Hallo Ivanko, Du schriebst am Sat, 18 Aug 2012 10:21:47 +0500:
> interval _that_ long! Milliseconds would be a far better fit, > ============== > Then CPUs would be used less efficient because of higher relative > losses on context switching. Me even think that the current switch of That's right, but at clock speeds of around 3 GHz (equivalent to cycle times of 0.3 ns - that means _millions_ of instructions to a couple ms. I.e. the overhead should be on the order of a few parts per a thousand. > few seconds is mainly designed to equalize temperatures of CPUs > (different zones of their crystal). It's a rather long time for that purpose - the dies are tiny and heat up in fractions of a second. Switching heat production from one part to another in seconds will constantly stress them thermally, somwhat like if you heat a big sheet of metal with a torch at one spot a time, constantly switching spots - that will warp your metal sheet into a mountain range. Now these crystal platelets are very brittle and don't like warping at all, they tend to break... (Maybe I really shouldn't use this machine under Linux any more, but switch to Windows... >-I But what about the other cores then? :-O ) -- (Weitergabe von Adressdaten, Telefonnummern u.ä. ohne Zustimmung nicht gestattet, ebenso Zusendung von Werbung oder ähnlichem) ----------------------------------------------------------- Mit freundlichen Grüßen, S. Schicktanz ----------------------------------------------------------- ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ mseide-msegui-talk mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mseide-msegui-talk
