ericm@ibmpa (02/16/90)
ok, here's the offical product announcement...
this'll give y'all more things to speculate on.
---------------cut here---------------
HIGH-PERFORMANCE RISC WORKSTATION FAMILY ANNOUNCED
IBM today announced the RISC System/6000* family -- a series of nine
high-performance workstations and servers that offer customers on the
leading edge of technology the power and solutions they are seeking
for computing's most advanced, sophisticated challenges.
Based on IBM's new POWER (Performance Optimization With Enhanced
RISC) Architecture, the family includes the industry's highest desktop
workstation performance at more than 27 million instructions per second
(MIPS) and 7 million floating-point operations per second (MFLOPS).
With complete system prices starting at $12,995, it also offers
industry-leading price/performance.
The nine POWERstations and POWERservers -- with top performance of
41 MIPS and 13 MFLOPS -- are designed to work in heterogeneous, open
systems networks with workstations from IBM and other manufacturers.
They will be supported by hundreds of third-party applications and
backed by IBM's service and support network.
This performance is supported by an aggressive software developer
program to aid software developers in porting their products to the
new family. To date, this program has resulted in commitments to port
more than 600 applications worldwide in areas ranging from mechanical
and electrical design to structural mechanics and computer-aided
software engineering.
"The RISC System/6000 family brings unprecedented power to the desks
of scientists, engineers, designers and other professionals," said
George H. Conrades, IBM senior vice president and general manager, US
Marketing & Services. "It enables our customers to do more, whether
their POWERstations are operating on their own, interconnected with a
mainframe or sharing the resources of a distributed computing
environment. Today, we are joined by hundreds of IBM Business Partners
and software developers in offering our customers the industry's most
advanced workstation solution."
The four-member POWERstation family and the five POWERservers
feature a high-performance implementation of IBM's Micro Channel bus
architecture, and are designed to exploit a new version of AIX --
IBM's implementation of the UNIX operating system -- also announced
today. The family is complemented by the new Xstation 120, a very
low-priced X server terminal, giving users concurrent access to a
variety of applications.
The RISC System/6000 family achieves its premium performance with
three major technical advances: a new superscalar processor capable
of executing multiple instructions in a single cycle; the industry's
most advanced RISC (Reduced Instruction Set Computer) floating-point
processor for numeric-intensive applications, such as quantitative
analysis, and optimized 3-D graphics capabilities for such complex
applications as visualization and mapping.
"The RISC System/6000 family is the latest example of our continuing
drive to bring the benefits of advancing technology to our customers,"
said IBM President Jack D. Kuehler. "Today's announcement
demonstrates our continuing commitment to open systems and industry
standards, and our determination to become a leader in the workstation
and open systems environment."
RISC System/6000 Family
The family begins with the RISC System/6000 POWERstation 320, the most
powerful desktop workstation available. It has a rated performance of
27.5 MIPS and 7.4 MFLOPS, with a complete system starting at $12,995.
The POWER Architecture and advanced IBM CMOS microprocessor design
combine to achieve this exceptional level of performance.
The workstation family also includes the deskside POWERstations 520
and 530, offering greater memory and disk storage capacity. These
systems provide performance of up to 34.5 MIPS and 10.9 MFLOPS.
The POWERstation 730, with an integrated Supergraphics Processor
Subsystem, is a deskside unit that combines extensive high-function,
3-D graphics with a high-performance superscalar RISC workstation.
Its capabilities include a shading processor for solid modeling and
other 3-D engineering designs.
The RISC System/6000 family is especially rich in graphics capability.
Four new graphics adapters range from grayscale to high-function 3-D,
allowing customers to implement solutions ranging from desktop
publishing to mechanical design to 3-D animated visualization of
scientific phenomena.
The RISC System/6000 family also includes five high-performance
servers with prices starting at $20,375. The POWERservers 320, 520,
530, 540 and 930 can be configured either as LAN-attached servers
for multiple users -- compute server or file server -- or as
multi-user systems using ASCII terminals or Xstation 120s. These
systems offer performance of up to 41 MIPS and 13 MFLOPS. The MFLOPS
rating is several times faster than other servers in this price class.
New AIX Version
AIX Version 3 for RISC System/6000 is an outstanding UNIX operating
environment and software development platform. While conforming to
open software standards, it includes many new and improved features,
such as physical disk space management, advanced file system and
program management facilities, extended realtime support and enhanced
virtual memory.
The new UNIX implementation will provide a windowed hypertext
retrieval system that offers a versatile and innovative approach to
providing customers with access to online publications and help
information. Documentation is available on a 5.25-inch compact disc
(CD-ROM) that may be accessed from the user's POWERstation or from a
network POWERserver.
User Interfaces
IBM announced industry-leading graphical user interfaces: AIXwindows
Environment/6000, based on the OSF/Motif graphical user interface
from the Open Software Foundation; and AIX NextStep Environment/6000,
based on the innovative NextStep environment from NeXT Inc.
AIXwindows Environment/6000 and AIX NextStep Environment/6000 will be
available as separate licensed programs. Customers can choose the
interface most applicable to their environments.
Service and Support
Each POWERstation and POWERserver will be backed by IBM service and
support. IBM announced a one-year warranty that includes 24-hour,
seven-day-a-week availability of hardware service. Full software
service is included in the license charge.
In addition, IBM has significantly increased its number of skilled
systems engineers to assist customers in the design, configuration,
installation and support of the RISC System/6000 family.
IBM also announced SystemXtra for the RISC System/6000 family, a fee
service offering that provides a total service solution after system
installation. It includes a 24-hour, seven-day-a-week telephone hot
line for hardware and software problems and usage assistance, in
addition to onsite installation of software maintenance.
A number of IBM manufacturing and development sites were involved in
today's announcement, including: Austin, Texas; Basingstoke, United
Kingdom; Boca Raton, Florida; Bordeaux, France; Burlington, Vermont;
Charlotte, North Carolina; East Fishkill, New York; Endicott, New York;
Essonnes, France; Fujisawa, Japan; Haifa, Israel; Havant, United
Kingdom; Hursley, United Kingdom; Kingston, New York; Lexington,
Kentucky; Manassas, Virginia; North Harbour, United Kingdom;
Poughkeepsie, New York; Raleigh, North Carolina; Rochester, Minnesota;
San Jose, California; Santa Palomba, Italy; Sindelfingen, West Germany;
Toronto, Canada; Vimercate, Italy; Yamato, Japan and Yorktown, New York.
*Trademarks: RISC System/6000 is a trademark of International
Business Machines Corporation.
