From owner-freebsd-hardware Wed Oct 2 11:55:24 1996 Return-Path: owner-hardware Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id LAA27710 for hardware-outgoing; Wed, 2 Oct 1996 11:55:24 -0700 (PDT) Received: from lserver.infoworld.com (lserver.infoworld.com [192.216.48.4]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id LAA27704 for ; Wed, 2 Oct 1996 11:55:22 -0700 (PDT) From: BRETT_GLASS@infoworld.com Received: from ccgate.infoworld.com (ccgate.infoworld.com [192.216.49.101]) by lserver.infoworld.com (8.7.5/8.7.3/GNAC-GW-1.2) with SMTP id LAA17747; Wed, 2 Oct 1996 11:52:02 -0700 (PDT) Received: from ccMail by ccgate.infoworld.com (SMTPLINK V2.11) id AA844282062; Wed, 02 Oct 96 12:09:47 PST Date: Wed, 02 Oct 96 12:09:47 PST Message-Id: <9609028442.AA844282062@ccgate.infoworld.com> To: "Michael L. VanLoon -- HeadCandy.com" Cc: rgrimes@gndrsh.aac.dev.com, narvi@haldjas.folklore.ee, obrien@nuxi.cs.ucdavis.edu, freebsd-hardware@freebsd.org Subject: Re: H/W recommendation Sender: owner-hardware@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >> But the bargain >> basement version of the Pentium Pro, with the 256 KB cache, will drag in >> the same configuration. Unfortunately, far too many clone vendors just >> HAPPEN not to mention in their ads that they're including the cheaper >> CPU. > Do you have any proof of this? I think you're speculating. I've done benchmarks of heavily-loaded servers with and without the proper chip. Intel originally wanted *all* Pentium Pros to have half a meg of cache on board (see their earliest announcements), and for good reason: it's needed. Otherwise, the ratio of internal to external bus speeds is a killer. > And, I don't know how you can call the 256K P6 the "bargain basement" > when it costs as much as the *highest* speed Pentium. Because it's under-engineered to pinch pennies. Again, Intel originally specified the parts to ALL have 512K of cache. > ... But if you like paying $1200 for your processors, you can think of > them any way you like. :-) Usually, a high price for a chip means either great demand or yield problems. I can't speculate on Intel's rationale, but in this case, both factors could be important. > Besides, as already pointed out above, the 256K Pentium Pro cache is > already *way* faster than any size cache on a Pentium. Plus, with a 256K > cache you're already taking about hits in the 90+% range. On a personal machine, maybe. But not on a UNIX server with a substantial working set. Or on a multiprocessing machine. > 512K would be nice, but its definitely at the far side of diminishing > returns. Not when the penalty for using the external bus is so high. When misses carry a high cost, a higher hit rate is required. --Brett