From owner-freebsd-hackers Thu May 16 12:00:57 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id MAA24959 for hackers-outgoing; Thu, 16 May 1996 12:00:57 -0700 (PDT) Received: from haldjas.folklore.ee (Haldjas.folklore.ee [193.40.6.121]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id MAA24952 for ; Thu, 16 May 1996 12:00:49 -0700 (PDT) Received: (from narvi@localhost) by haldjas.folklore.ee (8.6.12/8.6.12) id WAA28287; Thu, 16 May 1996 22:05:44 +0300 Date: Thu, 16 May 1996 22:05:43 +0300 (EET DST) From: Narvi To: Terry Lambert cc: hackers@FreeBSD.org Subject: Re: EDO & Memory latency In-Reply-To: <199605161827.LAA17550@phaeton.artisoft.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-hackers@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk On Thu, 16 May 1996, Terry Lambert wrote: > > > > I have just tried lmbench and the numbers it gives are looking > > > > slightly strange for me. It shows memory latency upto 500ns while > > > > I have 60-ns EDO memory in a Pentium/75 box. Okay, its external > > > > clock is 25MHz, this gives 40ns, one wait state, it gives another 40ns, > > > > it gives 80ns, but why the overhead is over 400ns ? > > > > > > 25 * 3 = 75 > > > > > > Multiply access latency by 3. > > > > Contrary to what some prefer to think, Pentium 75 run on the external > > speed of 50 Mhz - 1.5*50=75. So how about looking the figures over once > > more time? > > Was this supposed to be a memory bus or a data bus speed? I guess it was the memory speed and latency - I doubt he had his memory on the other side of the PCI bus. In the opposite case no-one would say that 60ns EDO memory is recomemded for 66Mhz external bus as the memory access would in any case be done at 33Mhz. > > PCI can't run faster than 33 -- on a 75, that means 25. > > I feel I can't argue against this - though why aren't they running it at some other speed than 1/2 of external and not 33/50 of it? I hate Pentium 75-s which lose to 486DX-80s in one of the components of graphics winmark... (everything else was exactly the same) :-( > Terry Lambert > terry@lambert.org > --- > Any opinions in this posting are my own and not those of my present > or previous employers. > Sander .sigless on the moment