---------------cut here---------------
eric murray ibmsupt!ericm@uunet.uu.net {ucbvax,uunet}!ibmsupt!ericm
KILL YOUR TELEVISION!gnu@hoptoad.uucp (John Gilmore) (02/19/90)
> . . . POWERstations and POWERservers . . . I like the way they adopted the Sun Marketing naming convention (SPARCstations and SPARCservers). > . . . a complete system starting at $12,995. Including OS and window system? Manuals? Ethernet? Compilers? In the original RT/PC announcement a few years ago, the lowball price didn't even include a keyboard or monitor, which cost $thousands! One analysis showed that if Sun had charged as much for Unix as IBM, they could've thrown in the workstation for free! Also, one breakdown I saw was that this system only has four slots -- and all of them are in use. No expandability. > Documentation is available on a 5.25-inch compact disc > (CD-ROM) that may be accessed from the user's POWERstation or from a > network POWERserver. At extra cost > IBM announced industry-leading graphical user interfaces . . . > available as separate licensed programs. Customers can choose the > interface most applicable to their environments. At extra cost > IBM announced a one-year warranty that includes 24-hour, > seven-day-a-week availability of hardware service. Full software > service is included in the license charge. License charge? For a warranty? Oh, I get it! If you buy the Unix from them, you are forced to buy full software support for it! Well, no problem, I'm sure there are lots of third party vendors selling Unix for the RIOS... > In addition, IBM has significantly increased its number of skilled > systems engineers to assist customers in the design, configuration, > installation and support of the RISC System/6000 family. I heard from a usually reliable source that there are TWO people in the Bay Area who are trained to configure and repair these machines. That's a "significant increase" from last year, when there were NONE! > IBM also announced SystemXtra for the RISC System/6000 family, a fee > service . . . . . . . . . . . . total service solution . . . At extra cost. -- John Gilmore {sun,pacbell,uunet,pyramid}!hoptoad!gnu gnu@toad.com Just say *yes* to drugs. If someone offers you a drug war, just say no.
adam@ncifcrf.gov (Adam W. Feigin) (02/19/90)
In article <10307@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >> . . . POWERstations and POWERservers . . . > >I like the way they adopted the Sun Marketing naming convention >(SPARCstations and SPARCservers). Well, since they licensed the graphics technology from SGI, and they call their machines the "POWER Series" and IBM is going after Sun and their SPARCstations/servers, its natural that IBM should call their machines in this manner. > >> . . . a complete system starting at $12,995. > >Including OS and window system? Manuals? Ethernet? Compilers? In >the original RT/PC announcement a few years ago, the lowball price >didn't even include a keyboard or monitor, which cost $thousands! One >analysis showed that if Sun had charged as much for Unix as IBM, they >could've thrown in the workstation for free! Correct. The base price does NOT include OS & Window system or ethernet (I assume that you get manuals & compilers with the OS, but I could be wrong, and I certainly wouldn't put it past IBM NOT to include them, and charge extra for them, as they have a nasty habit of doing). I did notice that in the glossy that the 3-button mouse and keyboard are marked as "optional products".. Nice, very nice, looks like IBM is up to their old tricks again. They never learn, do they ?? (BTW the price sheet I have show the OS + X-Window System at $2000) >Also, one breakdown I saw was that this system only has four slots -- >and all of them are in use. No expandability. I'm not sure about this, but the base system comes with a 120MB DBA ("Direct Bus Attached" -- whatever that means) disk; I dont know if you need a slot for a controller, but if you want to add more disk, you gotta buy a controller. Lets not forget a slot for ethernet (optional), a slot for graphics (you really dont want to see anything, do you ??, you can just watch the blinkin' lights...) >> Documentation is available on a 5.25-inch compact disc >> (CD-ROM) that may be accessed from the user's POWERstation or from a >> network POWERserver. > >At extra cost Of course !! It will be a cold day in hell when the temperature is below freezing when IBM gets serious about the Unix/Workstation market. -- Internet: adam@ncifcrf.gov Adam W. Feigin UUCP: {backbonz}!ncifcrf!adam Senior Systems Manager Mail: P.O. Box B, Bldg 430 National Cancer Institute-Supercomputer Center Frederick, MD 21701 Frederick Cancer Research Facility
mccarty@hpuxa.ircc.ohio-state.edu (James E. McCarty) (02/20/90)
There seems to be some confusion about what is and isn't included
in the basic POWERstation 320 configuration. This information is
from the "RISC System/6000 Quick Pricer" (IBM Manual G320-9881).
"POWERstation 320 -- Disktop Workstations
The following configurations include, unless otherwise noted, a
1280 x 1024 graphics display, Ethernet, keyboard, 3-button mouse,
AIX Version 3 for RISC System/6000 operating system (1-2 user)
and graphical user interface, software service and one-year
hardware warranty service.
Description Price
Monochrome Workstations
1. POWERstation 320, 19" Mono Display, Grayscale $12,995
Graphics Display Adapter, 8MB RAM, 120MB Disk
Color Workstations
2. POWERstation 320, 16" Color Display, Color $15,030
Graphics Adapter, 8MB RAM, 120MB Disk
3. POWERstation 320, 19" Color Display, Color $15,330
Graphics Adapter, 8MB RAM, 120MB Disk"
Also, the information I received indicated that the 4 slots in
the base model were all available.
