Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10 beta 3/9/83; site xios.UUCP Path: utzoo!dciem!nrcaer!xios!tony From: tony@xios.UUCP (Keeper Of News) Newsgroups: net.micro.att Subject: Re: Orphaned Response Message-ID: <258@xios.UUCP> Date: Fri, 2-May-86 11:24:24 EDT Article-I.D.: xios.258 Posted: Fri May 2 11:24:24 1986 Date-Received: Tue, 6-May-86 08:39:12 EDT References: <147@sdsioa.UUCP> <12700002@gorgo.UUCP> Reply-To: tony@xios.UUCP (Keeper Of News) Organization: Xios Systems Corp., Ottawa, Ont. Lines: 68 Summary: I tried mailing this, but one of the machines (sdcc3) on my way to him doesn't seem to be talking. There was another posting which said that it was a reasonable choice for your application. Maybe I'm spoiled, but I find it rather slow. Anyway here is the text of may mail message. ========================================================================== Actually, I find it mighty slow with just one user. Looking at the results of a certain benchmark set ( the results of which can't be published due to our licence ), it is better than a 3b2/300, about equivalent to a Spectrix Series 30 ( 8MHz 68000 on a multibus ) or a 68010 NCR Tower. We tested a MicroVax II, and it is about twice as fast, and a Convergent Technologies Mightyframe which is about 3 - 10 times faster. All these machines are running System V, except the Spectrix, which runs V7, and the MicroVax, which runs Ultrix 1.1 (or is it 1.2). Now all these tests were done in single user mode, so besides the usual caveates about benchmarks, they don't say much about how they react when loaded with users. However... The 3B2 machines do a wee bit of processing on the ports board, which means that under a high i/o load it would suffer less than the MicroVax (depending on which boards you are using) or the Tower, though not much. Now the Tower can do DMA from the frontend straight into user memory, and they have a 68020 model, so that may be much better than indicated, especially for the kind of application you are interested in. Now the Mightyframe has an i/o processor board ( an added option ) which off-loads most of the work from the main CPU (68020). So for your application you are unlikely to see much degradation from the above results. Insofar as using them for normal things, with 1-5 users, I have the following subjective results from doing things like mail, vi, etc.: 3B2 (both 300 and 400 ) - pigs Spectrix, MicroVax - reasonable with one or two people, slowing down mightily with more. Mightyframe - blindingly fast. Most commands start with no visible delay, piggish programs like rn, vi, SuperRogue, less than a second. Rnews doesn't even make a dent. So, in summation, the 3b2 starts out as a pig, and will get worse as you add users to it (when I mentioned 20 users on a 3B, some one asked if it was a joke). The Mightyframe flys like the wind, and I've seen no visible degradation with up to five people and an rnews going at the same time. (And according to the sales pitch, this should be true for 20). Though I'm not in the habbit of hawking other people's equipment, I'd suggest you look at the Mightyframe. For about the same price as the 3b2, it's a lot more bang for the buck. (It also has TCP/IP a la sockets and other Berkley goodies). Though I don't have a phone number handy, they are located in San Jose, Ca. #include -- ------------------------------------------------------------------------------- Tony Lill Keeper of News @ Xios Systems Corporation 1600 Carling Avenue, Suite 150, Ottawa, Ontario, Canada, K1Z 8R8 (613) 725-5411 utzoo!dciem!nrcaer!xios!tony ------------------------------------------------------------------------------- Not the edge of the world, but we can see it from here.