Wikipedia:Reference desk/Archives/Computing/2016 July 21

= July 21 =

Increasing computer speed
Where are scientists working on to make computations faster? Where is the bottle neck in computer nowadays? Where was the bottle neck 10 or 20 years back? — Preceding unsigned comment added by Hofhof (talk • contribs) 00:02, 21 July 2016 (UTC)


 * That's a massive topic... how about checking out the most recent issue of XRDS, the ACM's magazine for students interested in computing machinery? It's the premiere review magazine for students in this field, and a good place to start to help you get oriented towards more specific current research topics.  Nimur (talk) 00:58, 21 July 2016 (UTC)


 * CPU power dissipation is the main obstacle to faster sequential computation. The main obstacle to parallel computation is that we're bad at designing parallel algorithms (except for embarrassingly parallel problems). -- BenRG (talk) 07:07, 21 July 2016 (UTC)
 * Alternatively, we could say that the obstacle is that many interesting problems are simply not amenable to parallelization. Kind of tough to say what could be done if we were cleverer, and could develop parallel routines analogous to e.g. god's algorithm. SemanticMantis (talk) 15:48, 21 July 2016 (UTC)
 * You might be interested in Supercomputer. Vespine (talk) 02:11, 22 July 2016 (UTC)