---------------------------------------------------------------------
Jim McCarty This are my opinions only and do not reflect
Leader VM Systems the opinions of my employer or IBM
IRCC, OSU
----------------------------------------------------------------------schafer@brazos.rice.edu (Richard A. Schafer) (02/20/90)
In article <10307@hoptoad.uucp>, gnu@hoptoad.uucp (John Gilmore) writes: > > . . . POWERstations and POWERservers . . . > > I like the way they adopted the Sun Marketing naming convention > (SPARCstations and SPARCservers). > > > . . . a complete system starting at $12,995. > > Including OS and window system? Manuals? Ethernet? Compilers? In > the original RT/PC announcement a few years ago, the lowball price > didn't even include a keyboard or monitor, which cost $thousands! One > analysis showed that if Sun had charged as much for Unix as IBM, they > could've thrown in the workstation for free! I just signed on IBMLINK and ran a configuration on the standard grayscale configuration. I don't come up with the $12,995 price, but did produce the following configuration: Powerstation 320 $7475 120MB disk $1950 Grayscale adapter $1395 keyboard $ 255 mouse $ 130 mono display $1295 ------ Hardware total, $12,500 AIX 3.0 $1250 AIX Xwindows $ 500 ------ Software total, $ 1,750 Package total, $14,250 > Also, one breakdown I saw was that this system only has four slots -- > and all of them are in use. No expandability. The system described above has 3 slots still available. > > Documentation is available on a 5.25-inch compact disc > > (CD-ROM) that may be accessed from the user's POWERstation or from a > > network POWERserver. > > At extra cost Or in the traditional hard-copy form for no additional cost. > > IBM announced industry-leading graphical user interfaces . . . > > available as separate licensed programs. Customers can choose the > > interface most applicable to their environments. > > At extra cost Yep, since they allow you to choose Ethernet or Token-Ring adapters, these are optional items. In addition, you can configure the machine as an asynch timesharing machine, which a host of asynch adapters. So I don't think it's unreasonable to do it this way. > > > IBM announced a one-year warranty that includes 24-hour, > > seven-day-a-week availability of hardware service. Full software > > service is included in the license charge. > > License charge? For a warranty? Oh, I get it! If you buy the Unix > from them, you are forced to buy full software support for it! Well, > no problem, I'm sure there are lots of third party vendors selling Unix > for the RIOS... Excuse me? You're objecting to having software support included as part of the $1,250 purchase price for AIX? Given what I remember about SUN pricing, software support included in a $1,250 purchase price for AIX doesn't sound too bad to me. > > > In addition, IBM has significantly increased its number of skilled > > systems engineers to assist customers in the design, configuration, > > installation and support of the RISC System/6000 family. > > I heard from a usually reliable source that there are TWO people > in the Bay Area who are trained to configure and repair these machines. > That's a "significant increase" from last year, when there were NONE! > > > IBM also announced SystemXtra for the RISC System/6000 family, a fee > > service . . . . . . . . . . . . total service solution . . . > > At extra cost. Yep, and considering what SystemXtra is, they'd be crazy to give it away. SystemXtra is essentially a "we do all the hardware and software maintenance for you, answer end-user questions, etc." arrangement. While I have no interest in it, if I wanted that kind of hand-holding, I'd sure be prepared to pay for it. > -- > John Gilmore {sun,pacbell,uunet,pyramid}!hoptoad!gnu gnu@toad.com > Just say *yes* to drugs. If someone offers you a drug war, just say no. I just ran a slightly different configuration, getting a non-standard package with an Ethernet adapter and the CD-ROM player thrown in. Here's what you get: Powerstation 320 $7475 120MB disk $1950 grayscale adapter $1395 scsi controller $1200 ethernet adapter $ 695 keyboard $ 255 mouse $ 130 CD-ROM drive $1695 mono display $1295 ----- Hardware total, $16,090 (this configuration leaves 1 slot available, having added the Ethernet card and SCSI card to the previous configuration.) AIX $1250 CD-ROM hypertext $ 115 AIXWindows $ 500 ----- Software total, $1,865 (Interesting note: the standard shipment comes with the software preloaded, and you do not get actual diskette copies unless you make them yourself. You can purchase the "backup copy option" for the software listed above for an additional $400, which gets you diskette copies sent with the shipment.) Package total, $17,955 I still haven't found what they're including in the $12,995 price, but presumably it's the hardware listed in the first configuration above, plus something that costs $495. :-) Richard
schafer@brazos.rice.edu (Richard A. Schafer) (02/20/90)
In article <1514@fcs280s.ncifcrf.gov>, adam@ncifcrf.gov (Adam W. Feigin) writes: > Correct. The base price does NOT include OS & Window system or > ethernet (I assume that you get manuals & compilers with the OS, but I > could be wrong, and I certainly wouldn't put it past IBM NOT to > include them, and charge extra for them, as they have a nasty habit of > doing). I did notice that in the glossy that the 3-button mouse and > keyboard are marked as "optional products".. Nice, very nice, looks You get a C compiler with AIX, and perhaps f77; if you want their better fortran compiler, you have to order it separate. Pascal, COBOL, and ADA compilers are also separately packaged. > like IBM is up to their old tricks again. They never learn, do they ?? I don't know about that. For the performance quoted, a reasonably configured system still looks like a pretty good deal compared to SUN or DEC prices. (I don't yet know what our educational discount will be.) > (BTW the price sheet I have show the OS + X-Window System at $2000) > > >Also, one breakdown I saw was that this system only has four slots -- > >and all of them are in use. No expandability. Yes, four slots. No, not all in use. For $12,500, you can get a system with 3 slots available that has a processor, keyboard, display, mouse, and 120MB external disk. Adding an Ethernet adapter takes one more, still leaving 2. > I'm not sure about this, but the base system comes with a 120MB > DBA ("Direct Bus Attached" -- whatever that means) disk; I dont know > if you need a slot for a controller, but if you want to add more disk, > you gotta buy a controller. Lets not forget a slot for ethernet > (optional), a slot for graphics (you really dont want to see anything, > do you ??, you can just watch the blinkin' lights...) Yes, the disk does seem to take a slot, as well as the ethernet adapter. If you want to add more disk, I'd suggest a SCSI controller, rather than adding a direct attach disk controller per disk. The graphics adapter does not appear to take a "slot", from running the configuration program. > >> Documentation is available on a 5.25-inch compact disc > >> (CD-ROM) that may be accessed from the user's POWERstation or from a > >> network POWERserver. > > > >At extra cost You'd expect it for free, maybe? Hardcopy docs in the traditional binder format are provided with the product. Only the hypertext software and CD-ROM player are at extra cost. The hypertext software does allow you to create your own hypertext files and store them either on disk or (if you're wealthy enough create a CD-ROM of your own and read them from there.) I saw an early version of this code demonstrated, and it looked pretty slick. > Of course !! > > It will be a cold day in hell when the temperature is below freezing > when IBM gets serious about the Unix/Workstation market. I sure don't understand why you think this is a non-serious attempt. Given the prices versus performance, the product looks pretty serious to me. Richard
