Monday 8 October 2007
Fish, Water
Surprisingly, I will be giving a talk tomorrow at the Auckland University Bioengineering Institute, titled "Small, Hot And Parallel: The Future Of Web Browsing (And Computing)".
Since I don't know anything about biology, I'll talk about Web browsers instead. I'm planning to borrow shamelessly from Ras & Co's slides (with permission) to speculate about how one might build an awesome browser for small devices given the physical constraints of both humans and computers. Should be fun.
Comments
I'm thinking that target-ting for 8 cores (for example) is probably too small. I think it'd be too tempting for a team in a rush to unintentionally leave nasty bottlenecks in their design. I've a feeling that hacks to a current browser could probably reach the point where it was using 8 cores fairly evenly. Best to target a "silly" number like, say, 100 cores.
Otherwise we'll have the same problem again in fiteen years when the new kilo-core machines come out.
I was also wondering if you could do much to improve browser performance by running predictive branches in parallel. eg. when the layout code hits a box to fill, it forks into two. One assumes the box will not need to resize, the other assumes it will overflow. Then when enough information comes in later you can kill the one that took the wrong path.
I guess it's not very efficient, and would need fast copy, but I'm just bouncing ideas around :)
- Colin
James
My slides don't really add anything beyond Ras' slides, which are online, and my comments in my blog, and since they contain a lot of Ras' slides, I'm a little reluctant to redistribute them.