Reply to thread

Message

[USER=259]@aquelito[/USER] Are you running without lid?

[USER=297]@Ceros_X[/USER] I guess you missed that Moore's law is failing. It will be at least 20 years before any alternative to electric transistors are developed and commercialized, if ever. Next step for Intel is 7nm and the time they will take to get to market will probably be 2-3 times the 14nm, I'm pretty sure those processors will not be more powerful than the current generation, they might save you a few watts, but they will not make an upgrade worthwhile.

You can clearly see this in the progress of C++ which started increasing again when individual processor cores peaked in 2003.

As to why we can't keep piling cores there are good reasons for that too... I was wondering where my 16 core desktop Intel processor was... and even if we had more cores writing async software to scale inside of one process is hard, I know since I wrote my own NIO webserver.

Next step for processors is probably async processors, but the same problem there; async is hard.

<blockquote data-quote="rupy" data-source="post: 36883" data-attributes="member: 1619">[USER=259]@aquelito[/USER] Are you running without lid?[USER=297]@Ceros_X[/USER] I guess you missed that Moore's law is failing. It will be at least 20 years before any alternative to electric transistors are developed and commercialized, if ever. Next step for Intel is 7nm and the time they will take to get to market will probably be 2-3 times the 14nm, I'm pretty sure those processors will not be more powerful than the current generation, they might save you a few watts, but they will not make an upgrade worthwhile.You can clearly see this in the progress of C++ which started increasing again when individual processor cores peaked in 2003.As to why we can't keep piling cores there are good reasons for that too... I was wondering where my 16 core desktop Intel processor was... and even if we had more cores writing async software to scale inside of one process is hard, I know since I wrote my own NIO webserver.Next step for processors is probably async processors, but the same problem there; async is hard.</blockquote>

[QUOTE="rupy, post: 36883, member: 1619"] [USER=259]@aquelito[/USER] Are you running without lid? [USER=297]@Ceros_X[/USER] I guess you missed that Moore's law is failing. It will be at least 20 years before any alternative to electric transistors are developed and commercialized, if ever. Next step for Intel is 7nm and the time they will take to get to market will probably be 2-3 times the 14nm, I'm pretty sure those processors will not be more powerful than the current generation, they might save you a few watts, but they will not make an upgrade worthwhile. You can clearly see this in the progress of C++ which started increasing again when individual processor cores peaked in 2003. As to why we can't keep piling cores there are good reasons for that too... I was wondering where my 16 core desktop Intel processor was... and even if we had more cores writing async software to scale inside of one process is hard, I know since I wrote my own NIO webserver. Next step for processors is probably async processors, but the same problem there; async is hard. [/QUOTE]

Verification