dcm@toysrus.uucp (dcm) (02/20/90)
In article <10307@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: > >I like the way they adopted the Sun Marketing naming convention >(SPARCstations and SPARCservers). Who knows? Maybe someone at Sun heard about the new IBM machines and thought "what a great naming convention!" :-) >> . . . a complete system starting at $12,995. > >Including OS and window system? Manuals? Ethernet? Compilers? In >the original RT/PC announcement a few years ago, the lowball price >didn't even include a keyboard or monitor, which cost $thousands! One Have anyone seen a pricelist yet? Does anyone know what the minimal configuration includes? I don't... The nice thing about packaging is that it gives customers complete flexibility about what they get. The other option, which some computer manufacturers have taken advantage of, is to charge $X over the 'minimal configuration' and include the world, thereby forcing each and every customer to pay the $X regardless of what they want. Which way is better? Hell, I don't know. I personally like the packaging. As long as it's easy to get the packages I need and they're affordable, who needs *everything*? Not every customer... >analysis showed that if Sun had charged as much for Unix as IBM, they >could've thrown in the workstation for free! Well, as everyone has pointed out, the RT was probably a flop. IBM is at least attempting to learn from their mistakes and do it better this time. I can tell you they're serious about it. All we ever hear around here is "got to be better than the RT in every way". I'm almost getting sick and tired of hearing it. It *is* a major consideration. >Also, one breakdown I saw was that this system only has four slots -- >and all of them are in use. No expandability. What are the four slots being used for in the configuration you saw? >> Documentation is available on a 5.25-inch compact disc >> (CD-ROM) that may be accessed from the user's POWERstation or from a >> network POWERserver. > >At extra cost I'm pretty sure they'll have on-line man pages too. I don't think they'll force you to buy the CD-ROM to run 'man'. > >> IBM announced industry-leading graphical user interfaces . . . >> available as separate licensed programs. Customers can choose the >> interface most applicable to their environments. > >At extra cost Well, at least the customer can *choose* their interface. They're not forced to pay for one(s) they don't want. > >> IBM announced a one-year warranty that includes 24-hour, >> seven-day-a-week availability of hardware service. Full software >> service is included in the license charge. > >License charge? For a warranty? Oh, I get it! If you buy the Unix >from them, you are forced to buy full software support for it! The way I read this paragraph is "the AIXV3 license charge will include full software service". I'm not sure about the hardware service. I imagine they'll do what everyone does: offer a service contract. >I heard from a usually reliable source that there are TWO people >in the Bay Area who are trained to configure and repair these machines. >That's a "significant increase" from last year, when there were NONE! Then again, the RS/6000 isn't being released in volume yet. Give them a chance to get service people in place around the country. They're smart enough to know that service is very important. Possibly important enough to make or break the product. >> IBM also announced SystemXtra for the RISC System/6000 family, a fee >> service . . . . . . . . . . . . total service solution . . . > >At extra cost. Don't most computer manufacturers offer a service program at a cost? Or are they all free? >John Gilmore {sun,pacbell,uunet,pyramid}!hoptoad!gnu gnu@toad.com My point was not to stick up for IBM (or bash John). At this point, I have nothing to gain from the success/failure of the RS/6000. I think we should wait and see how it turns out. Maybe IBM did learn something from the RT experience. Maybe not. If it flops, then bash all you want. I'll join you! But at least give them a chance... Craig Miller p.s. obviously, I'm not a spokesman for IBM. Anything I said is IMHO. -------- Craig Miller contractor @ IBM Austin UUCP: ..!cs.utexas.edu!ibmaus!auschs!toysrus.austin.ibm.com!dcm "I don't believe in .signatures."
dcm@toysrus.uucp (dcm) (02/20/90)
In article <1514@fcs280s.ncifcrf.gov> adam@fcs260c.UUCP (Adam W. Feigin) writes: > >It will be a cold day in hell when the temperature is below freezing >when IBM gets serious about the Unix/Workstation market. Do you have any idea how many $$ they've spent on this one? How many people? How much time? Where you ever associated with this project? Anyone who was associated will tell you the same thing: they're *very* serious about this one. They really do want to do it right this time. IBM recognizes this will change their future, succeed or fail. Craig p.s. I'm not a IBM spokesman. I speak only for myself. -------- Craig Miller contractor @ IBM Austin UUCP: ..!cs.utexas.edu!ibmaus!auschs!toysrus.austin.ibm.com!dcm "I don't believe in .signatures."
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) (02/20/90)
From the marketing announcement that I have, it looks like a base workstation package will run $14,250+$695=$14945 for the base desktop model with (8MB RAM, 120MB Hard Disk, Keyboard, Mouse, AIX3, AIXwindows, NFS, 4-bit greyscale adapter, 19-inch monochrome monitor). I assume that, as was with the RT, manuals and all such are included with this. Oh yeah, one year warranty (not bad considering the 90day standard) tossed in with all this. Not a bad price, even so, considering 27.5 MIPS (peak 100) and 7.5MFLOPS. That works out to be $543.45/MIP system cost which is the lowest in the industry, even among name brand 386/486 PCs, I believe. Data General seems to be runnerup with their 88K based system. Remeber IBM also offers the JMB and NeXtStep interfaces (additional charge, but available). +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan (IBM VNET)WOAN AT AUSTIN, (AUSTIN)ron@woan.austin.ibm.com + + outside of IBM @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron + + last resort woan@peyote.cactus.org +
jet@karazm.math.uh.edu (J. Eric Townsend) (02/20/90)
First off, the sales rep told me that I could buy GNU Emacs and that X for AIX 3.0 would cost around $500. I think IBM sales reps don't know how to deal with $0.00 in the price column, or what public domain (or freely distributable) code is.. In article <10307@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >> . . . a complete system starting at $12,995. >Including OS and window system? Manuals? Ethernet? Compilers? In Nope. "X server software" is $500, plus $50 for each terminal... -- J. Eric Townsend University of Houston Dept. of Mathematics (713) 749-2120 jet@karazm.math.uh.edu Skate UNIX(tm).
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) (02/20/90)
In article <1559@awdprime.UUCP>, ron@woan.austin.ibm.com (R Woan/2100000) writes: |>From the marketing announcement that I have, it looks like a base |>workstation package will run $14,250+$695=$14945 for the base desktop |>model with (8MB RAM, 120MB Hard Disk, Keyboard, Mouse, AIX3, |>AIXwindows, NFS, 4-bit greyscale adapter, 19-inch monochrome monitor). Whoops, I forget to mention that the $695 is for the Ethernet card. By the way this leaves two slots free on the micro-channel bus. Ron +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan (IBM VNET)WOAN AT AUSTIN, (AUSTIN)ron@woan.austin.ibm.com + + outside of IBM @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron + + last resort woan@peyote.cactus.org +
jason@cs.utexas.edu (Jason Martin Levitt) (02/20/90)
In article <5098@brazos.Rice.edu> schafer@brazos.rice.edu (Richard A. Schafer) writes: >I just signed on IBMLINK and ran a configuration on the standard grayscale >configuration. I don't come up with the $12,995 price, but did produce the >following configuration: > >Powerstation 320 $7475 >120MB disk $1950 >Grayscale adapter $1395 >keyboard $ 255 >mouse $ 130 >mono display $1295 > ------ Hardware total, $12,500 >AIX 3.0 $1250 >AIX Xwindows $ 500 > ------ Software total, $ 1,750 > Package total, $14,250 > Let's get real on these prices folks. A 120 MB disk? That *might* work if your user directories are mounted with NFS. I don't know the exact numbers, but I bet after AIX 3.x and Xwindows is loaded plus swap space, 120MB is barely adequate. Add a SCSI adapter and large hard disk please == $$$$$. I doubt IBM will have their diskless version working well for quite a while. Anyone know the exact disk space requirements? Recommended swap space? ---Jason ----- Jason Martin Levitt P.O. Box 49860 Austin, Texas 78765 (512) 459-0055 Internet : jason@cs.utexas.edu | "Toroidal carbohydrate modules? UUCP : ...cs.utexas.edu!hackbox!jason | Make mine glazed!" BIX : jlevitt | -- Zippy
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) (02/20/90)
In article <1990Feb19.213357.7340@lavaca.uh.edu>, jet@karazm.math.uh.edu (J. Eric Townsend) writes: |>First off, the sales rep told me that I could buy GNU Emacs and that |>X for AIX 3.0 would cost around $500. I didn't see the charge for GNU EMACS, but if I understand the GPL, we can't charge more than a minor media and handling fee without restriction of copying from your neighbor. Youch! I looked and our distrbution charge is $125! I guess we have a lot of distribution overhead (manual maybe?) Heck I would suggest that you just roll your own or get it from your neighbor ($125 over 25 workstations wouldn't be bad)... |>"X server software" is $500, plus $50 for each terminal... Every major manufacturer charges for their own X-Window product to cover porting (one heck of an optimization job too) and distribution costs (don't forget full manual set). The $50/terminal results from the brain-dead Motif licensing fees (call OSF up, they'll say they charge $50/term no matter the platform too), so no profit there. All I can say is that my earlier numbers were correct and nowhere will you find a cheaper buck/MIP ratio from a major manufacturer, and our service will be second to none. Ron +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan (IBM VNET)WOAN AT AUSTIN, (AUSTIN)ron@woan.austin.ibm.com + + outside of IBM @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron + + last resort woan@peyote.cactus.org +
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) (02/20/90)
********************* DISCLAIMER ************************* * The following posting is done on my own to provide * * a source of information for people interested in the * * RISC System/6000. Nothing that follows should be * * considered a stance, statement, or position by IBM. * * Contact your IBM representative to get an authorized * * position. I just work here. * ****************** End of Disclaimer ******************** There has been some speculation and confusion in this newgroup about what comes with a Risc System/6000 system. Hopefully, that confusion may be alleviated by an available document: There is a document available from IBM called the 'Risc System/6000 Quick Pricer' (document number G320-9881-0) which provides information on the RISC System/6000 family and AIX Version 3 for RISC System/6000 and US prices. Requests for copies of that publication should be made to your IBM Authorized Dealer or your IBM Marketing Representative. The prices in the document are NOT to be used in lieu of those in the Sales Manual, HONE Configurator, INFOLink, or announcement letters. Anyway, the above-mentioned document contains unofficial prices, as all the information is subject to change. Now that that is laid out, the Quick Pricer says that for a $12,995 POWERstation 320 Desktop Workstation you get: (Under a heading which states that the configuration includes Ethernet, keyboard, 3-button mouse, AIX Version 3 for RISC System/6000 (1-2 user), and graphical user interface), 1. POWERstation 320, 1280x1024 19" Mono Display, Grayscale Graphics Display Adapter, 8 MB RAM, and 120 MB Disk The graphical user interfaces mentioned in the document are AIXwindows Environment/6000 and AIX NextStep Environment/6000. I don't know if the configuration would include your choice of these or what (call your IBM rep). No FORTRAN compiler is supplied with the base OS. FORTRAN, COBOL, Pascal, and Ada are listed in the back as separate items. Hope this helps, but like I said, nothing I posted is official. Yes, I might be paranoid about it, but I want it to be perfectly clear. Marc Stephenson (marc@stingray.austin.ibm.com) Location: F57/992, (79)3-3796, ZIP 2401, 1C-22/992, Austin, Texas Internal: marc@stingray.austin.ibm.com VNET: MARC at AUSVM6 External: uunet!cs.utexas.edu!ibmaus!auschs!stingray.austin.ibm.com!marc
grunwald@foobar.colorado.edu (Dirk Grunwald) (02/20/90)
>> I didn't see the charge for GNU EMACS, but if I understand the GPL, we >> can't charge more than a minor media and handling fee without >> restriction of copying from your neighbor. Youch! I looked and our >> I don't think that the GPL restricts you from charging for Emacs, but you had better distribute (or offer to distribute) the source for the version being sold.
dmcanzi@watserv1.waterloo.edu (David Canzi) (02/20/90)
In article <1565@awdprime.UUCP> @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron writes: >|>"X server software" is $500, plus $50 for each terminal... > >Every major manufacturer charges for their own X-Window product to >cover porting (one heck of an optimization job too) and distribution >costs (don't forget full manual set). The $50/terminal results from >the brain-dead Motif licensing fees (call OSF up, they'll say they >charge $50/term no matter the platform too), so no profit there. Given the ability, inherent in the way X windows works, to use window managers over the network, how can any sane procedure be devised to count and charge for the number of terminals the window manager can (and will) be used on? -- David Canzi
mjt@nagshead.ncsc.org (Mike Tighe) (02/20/90)
In article <1990Feb19.213357.7340@lavaca.uh.edu> jet@karazm.math.uh.edu (J. Eric Townsend) writes: > > >First off, the sales rep told me that I could buy GNU Emacs and that Hmmm. I did not think the GNU Public License allowed allowed the sale of GNU Emacs. It is free software. Also, the source code must also be provided too. -- --- Michael Tighe, mjt@ncsc.org
ehrlich@cs.psu.edu (Daniel Ehrlich) (02/20/90)
In article <1148@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: Jason> In article <5098@brazos.Rice.edu> schafer@brazos.rice.edu (Richard A. Schafer) writes: >I just signed on IBMLINK and ran a configuration on the standard grayscale >configuration. I don't come up with the $12,995 price, but did produce the >following configuration: > >Powerstation 320 $7475 >120MB disk $1950 >Grayscale adapter $1395 >keyboard $ 255 >mouse $ 130 >mono display $1295 > ------ Hardware total, $12,500 >AIX 3.0 $1250 >AIX Xwindows $ 500 > ------ Software total, $ 1,750 > Package total, $14,250 > Jason> Let's get real on these prices folks. A 120 MB disk? That *might* Jason> work if your user directories are mounted with NFS. I don't know the exact Jason> numbers, but I bet after AIX 3.x and Xwindows is loaded plus swap Jason> space, 120MB is barely adequate. The 120Mb disk is less than adequate. The software will barely fit on the 320Mb disk. The on-line hytertext documentation takes up 120Mb alone. Jason> Add a SCSI adapter and large hard disk please == $$$$$. The SCSI adapter card is $US 1200. To upgrade the 120Mb disk to a 320Mb disk will cost $US 3,000. An additional 320Mb disk drive is $US 5,500. -- Dan Ehrlich <ehrlich@cs.psu.edu> Voice: +1 814 863 1142 FAX: +1 814 865 3176
schafer@brazos.rice.edu (Richard A. Schafer) (02/21/90)
Reading the Wall Street Journal this morning, I saw the IBM ad
(3 full pages) for the new systems, and finally found the
fine print that describes what you get for the oft-quoted
$12,995 price.
