Re: [hlds_linux] Deficient Linux performance (was: hardware?)
-- [ Picked text/plain from multipart/alternative ] The windows server version is generaly better because it just is. Valve designed HLDS for windows. And then made a linux port. On 9/25/05, ScratchMonkey <[EMAIL PROTECTED]> wrote: > > --On Saturday, September 24, 2005 8:35 PM -0500 "Eric (Deacon)" > <[EMAIL PROTECTED]> wrote: > > > on the same hardware, a Windows HLDS install > > will perform better than a Linux hlds installation > > As a wild guess, I'll assume you mean some kind of benchmark within the > binary gives better numbers on Win32. Which Win32? What do you suppose > accounts for the difference? Poor compiler? Choice of optimizations? > Specific other processes stealing cycles? (Recall that neither Win32 nor > Linux are monolithic and exact objects; both are assemblies of varying and > versioned components.) > > If a Windows install is better, we should determine exactly what the issue > is that keeps Linux from matching it. The first step to that objective is > to establish exactly what we're measuring. > > ___ > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > http://list.valvesoftware.com/mailman/listinfo/hlds_linux > -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Deficient Linux performance
In a bold display of creativity, ScratchMonkey wrote: If a Windows install is better, we should determine exactly what the issue is that keeps Linux from matching it. The first step to that objective is to establish exactly what we're measuring. By the way, IIRC the conclusion reached by people who know far more about such things than I ever will is that the whole thing was and continues to be written on Windows for Windows and is simply ported to Linux, and somewhere in there there's a difference in optimizations and what things can be taken advantage of on one platform versus another...yada yada yada. In other words, it's not to say there's anything inherent to Windows itself that provides better performance, nor is there anything inherent to Linux that makes it worse, but rather that being a native Win32 app ported to Linux winds up with the native Win32 version performing better on the same hardware. But again, this isn't something I bleed over, nor does it keep me up a night, nor do I care to be involved in any "we" that sets out to solve this mysterious riddle. -- Eric (the Deacon remix) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] Deficient Linux performance
In a bold display of creativity, ScratchMonkey wrote: What do you suppose accounts for the difference? Poor compiler? Choice of optimizations? Specific other processes stealing cycles? (Recall that neither Win32 nor Linux are monolithic and exact objects; both are assemblies of varying and versioned components.) I'd suppose it's the magical flying monkeys that just came shooting out of my ass and are now fluttering about the room painting murals on my walls with their feces. I don't know, and I don't care, though you're welcome to try to throw around things you may or may not understand (most of it's over my head) about things which you certainly don't know the specifics (closed source, sir). It seems to be inherent to the system regardless of kernel versions or other processes on the box, etc. If a Windows install is better, we should determine exactly what the issue is that keeps Linux from matching it. The first step to that objective is to establish exactly what we're measuring. Have fun with that. -- Eric (the Deacon remix) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux