Humm,
that can be causing some damages. Not sure, though. Since i have posted the message i didn't try it anymore in the dual hyperthreading machine. The machine is to debug applications and the other guys using almost kill me! My application was simply destroying the server. In a dual pIII it seams there are no problem. The application stays stable between 80MB and 120MB (the same is true for any machine with a single processor or with a single hyperthreading one). Therefore, this is one of those problems to solve somewhere in the future. thanks, Manuel ----- Original Message ----- From: "Stoyan Damov" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, August 18, 2003 4:48 PM Subject: Re: [ADVANCED-DOTNET] GC,multithreading and multiprocessors > I have exhausted the physical memory once, when I was subscribing a new > instance of an object in a loop (not exactly a tight one), but forgot to > unhook the event handler in the finalizer (or earlier). Because the event's > delegates had a strong reference to each instance (via the Target property), > the objects I expected to be GCed quickly ate the whole memory. Look at your > code closely for similar scenarios... > > HTH, > > Stoyan Damov > > -----Original Message----- > From: Moderated discussion of advanced .NET topics. > [mailto:[EMAIL PROTECTED] Behalf Of Manuel Costa > Sent: 07/23/2003 01:16 > To: [EMAIL PROTECTED] > Subject: [ADVANCED-DOTNET] GC,multithreading and multiprocessors > > > Hi, > > I've been experiencing for some time problems with garbage collector in a > machine with 2 hyper threaded processors. > The application processes network messages for an online game > implementation. Under strong loads and in the multi-processor machine, the > application memory grows up to values far beyond the physical memory. Of > course, this destroys its performance! On the other hand, if the same > application runs on a single processor this behavior no longer exists. > My first thought when i saw this behavior was that i had some > synchronization problem, which was disguised in the single processor > machine. However, i've been looking to the code again and again and it seams > ok. Therefore, i think that a successive and intensive multi-threaded > environment of memory allocations and reallocations causes the problem. > > Does anyone have a reasonable explanation for this behavior? Does it make > sense or should i burn my eyes again and again on the developed code? > > > Manuel > > =================================== > This list is hosted by DevelopMentor� http://www.develop.com > NEW! ASP.NET courses you may be interested in: > > 2 Days of ASP.NET, 29 Sept 2003, in Redmond > http://www.develop.com/courses/2daspdotnet > > Guerrilla ASP.NET, 13 Oct 2003, in Boston > http://www.develop.com/courses/gaspdotnet > > View archives and manage your subscription(s) at http://discuss.develop.com =================================== This list is hosted by DevelopMentor� http://www.develop.com NEW! ASP.NET courses you may be interested in: 2 Days of ASP.NET, 29 Sept 2003, in Redmond http://www.develop.com/courses/2daspdotnet Guerrilla ASP.NET, 13 Oct 2003, in Boston http://www.develop.com/courses/gaspdotnet View archives and manage your subscription(s) at http://discuss.develop.com