PowerStation 320 with 8MB of RAM
120MB disk
19" grayscale display and graphics adapter
keyboard
mouse
Ethernet adapter
AIX
"user interface environment" (I don't think this means X-windows,
but don't know what it *does* refer to.)
NFS (it's part of AIX, I think)
hypertext document search & retrieval capability. (I think this
means the software support for hypertext; it clearly does not
refer to the CD-ROM player.
While there may be questions about how much use a 120MB disk
systems is standalone, as a workstation connected to a server,
this sounds like it could be quite interesting. While I wasn't
able to come up with this price in any configuration I did
yesterday, I have no doubt that the price is real.
There have been lots of comments about whether IBM is serious
or not in this market. Can anyone post the commercial prices
of an equivalently-powered workstation for comparison?
My October 1, 1989 SUN pricelist quotes
SparcStation 1 (4/60M1-8-P3)
104MB SCSI internal disk
1.44 diskette drive
19" Mono monitor
8MB RAM $11,495
SunOS Current Standard Release
(no manuals) (SS2-13) $ 450
SparcStation 1 Full System
Documentation Set (SS-09) $ 450
-------
$12,395
(I presume this includes a keyboard and mouse, even though
the price book doesn't mention it. Although we have lots
of Sun's, I don't usually do the configurations myself.)
Given the performance numbers quoted by IBM of the PowerStation
320 versus a SparcStation 1, the fact that the above price has
a *smaller* disk than the one claimed to be totally useless by
Dan Ehrlich, and that the IBM price for AIX includes software
support services (if I read the Sun price list right, an
equivalent level of software support to IBM's would cost you
$95/month = $1140/year, almost the price of AIX every year),
plus a longer warranty on the IBM equipment (1 year versus
90 days on Sun's), my calculations seem to suggest that the
(much?)-less-powerful SparcStation 1 actually costs a little bit *more*
then the equivalently configured PowerStation 320.
That's the only prices I have available. Anyone else care to
provide DEC, etc., equivalents?
Richardgriefer@adg.almaden.ibm.com (Allan D. Griefer) (02/21/90)
The following prices are accurate to the best of my knowledge, but should
not be used without checking with an IBM Salesperson.
UNIT MDL/FC DESCRIPTION QTY COST
7012-320 IBM RISC/6000 PWRSTA/SVR 320 1 7475.00P
2760 GRAYSCALE GRAPH DISP ADPT 1 1395.00P
2980 ETHERNET HIGH PERF LAN ADPTR 1 695.00P
6010 KEYBD 101 KEYS U.S. 1 255.00P
6041 3-BUTTON MOUSE 1 130.00P
8508-001 PS/2 MONOCHROME DISPLAY 1 1295.00P
HARDWARE TOTALS 11245.00*
5756-030 AIX V3 FOR RISC SYSTEM/6000 1
0200 BASIC OTC 1-2 UL GRP E5 1 1250.00
5151 ENCRYPTION FEATURE 1
5601-257 AIXWINDOWS ENVIRONMENT/6000 1
0173 BASIC OTC GROUP E5 1 500.00
SOFTWARE TOTALS 1750.00*
GRAND TOTAL 12995.00
This is not a complete list of order features as it doesn't include a bunch
of no cost option information. This configuration has 120MB disk, 8MB of
memory and two free slots. It comes with the complete AIX 3 for 1-2 users
and includes AIXWindows which includes X and MOTIF.
I hope this straightens out some of the misconceptions about what you get
for $12,995. It is, in fact a complete system.
Opinions are strictly my own,
Allan D. Griefer, IBM Almaden Research Center, San Jose, CA
BITNET: GRIEFER at ALMADEN Internet: griefer@ibm.com
UUCP: ...!uunet!ibmarc!griefer mcimail: 398-8024jsw@xhead.SGI.COM (Jeff Weinstein) (02/21/90)
Could someone please post the prices for the various 3D graphics options and systems? Jeff Weinstein - X Protocol Police Silicon Graphics, Inc., Entry Systems Division, Window Systems jsw@xhead.esd.sgi.com Any opinions expressed above are mine, not sgi's.
griefer@adg.almaden.ibm.com (Allan D. Griefer) (02/21/90)
There's a $125 distribution charge for GNU EMACS, but, last time I looked, FSF was charging $150 for distribution. Opinions are strictly my own, Allan D. Griefer, IBM Almaden Research Center, San Jose, CA BITNET: GRIEFER at ALMADEN Internet: griefer@ibm.com UUCP: ...!uunet!ibmarc!griefer mcimail: 398-8024
griefer@adg.almaden.ibm.com (Allan D. Griefer) (02/21/90)
It seems that there is a little confusion on the "X server code." This code is to support Xstation 120s since they load over the network and the $50 charge is the license fee for each of the Xstations. AIXwindows Environment 600, which also costs $500 is the IBM version of X plus MOTIF plus other enhancements and doesn't have a multiple terminal fee as far as I can tell. Opinions are strictly my own, Allan D. Griefer, IBM Almaden Research Center, San Jose, CA BITNET: GRIEFER at ALMADEN Internet: griefer@ibm.com UUCP: ...!uunet!ibmarc!griefer mcimail: 398-8024
rcd@ico.isc.com (Dick Dunn) (02/21/90)
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) writes: ...[price info deleted] > Not a bad price, even so, considering 27.5 MIPS (peak 100) and > 7.5MFLOPS. That works out to be $543.45/MIP system cost which is the > lowest in the industry, even among name brand 386/486 PCs, I believe. Let me start by saying that the new IBM boxes look like they'll actually come in at believable prices when they're available. That nicety out of the way,... 27.5 MIPS is a junk number. There is no useful meaning to "MIPS" which justifies expressing three significant digits, especially with a variable instruction-issue rate. ONE significant digit is pushing it. $543.45 (FIVE figures?!?!?!?) is doubly bogus, since it's calculated by bashing the barely-meaningful MIPS number against an even-less-relevant cost. Think about it: The easiest way to reduce the $/MIP ratio is to offer a more stripped-down configuration! Then there's the business of reporting Dhrystone 1.1 numbers. That's just a disgrace. I have to wonder if it took some special effort to dig up a copy of the old Dhrystone 1.1 benchmark instead of using a "current" (can we call it current even tho it's more than a year old?) version! And, it seems, the MIPS numbers are calculated off these lousy Dhrystone values. (I *hope*--for IBM's sake--that they're lousy! Their compiler had better be good enough to over-optimize Dhrystone 1.1.:-) If we're going to discuss benchmarks on these machines, let's start with something halfway reasonable, like SPEC. If you want some meaningful numbers, you're going to have to think a little bit. You may even find--horror of horrors!--that you won't be able to describe the cost-effectiveness of a machine, compared to all others in the market, with a single number. -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...Don't lend your hand to raise no flag atop no ship of fools.
