From: Brian Osman Newsgroups: comp.os.msdos.djgpp Subject: Re: 686 Date: Sun, 23 Feb 1997 13:13:17 -0500 Organization: Rensselaer Polytechnic Institute, Troy NY, USA Lines: 84 Message-ID: <331088BD.5832@rpi.edu> References: <5e8keu$cp3 AT chronicle DOT concentric DOT net> <5ebhn2$lsl AT freenet-news DOT carleton DOT ca> <330A2003 DOT 6191 AT post DOT comstar DOT ru> <01bc218d$be63fb20$b7407e82 AT thalidomide> Reply-To: osmanb AT rpi DOT edu NNTP-Posting-Host: darkwing.stu.rpi.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Steven Engelhardt wrote: > > Well I know I'm going to start some kind of nasty thread here, but I have > to defend my system. Well, I own a 6x86, and I don't want to start the thread either, because you're mostly right. And the original post was entirely wrong. > A Pentium Pro 200 will be faster on all accounts than the equivalent 6x86 > running a fully 32-bit OS like WinNT, Linux, etc. On a 16-bit OS the 6x86 > will have (marginal, virtually unnoticeable) better integer performance. > But that's not the killer. True. All of it. Pentium Pro is a faster chip. The 6x86 isn't meant to compete with it. It's designed to compete with the Pentium. > The worst part about the 6x86 is its ghastly FPU performance. A Cyrix 6x86 > 166 has approximately the floating point performance of a Pentium 90. Try > running Quake or a 3-D application and you'll notice a huge difference. The FPU is pretty slow, but not quite that bad. My P150 (clock speed 120) goes about like a Pentium 120 or 100. Although it is very difficult to determine all of this exactly. The 6x86 takes a very different approach to the FPU from the Pentium. Cyrix explains it quite honestly with benchmarks and all (from 3-D games and business apps) on their web page. http://www.cyrix.com/ > Another terrible problem with the 6x86 is its massive heat generation which > makes it very sensitive to reflections off of the bus. Microsoft > encountered this problem frequently enough to justify disabling the L1 > cache on the 6x86 in NT for stability reasons ; this led to huge > performance losses in the 6x86. There is a patch for the 6x86 under NT, > but it can introduce system instability. Of course, the 6x86L fixed many > of these problems. Yes. My chip was of a late enough revision that it did not have the problem. And I've heard that all it took was a call to Cyrix to get the latest version. Still, I can hear when my fan goes on or off due to excessive heat, and it's off most of the time. (I run my system about 12-14 hours at a time.) > There are also a few, relatively unfrequent incompatibility issues with the > 6x86 where applications refuse to run. I haven't hit this yet, although I know it's true. The worst part is when apps check the CPU ID and see that it's a 486. (CyrixInstead!) And say that the game (usually games in this case) needs a Pentium. I just ignore and continue. > The AMD K5 (the K6 isn't even out yet, but it promises to be killer) is a > great chip, but I believe they only have up to P-133 equivalent chips up to > now, so that puts them out of the race. The M2 should be a great chip. No experience with the K5, but I know that AMD had a few setbacks resulting in diminished clock speeds. More on the M2 in my next paragraph... > Cyrix's and AMD's are great for the money, but I would never choose it > deliberately over an Intel... yet. The competition promises to keep Intel > honest and working hard though. > > Steven Engelhardt > sengelha AT uiuc DOT edu When I bought this system last fall, the M2 (mentioned above) was one of the reasons I purchased it. I really couldn't afford a PPro, and comparing a Pentium to a 6x86 was a no-brainer. The 6x86 is (generally) faster at the same P-rating, and much cheaper. Plus, the Cyrix M2 is going to be amazing. Very large L1 cache, 32-bit optimizations (I run NT4...), MMX, etc. All socket compatible with my chip! If I bought a Pentium, I could get MMX later, but going to a PPro, (which the M2 will compete with) would require a new motherboard. Oh well, just my two cents... Brian