop. Quite frankly this is the most cost effective way to improve performance on any machine. CPU speeds have grown dramatically over the last few years, so has memory capacity and CPU to memory bandwidth, but disk More >
Cache optimized scanning of pairwise combinations of values
about 4 years ago - No comments
Suppose you wanted to calculate all of the pairwise combinations of the values in a vector of n elements. The result would be a 2D matrix of dimensions n squared. A naive approach to this would be two nested for loops evaluating each of the n squared values. But now suppose that the values in More >
Amazon High-CPU Medium Instance is about 30% slower than a dedicated core 2 duo server