shair@ux1.cso.uiuc.edu (02/21/90)
At the risk of being accused of advertising, I'd like to clarify a couple of issues about disk drive support on the desktop RIOS (POWERStation 320). The 320 has room for the installation of two 3.5" disk drives. The standard configuration comes with a single 120MB ESDI drive. This is the same drive used in the PS/2 model 70 (23ms avg seek, 1.3MB/sec transfer) Feature number 2120, 120MB disk drive, adds the SECOND drive for $1,950. The configurations including that have had a total of 240MB. Alternatively one can purchase Feature number 2828 (SCSI adapter) for $1,200 and Feature number 2541 (Select 320MB disk) for $3,000. These two features provide one high-performance (12.5ms avg seek, 2MB/sec xfer) 320MB disk. Once one has the SCSI adapter, any number of other interesting devices can be installed, including (exclusive of cable costs) Second 320MB disk $5,500 External CD ROM $1,695 External QIC tape $1,995 External 8mm tape $6,650 (but this is really beginning to look like advertising...) The 120MB disks do NOT take one of the microchannel slots, so systems with these disks, display and ethernet have two slots open. The SCSI adapter would take one of them if 320MB disks are used. Having one of each disk type appears to be a valid configurations. My own machine has a 320 and a 120.
shair@ux1.cso.uiuc.edu (02/21/90)
Oops, forgot to sign.... as someone noted earlier in this string, we in the IBM marketing force are still not completely nativized to the unix environment. Bob Shair Scientific Computing Specialist IBM Champaign
emv@math.lsa.umich.edu (Edward Vielmetti) (02/21/90)
[ disk options on the RS/6000 aha, it's an esdi drive that's built in. I wondered. 23ms, 1.3MB/sec transfer is wimpy for a fast machine. In this configuration the machine is going to be seriously i/o bound, without a doubt. 12.5ms, 2MB/sec transfer on the 320MB disk is better, but that's not top of the line these days -- not for SCSI (sync will go faster) and certainly not for disk in general. I don't see a real fast disk for these things. Can we get a real word on the scsi adapter -- i.e. - is it a part that's being sold now for the microchannel, or a new thing ? - is it SCSI-1 or SCSI-2, does it support synchronous operation, etc. - are there any problems with dropping in a microchannel SCSI adapter except perhaps that of getting device driver support? The SCSI adapter I saw had an 80C186 and a big IBM chip (must be some ASIC thing) on it, the copyright on the 80C186 was 1979 -- hardly state of the art in chips ! --Ed
woan@peyote.cactus.org (Ronald S. Woan) (02/21/90)
In article <1990Feb21.012432.22401@ico.isc.com>, rcd@ico.isc.com (Dick Dunn) writes: > ron@woan.austin.ibm.com (Ronald S. Woan/2100000) writes: > ...[price info deleted] > > Not a bad price, even so, considering 27.5 MIPS (peak 100) and > > 7.5MFLOPS. That works out to be $543.45/MIP system cost which is the > > lowest in the industry, even among name brand 386/486 PCs, I believe. > > Let me start by saying that the new IBM boxes look like they'll actually > come in at believable prices when they're available. That nicety out of > the way,... > > 27.5 MIPS is a junk number. There is no useful meaning to "MIPS" which You are definitely right. I didn't look to carefully at how they came up with the numbers and Dhrystone 1.1 is definitely a wimpy way to go (yeah, I can flame my cohorts). Anyway, I got a little carried away at the marketing hype and the numbers that Data General loves to quote. But a better mark I guess would be the SPECS 20-30 (I believe), so for the $12995 deal in the Wall Street Journal (is this a limited time offer or where did they arrive at that price?) you get a system at $650/SPEC or so. Why are SPECs the geometric mean rather than average by the way? As with all hardware, try before you buy; I think that you will be impressed especially with the graphics workstation model. One last marketing slogan before I go that seems a little more apt than the buck/MIP stuff: "buy a solution not just a hunk of hardware." As for the question of service, I think that we are really commited to it this time and are improving problem and system tracking as well as beefing up trained personnel. Some people don't realize it, but internal IBM'ers have to call the same service people as our customers, so for our own sakes I know we are trying our best (yeah, I've fried an RT and two mice in last eight months and some of the service personnel never seemed to have heard of an RT, not much enough units out there I guess). My apologies all, Ron -- +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan @cs.utexas.edu:romp!auschs!woan.austin.ibm.com!ron + + second choice: woan@peyote.cactus.org +
peter@tern.stars.flab.Fujitsu.JUNET (Peter S. Ford) (02/21/90)
IBM's strategy for unbundling LAN and disk adaptors and software is not uncommon for the industry. I can never understand why people see unbundling as the GREAT evil. Is it that hard to read the price list and sum up the total cost and THEN do your comparisons? I seem to remember people slamming NeXT for not letting you buy a machine w/o an optical disk. wrt benchmarks. We have run over 10 science codes on the RS6000 and it is usually 4 times faster than a Sparcstation. No SPECmarks, no FLOPS, no XXXpacks, just 4 times faster. Your mileage may vary. This is the 520 model, which is the 20 MHz version of the machine and has a shallower data cache than the 25 & 30 MHz models. re screen sizes. All screens are 1280x1024. If you want a cheaper (albeit slower) machine you still buy a 3100 or a SparcStation. If you need floating pt crunch, it looks like the RS 6000 is a good choice (you may want to llok at the MIPS R/6000 as well, although not as a desktop workstation). Peter Ford, Center for Nonlinear Studies, Los Alamos -- Peter Ford MS-B258 Center for Nonlinear Studies Los Alamos, NM 87545 peter@lanl.gov
windley@cheetah.ucdavis.edu (Phil Windley/20000000) (02/22/90)
In article <5098@brazos.Rice.edu> schafer@brazos.rice.edu (Richard A. Schafer) writes:
I just signed on IBMLINK and ran a configuration on the standard grayscale
configuration. I don't come up with the $12,995 price, but did produce the
following configuration:
Powerstation 320 $7475
120MB disk $1950
Grayscale adapter $1395
keyboard $ 255
mouse $ 130
mono display $1295
------ Hardware total, $12,500
AIX 3.0 $1250
AIX Xwindows $ 500
------ Software total, $ 1,750
Package total, $14,250
An IBM ad in the San Francisco Chronicle yesterday quoted the $12995 price
for a 320 with gretscale monitor, 8Mb RAM, 120 Mb disk, ethrenet, AIX, and
AIX Windows. I don't know what kind of pricing arrangement it is, but
apparently, they're selling a "complete" machine for that price.
