jaw@ames.arpa (James A. Woods) (07/16/87)
During Phoenix USENIX, at the Unisoft booth, on the Mac II,
a "time cat /usr/dict/words > /dev/null" took about four seconds
of clock time. Staffers knew little about the disk or controller,
except that it was made by Qualogy.
Is the sluggishness due to:
(a) xfer rate of the disk (the Dataframe, for example is
spec'ed at just 5 Mbits/sec)?
(b) an underpowered controller?
(c) the Sys5-derived filesystem (BSD would do better)
(d) suboptimal interleave factors for above combination?
The informal comparison between the new Sony 68020-based machine
and the Mac II in the August Unix/World also hints at unbalanced I/O
on the Mac II.
Would disk performance under the default Mac OS be similar?
ames!jawjww@sdcsvax.UCSD.EDU (Joel West) (07/16/87)
A/UX is a System V kernel with BSD extensions. An
earlier posting said it has the cursed 14-character file name
limit, so that's an indication of how ATT-like it is.
I was under the impression that the only disk now supported by A/UX
is a Quantum, provided in the Apple 80 and Jasmine 80.
--
Joel West, Palomar Software, Inc. (c/o UCSD)
{ucbvax,ihnp4}!sdcsvax!jww or jww@sdcsvax.ucsd.eduelwell%tut.cis.ohio-state.edu@osu-eddie.UUCP (Clayton Elwell) (07/16/87)
In article <3465@sdcsvax.UCSD.EDU> jww@sdcsvax.UCSD.EDU (Joel West) writes: > >I was under the impression that the only disk now supported by A/UX >is a Quantum, provided in the Apple 80 and Jasmine 80. >-- > Joel West, Palomar Software, Inc. (c/o UCSD) > {ucbvax,ihnp4}!sdcsvax!jww or jww@sdcsvax.ucsd.edu Not so--the CDC Wren works quite nicely, even though it doesn't fit inside the case... -=- Clayton Elwell Arpa/CSNet: Elwell@Ohio-State.ARPA UUCP: ...!cbosgd!osu-eddie!elwell Voice: (614) 292-6546
gustav@swanee.OZ (Gustav) (07/22/87)
In article <2363@ames.arpa>, jaw@ames.arpa (James A. Woods) writes: > During Phoenix USENIX, at the Unisoft booth, on the Mac II, > a "time cat /usr/dict/words > /dev/null" took about four seconds > of clock time. > ... > Is the sluggishness due to... Well, let me compare this to our VAX 11/750 running 4.2BSD. The disk we use is RA81: Script started on Wed Jul 22 22:24:23 1987 103 $ time cat /usr/dict/words > /dev/null real 0m31.53s user 0m4.61s sys 0m2.26s 104 $ ^D script done on Wed Jul 22 22:25:01 1987 The system at the moment has 5 users and the load is 4.2. Mac II wins a long way!
cbenda@unccvax.UUCP (carl m benda) (07/24/87)
In article <344@swanee.OZ>, gustav@swanee.OZ (Gustav) writes: > In article <2363@ames.arpa>, jaw@ames.arpa (James A. Woods) writes: > > During Phoenix USENIX, at the Unisoft booth, on the Mac II, > > a "time cat /usr/dict/words > /dev/null" took about four seconds > > of clock time. > Well, let me compare this to our VAX 11/750 running 4.2BSD. The disk > we use is RA81: > > Script started on Wed Jul 22 22:24:23 1987 > 103 $ time cat /usr/dict/words > /dev/null > > real 0m31.53s > user 0m4.61s > sys 0m2.26s > 104 $ ^D > > script done on Wed Jul 22 22:25:01 1987 > > The system at the moment has 5 users and the load is 4.2. > Mac II wins a long way! Please multiply Mac II time by 5 then add the operating system overhead time necessary for time slicing to simulate the time necessary for five users... I find your comparisons of multi-user operating systems with non-multi user systems ill consieved. If your going to compare one machine with another, it should be done under similar conditions. There were five users logged on and God knows how many system jobs being performed. The question remains: Why is AUX disk I/O ONLY 50KB/second? /Carl ...mcnc!unccvax!cbenda
tim@ism780c.UUCP (Tim Smith) (07/24/87)
In article <777@unccvax.UUCP> cbenda@unccvax.UUCP (carl m benda) writes: < In article <344@swanee.OZ>, gustav@swanee.OZ (Gustav) writes: < > In article <2363@ames.arpa>, jaw@ames.arpa (James A. Woods) writes: < > > During Phoenix USENIX, at the Unisoft booth, on the Mac II, < > > a "time cat /usr/dict/words > /dev/null" took about four seconds < > > of clock time. < > Well, let me compare this to our VAX 11/750 running 4.2BSD. The disk < > we use is RA81: < > real 0m31.53s < > user 0m4.61s < > sys 0m2.26s < > The system at the moment has 5 users and the load is 4.2. < > Mac II wins a long way! < < Please multiply Mac II time by 5 then add the operating system overhead time < necessary for time slicing to simulate the time necessary for < five users... Nope, you don't have to do that. The significant time on the multi-user system is user + system time. That is relatively independant of load. The load affects the real time. Thus, the above vax would not get much better than 4.6 + 2.3 seconds even if it were single user. I don't know why that vax is so slow. Here is what I get on a 780 running 4.3: Script started on Fri Jul 24 13:34:12 1987 388. wc /usr/dict/words 24259 24259 198596 /usr/dict/words 389. time cat /usr/dict/words > /dev/null real 0m1.00s user 0m0.00s sys 0m0.40s 390. r time cat /usr/dict/words > /dev/null real 0m1.38s user 0m0.01s sys 0m0.30s 391. r time cat /usr/dict/words > /dev/null real 0m0.96s user 0m0.01s sys 0m0.26s 392. exit script done on Fri Jul 24 13:34:45 1987 -- Tim Smith, Knowledgian {sdcrdcf,seismo}!ism780c!tim
