2 ms·
"For example, algorithms and data structures generally accepted as core knowledge do not take enough advantage of today's relatively high RAM availability." Wo
by plussed_reader 10y ago
"For example, algorithms and data structures generally accepted as core knowledge do not take enough advantage of today's relatively high RAM availability."
Wouldn't it be prudent to keep your algorithms as lithe as possible? I would think an algorithm/scheme that chews up large amounts of ramm would be perceived as inefficient?
Kinda like the dumpster fire that is Chrome, currently.
- ilaksh 10y agoCase in point. Read the Google engineering articles describing the work optimizing V8's memory usage. That primarily involved using profiling tools on lots of web pages and then changing the initial size of buffers. But they still have to use relatively large buffers from an old-fashioned CS algorithm perspective and they couldn't make them smaller without impacting performance the other way. And they can't know that without doing the profiling on the whole system. So yes you want functions to be lithe but you have to also focus some time, skills and knowledge on optimizing the entire system using tools and real-world data.