From owner-freebsd-hardware Sun Feb 18 11:28:57 1996 Return-Path: owner-hardware Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id LAA29289 for hardware-outgoing; Sun, 18 Feb 1996 11:28:57 -0800 (PST) Received: from Sysiphos (Sysiphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id LAA29284 for ; Sun, 18 Feb 1996 11:28:54 -0800 (PST) Received: by Sysiphos id AA04255 (5.67b/IDA-1.5 for freebsd-hardware@FreeBSD.org); Sun, 18 Feb 1996 20:25:17 +0100 Message-Id: <199602181925.AA04255@Sysiphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Sun, 18 Feb 1996 20:25:16 +0100 In-Reply-To: Ollivier Robert "Re: P55TP4XE issues, SP3G q's" (Feb 17, 19:13) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: Ollivier Robert Subject: Re: P55TP4XE issues, SP3G q's Cc: skynyrd@tahoma.cwu.edu, freebsd-hardware@freebsd.org Sender: owner-hardware@freebsd.org Precedence: bulk On Feb 17, 19:13, Ollivier Robert wrote: } Subject: Re: P55TP4XE issues, SP3G q's } It seems that Eric Chet said: } > slower than a pentium 133 on a Asus P55TP4 on a make world. The } > Pentium takes 3:40 the AMD-133 takes about 4:00, times are in } > HH:MM and are approximate. I have only timed the SP3G. } } I'm surprised. } } My DX4/100, 32 MB, BT-747S + AHA-1740A EISA system takes only 4:50 to } complete a "make world" (with /usr/src and /usr/obj mounted async). Well, what did you expect ? :) c't 3/96 mkwldstone :) source --------------------------------------------------- iDX2/66 1.00 (6:40 => 1.00) (se) aDX4/100 1.26 (5:15) iDX4/100 1.36 (4:50 => 1.38) (roberto) a5x86/133-WT 1.56 (4:20 => 1.54) (se) a5x86/133-WB 1.63 (4:00 => 1.67) (ec0) P75-PB-WB 1.69 (3:55) (ec0=Eric Chet, roberto=Ollivier Robert, se=Stefan Esser) The "c't 3/96" column lists some (of 18) results for i486 compatible CPUs. The Cyrix 5x86 results depend too much on the motherboard and chip revision to be useful as a system performance indicator (the 100MHz version seems to lie in the middle between iDX4/100 and AMD 5x86/133 at about 1.45 to 1.49). The c't numbers were obtained using the Babco test suite based on Windows programs. But since it nearly perfectly matches make world times, I think those values can be trusted anyway ... :) I've just found an offer for a Dirty Tag RAM at 13.50DM (about $8.00 taking local VAT into account) and think this is a good price for speeding up the make world by another 20 minutes ... BTW: The P75 with ppipelined burst cache is about as fast as a P90 with Neptun, SiS or Triton with asynch. cache SRAMS. Seems the AMD 5x86 really is a good buy at $80 (135,00DM). Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se