woody@tybalt.caltech.edu (William Edward Woody) (07/25/87)
In article <777@unccvax.UUCP> cbenda@unccvax.UUCP (carl m benda) writes: ... Deleted material comparing Mac II with VAX 11/750 with 5 people on it >I find your comparisons of multi-user operating systems with non-multi >user systems ill consieved. If your going to compare one machine with >another, it should be done under similar conditions. There were five >users logged on and God knows how many system jobs being performed. Uh, shouldn't you really be addressing computing power per dollar? I mean, sure, you can blow a Mac II out of the water with an unloaded VAX 11/750, but is a VAX 11/750 around here ever unloaded? And for 1 VAX 11/750, how many Mac II's can you buy? No flames as to "That's not how we compare computers" because That's How It Is Done In The Real World. (Except for government agencies who can buy millions of dollars of computers with pocket change... ;-) - William Woody Mac! > ][n && /|\ woody@tybalt.caltech.edu woody@juliet.caltech.edu
maiden@beowulf.ucsd.edu (VLSI Layout Project) (07/26/87)
>W.E. Woody writes: > ...sure with you can blow a Mac II out of the water with an unloaded > VAX 11/750... Is that really true? I would think that even an unloaded VAX 11-750 might have difficulty "blowing away" a Mac II. I do not have a Mac II running A/UX, but a 80386 machine is perceptably faster than a VAX 750 (a 780 might be a different matter). Not to bring up the old benchmark contraversy or anything, but not all VAX 11/750 's are alike (FPU, # DMA channels, etc.). The point, however, is a good one. Never have I seen an unloaded VAX. An unloaded SUN, perhaps, but not a VAX. Let's keep workstations as workstations, and minis as minis. E. Jung seismo!sdcsvax!maiden or maiden@sdcsvax.ucsd.edu
dennisg@pwcs.StPaul.GOV (Dennis Grittner) (07/28/87)
In article <344@swanee.OZ> gustav@swanee.OZ (Gustav) writes: >In article <2363@ames.arpa>, jaw@ames.arpa (James A. Woods) writes: >> During Phoenix USENIX, at the Unisoft booth, on the Mac II, >> a "time cat /usr/dict/words > /dev/null" took about four seconds >> of clock time. > >Well, let me compare this to our VAX 11/750 running 4.2BSD. The disk >we use is RA81: Lots of stuff removed about times, etc. > >The system at the moment has 5 users and the load is 4.2. >Mac II wins a long way! What you are witnessing on the 11/750 is called DEC never could do I/O. I can't recall his name , but at a Uniforum in D.C. the best comment I heard ( at a networking type forum ) was the famous " my dog could do a better job of designing I/O than DEC did on the Vax ". I haven't benched the latest Vaxen but I sure did the older ones and I bought a Pyramid just to get a Unix box that could do I/O. That's not a BSD I/O problem it's a Vax I/O limitation. -- Dennis Grittner City of Saint Paul, Minnesota (612) 298-4402 Room 700, 25 W. 4th St. 55102
dennisg@pwcs.StPaul.GOV (Dennis Grittner) (07/28/87)
In article <3327@cit-vax.Caltech.Edu> woody@tybalt.caltech.edu.UUCP (William Edward Woody) writes: >In article <777@unccvax.UUCP> cbenda@unccvax.UUCP (carl m benda) writes: >... Deleted material comparing Mac II with VAX 11/750 with 5 people on it > Lots of stuff removed >Uh, shouldn't you really be addressing computing power per dollar? I mean, >sure, you can blow a Mac II out of the water with an unloaded VAX 11/750, >but is a VAX 11/750 around here ever unloaded? And for 1 VAX 11/750, how More stuff removed - A Vax 11/750 can't really touch a Mac II in almost any area of performance. Certainly not simple computer speed, not MFLOPS, and probably not I/O under many conditions. The 750 ( or 780 ) is/was a dog ( performance wise ). -- Dennis Grittner City of Saint Paul, Minnesota (612) 298-4402 Room 700, 25 W. 4th St. 55102
rbl@nitrex.UUCP ( Dr. Robin Lake ) (07/30/87)
In article <3241@pwcs.StPaul.GOV> dennisg@pwcs.StPaul.GOV (Dennis Grittner) writes: >In article <344@swanee.OZ> gustav@swanee.OZ (Gustav) writes: >>In article <2363@ames.arpa>, jaw@ames.arpa (James A. Woods) writes: >>> During Phoenix USENIX, at the Unisoft booth, on the Mac II, >>> a "time cat /usr/dict/words > /dev/null" took about four seconds >>> of clock time. >> >>Well, let me compare this to our VAX 11/750 running 4.2BSD. The disk >>we use is RA81: > Lots of stuff removed about times, etc. > > >> >What you are witnessing on the 11/750 is called DEC never could >do I/O. ... In spite of the I/O rate deficiencies of the Unibus, consider the overhead of the device drivers and the entire I/O strategy of the operating system. In the mid-1970's, we did UNIX performance evaluation on a Unibus machine (PDP-11/45) using a solid-state disk versus a head-per-track fixed head disk and a moving head disk. Even though the transfer rate of the solid-state disk is 17,000 times faster (zero latency) than the fixed head disk, the device drivers consumed the vast majority of the transfer time. With DMA transfers, only a few words need be transfered to the device controller to initiate a transfer. However, many thousands of instructions need be executed before those few words are determined! Rob Lake decvax!cwruecmp!nitrex!rbl ihnp4!cbosgd!cwruecmp!nitrex!rbl Disclaimer: This work was done prior to joining my present employer.