3 ms·
And yet we do that! How fast is that CPU? You'll hear people quoting clock speed. You'll hear people quoting FPOS (floating point operations per second -- at l
by mcherm 10y ago
And yet we do that!
How fast is that CPU? You'll hear people quoting clock speed. You'll hear people quoting FPOS (floating point operations per second -- at least we USED to quote that). Both of which are single tasks.
It's not that we don't realize that the performance of a CPU varies depending on lots of things like instruction set design and (especially) memory pipelines and caching. It's just that there is not a general "does the stuff you want quickly" benchmark to measure (or rather, there ARE several such benchmarks, but each is skewed in its own way and not subtly, so that things like clock speed and FPOS are at least flawed in OBVIOUS ways). And there is some sort of very rough correlation: CPUs with greater clock speed do tend, as a general rule, to run most applications faster.
I think rating the spider's intelligence by giving them a web-making challenge was a really BRILLIANT idea, and provided a better assessment of intelligence than any other test I can imagine giving to a spider. Can you do better?
- cgriswald 10y agoI wasn't questioning the methodology. I think this was a good experiment and really interesting result. I was questioning the shocking surprise at the result as expressed by the article. A spider's brain is not a general purpose computing device. It has specific evolved functions. Performing that function well is necessary for the continuation of the species; both species have continued to survive, so both species probably perform the function well; no surprise.