From owner-freebsd-questions Fri Apr 20 1: 5:36 2001 Delivered-To: freebsd-questions@freebsd.org Received: from oahu.WURLDLINK.NET (oahu.WURLDLINK.NET [216.235.52.1]) by hub.freebsd.org (Postfix) with ESMTP id B69FA37B423 for ; Fri, 20 Apr 2001 01:05:30 -0700 (PDT) (envelope-from vince@oahu.WURLDLINK.NET) Received: from localhost (vince@localhost) by oahu.WURLDLINK.NET (8.9.3/8.9.3) with ESMTP id WAA10834; Thu, 19 Apr 2001 22:05:17 -1000 (HST) (envelope-from vince@oahu.WURLDLINK.NET) Date: Thu, 19 Apr 2001 22:05:17 -1000 (HST) From: Vincent Poy To: Charles Burns Cc: , , , , , , Subject: Re: the AMD factor in FreeBSD In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Fri, 20 Apr 2001, Charles Burns wrote: > > Hmmm, isn't there one more thing about the P4 that it uses the > >PII's FPU and is less advanced than the PIII's? > > The P2's FPU /is/ the P3's FPU. The P3 is just a P2 with some new > instructions pasted on (SSE) and now with a tweaked L2 cache. Hardly the > generational gap between the P1 and the P2. I swear Intel is getting almost > 10% as bad as Microsoft--letting the marketing team run the whole company. Hmmm, I just remember reading that the P4's FPU is like worst than the P3's so it may be one that was even before the P2. > > Okay, I guess the reason I asked earlier about using PC133 SDRAM > >versus DDR is that I have 768MB (3 256MB) PC133 modules already on my > >Intel platform and it seems like the DDR has 184 pins versus 168 pins so I > >have to make a new investment so that's why I was asking if the > >performance hit will be big if I got one of the VIA KT133A chipset > >motherboards and ran a AMD 1.33Ghz DDR CPU on it or would that be a big > >no? > > If you have 768MB of SDR RAM don't even touch DDR. DDR may be twice as fast > but here in the real world it is only 1% to 10% faster. Hardly a reason to > spend hundreds of dollars on more RAM, IMO. Of course you seem to have money > so you may just want to go for it. Hell, why not wait for the Tyan > dual-Athlon mobo then stick two 1533MHz Palomino Athlons in it and 4 256MB > modules. Do a make -j 10 world and don't tell me how fast it goes. Hmmm, that's a idea I guess.... What's currently the speed record for a make world build? > > True... I guess what Intel needs is to work on newer stuff in > >secret and then all of the sudden just release it rather than having the > >entire world expecting some announcement on so and so date.. heh. > > Then they'd have a great product that just the hardware enthusiasts would > know about. It's a marketing world and they have to hype products months > (indeed-Years) in advance to get the word out among the 95% of people that > don't know Megahertz from megabytes and think Windows is the only game in > town--but know that MSword 2010 won't take a week to load if they buy the > new whiz-bang Pentium-5. Yeah... I guess marketing and the software and hardware vendors helping each other to make each others products obselete. > Weird name, BTW. Pentium basically means "five-ium". We are now on the > Fiveium-four. Can't wait for the Fiveium-five. or the Fiveium-six! > > Maybe the Celeron 2 is really just P3's that don't past certain > >tests and instead of putting it in the trash or as problem chips, the > >marketing department thinks of selling it as a lower end CPU. > > Remember 486SX's? Intel originally destroyed all 486s the had bad FPUs. They > later decided to go ahead and try to market them and, to their surprise, > they sold very well. It was only many months later that they actually began > making 486s that had no FPUs. In fact, some SX's had WORKING FPU's that had > just been disabled to keep up with demand. That said, the Celeron 2's may be > dug out of Intel's stock of P3's that had some (not all) bad cache, saved > because of the 486SX lesson they learned. Yep, I guess they are taking advantage of all the sand used ;) > Too bad they use another 486 idea today--having only 8K of L1 cache memory > on the Pentium IV. Yeah... 8k is really bad. I guess with all these cost cutting, companies are just releasing things with less things in them. > If the trend of lowering performance continues, we might see marketing like: > > "The new Pentium 'VI'--runs the 'VI' editor without long processing delays!" > (...Then AMD market's the new AMD EMACS processor) > > Hmm, ya know I haven't written one word about FreeBSD in the last 10 or so > messages. This is getting really bad. Sorry for spamming everyone! Atleast vi is from Berkeley! Cheers, Vince - vince@WURLDLINK.NET - Vice President ________ __ ____ Unix Networking Operations - FreeBSD-Real Unix for Free / / / / | / |[__ ] WurldLink Corporation / / / / | / | __] ] San Francisco - Honolulu - Hong Kong / / / / / |/ / | __] ] HongKong Stars/Gravis UltraSound Mailing Lists Admin /_/_/_/_/|___/|_|[____] Almighty1@IRC - oahu.DAL.NET Hawaii's DALnet IRC Network Server Admin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message