Problem with XP scheduler?
in [Win32 Kernel]
|
From: Jonathan de Boyne Pollard on 18 Jan 2010 10:50 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> <blockquote cite="mid:%23gQQWIJlKHA.2188(a)TK2MSFTNGP04.phx.gbl" type="cite"> <p wrap="">I really am not trying to be antagonistic here, but you make it very difficult not to be.<br> </p> </blockquote> <blockquote cite="mid:uGhoWsglKHA.4408(a)TK2MSFTNGP04.phx.gbl" type="cite"> <p wrap="">I agree with Hector, he does not think appropriately or take appropriate action based on what he may or may not have read (okay, let's just say he doesn't read well).</p> </blockquote> <blockquote cite="mid:uKJVYUFlKHA.5020(a)TK2MSFTNGP02.phx.gbl" type="cite"> <p wrap="">You are now mentioning this in every post, [...] .<br> Most of your posts have been very vague along the lines of "Can't you see that Windows is terrible and all kinds of bad things happen?". [...]<br> I think we are all growing tired of blanket statements like this.<br> </p> </blockquote> <blockquote cite="mid:%23Q7PtKEmKHA.4148(a)TK2MSFTNGP05.phx.gbl" type="cite"> <p wrap="">If you didn't want to hear the details, you should not have followed up almost every one with an argument.<br> </p> </blockquote> <p>Oh for goodness sake! Can you <em>please</em> put two and two together and make four? You're being trolled. I pointed this out in this thread <em>last year</em>. M. Grigoriev <a href="news:///uyekvvPiKHA.2164(a)TK2MSFTNGP02.phx.gbl">pointed to</a> previous trolling threads, with the same "Goldmine" theme, at the end of December 2009. <strong>Please stop feeding.</strong> If what you wrote above isn't a clue to the fact that you're being trolled, then surely the presence of several <a href="http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/unmistakable-marks.html">unmistakable marks</a> and the fact that this person has repeatedly claimed not to "have time for this", whilst simultaneously posting up to 6 messages per day to the thread, should be. Leave the Programmers' Landmine troll alone. Please go and help someone who is <em>actually here for help</em>, such as M. Scott with xyr <a href="news:///6302F077-6D47-4EF9-9BC9-FB85364BBA61(a)microsoft.com">heap corruption problem</a> for example.<br> </p> </body> </html>
From: Hector Santos on 22 Jan 2010 05:25
tanix wrote: > > So, one more time for expecially gifted kind. > > It there a problem with XP scheduler on a multi-core system? YES! > > The answer is either "yes" or "no". YES!! > If yes, what IS the problem? I heard some people say it might be the battery on the motherboard. But that doesn't make sense, to me, because that would only mean that your computer will work slower when the battery is slow or if there has a short circuit, the heat flux would generate an alarm. Do you hear any beeping? Do you have the CPU and motherboard temperature monitors on? But I don't think its the battery, but something I call "Natural Black Space Quantum Holes" or NBQ for short. You see, ops codes have frequencies. When the sum of the op codes reach the value of a QUANTUM, there is a context switch. This is a natural quantum switch, not one generated by an interrupt or some forced residence time for threads. However, in a multi-core, I think there is a "Black Hole" during a context switch with a time so small, that might not be an issue most of the time and may not be detectable with properly below zero instruments that will slow down time. However, as more cores are added, this "black hole" frequency becomes deterministic - i.e, someone will feel it without instruments. Now, Explorer is more susceptible to NBQ holes because it relies on video interrupt frequencies lining up with context switching. So the more interrupts the wider the NBQ time spectum becomes. The windows scheduler could address this better synchronized the video interrupts with the CS generated NBQ holes. Everyone knows about it and it is well known there is a real simple fix, but I can only suspect MS won't fix it because APPLE has filed a patent 12312h232h433: "A solution for Multi-Core Video/NBQ Misfires On Windows Operating Systems." So maybe, and this is only a speculation, MS refuses to pony up money to license Apple's patent. Go Figure! -- HLS |