--
Phil Windley | windley@cheetah.ucdavis.edu
Division of Computer Science | ucbvax!ucdavis!cheetah!windley
University of California, Davis |
Davis, CA 95616 | (916) 752-6452 (or 3168)windley@cheetah.ucdavis.edu (Phil Windley/20000000) (02/22/90)
In article <1148@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes:
Let's get real on these prices folks. A 120 MB disk? That *might*
work if your user directories are mounted with NFS. I don't know the exact
numbers, but I bet after AIX 3.x and Xwindows is loaded plus swap
space, 120MB is barely adequate.
Add a SCSI adapter and large hard disk please == $$$$$.
I doubt IBM will have their diskless version working well for quite a
while. Anyone know the exact disk space requirements? Recommended swap
space?
First you complain about 120Mb being too small and then you want diskless.
I've used RT's with a 70MB disk with user files NFS mounted. I'd rather
have thsi configuration than a diskless product. System stuff goes much
faster.
Of course big disks cost money, but that's true of any system. So what's
your gripe?
--
Phil Windley | windley@cheetah.ucdavis.edu
Division of Computer Science | ucbvax!ucdavis!cheetah!windley
University of California, Davis |
Davis, CA 95616 | (916) 752-6452 (or 3168)jason@cs.utexas.edu (Jason Martin Levitt) (02/22/90)
In article <WINDLEY.90Feb21094531@cheetah.cheetah.ucdavis.edu> windley@cheetah.ucdavis.edu (Phil Windley/20000000) writes: >>In article <1148@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >> >> Let's get real on these prices folks. A 120 MB disk? That *might* >> work if your user directories are mounted with NFS. I don't know the exact >> numbers, but I bet after AIX 3.x and Xwindows is loaded plus swap >> space, 120MB is barely adequate. >> >> Add a SCSI adapter and large hard disk please == $$$$$. >> >> I doubt IBM will have their diskless version working well for quite a >> while. Anyone know the exact disk space requirements? Recommended swap >> space? > >First you complain about 120Mb being too small and then you want diskless. >I've used RT's with a 70MB disk with user files NFS mounted. I'd rather >have thsi configuration than a diskless product. System stuff goes much >faster. > >Of course big disks cost money, but that's true of any system. So what's >your gripe? > Someone responded to my question about disk space requirements and said that AIX 3.1 will barely fit on a 300mb hard disk. If it's possible to run some kind of minimal system configuration in 120mb, will someone please confirm it? BTW, they posted the reply to pc.rt or unix.aix. Yes, you can use AIX 2.2.1 with just a 70mb disk. But if you have NFS *and* X-WIndows and a reasonable amount of swap space, things are going to get very tight. But that's not the point. The point is that this is definitly *not* AIX 2.2.1, it's AIX 3.1. AIX 3.1 is a big, hairy, hulking, monster of an operating system. My gripe is that people are quoting this $12,000 price for a system that won't run. What is this 120mb disk for if you can't put the operating system on it? If it's a Sun, then it's local swap or other interesting things that a node which boots off of a server might need. Since IBM hasn't introducted any diskless technology yet, you have no choice, you must buy a SCSI adapter and big hard disk. If IBM's software technology can match their hardware, then their diskless node should be a wonderful, inexpensive workstation.....*if* their software technology can match their hardware........ ---Jason ----- Jason Martin Levitt P.O. Box 49860 Austin, Texas 78765 (512) 459-0055 Internet : jason@cs.utexas.edu | I UUCP : ...cs.utexas.edu!hackbox!jason | put the BIX : jlevitt | chic in geek.
dyer@spdcc.COM (Steve Dyer) (02/22/90)
In article <1152@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: > Someone responded to my question about disk space requirements and >said that AIX 3.1 will barely fit on a 300mb hard disk. If it's possible >to run some kind of minimal system configuration in 120mb, will someone >please confirm it? BTW, they posted the reply to pc.rt or unix.aix. This is one of the most ridiculous things I've ever heard. NO UNIX system requires 120mb of disk, and especially AIX, since it's broken into dozens of LPPs which may or may not be installed, depending on the user's available space and requirements. So don't load the Cobol compiler if you have it and don't need it. But don't blame IBM if you can't put everything they sell on a smallish disk. -- Steve Dyer dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer dyer@arktouros.mit.edu, dyer@hstbme.mit.edu
drake@sd2.almaden.ibm.com (Sam Drake) (02/22/90)
In article <1656@speedy.mcnc.org> mjt@nagshead.ncsc.org (Mike Tighe) writes: >In article <1990Feb19.213357.7340@lavaca.uh.edu> jet@karazm.math.uh.edu (J. Eric Townsend) writes: >>First off, the sales rep told me that I could buy GNU Emacs and that >Hmmm. I did not think the GNU Public License allowed allowed the sale of >GNU Emacs. It is free software. Also, the source code must also be provided >too. You are not allowed to SELL GNU Emacs, but you are allowed to charge a reasonable fee to duplicate it. IBM's fee to duplicate GNU Emacs for you ($125) is, I believe, less than the fee that the FSF charges to do the same thing, and this is perfectly fine according to the copyleft. Opinions are my own.... Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
dyer@spdcc.COM (Steve Dyer) (02/22/90)
8mb is probably tight for any current RISC machine. But IBM's
entry level system for $13K is at least usable (if slower than
optimal) out of the box unlike the ridiculous low-ball prices
for Sun and DEC diskless machines which aren't usable unless you
have another big machine acting as a server out on your net.
I think IBM should be as accountable as any of the other players,
but I'm a little tired of hearing them be dragged over the coals
for practices which, if anything, are more generous than their
competition. Sun OS 4.x has never been thought of as small,
yet you never heard the same bitching and kvetching when they
announced the SparcStation 1 with 8mb of memory and a 100+change
megabyte disk.
If you add more disk and memory, you still get something which
is as, if not more, competitive than what the competition offers now.
--
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer
dyer@arktouros.mit.edu, dyer@hstbme.mit.edujason@cs.utexas.edu (Jason Martin Levitt) (02/22/90)
In article <1791@ursa-major.SPDCC.COM> dyer@ursa-major.spdcc.COM (Steve Dyer) writes: >In article <1152@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >> Someone responded to my question about disk space requirements and >>said that AIX 3.1 will barely fit on a 300mb hard disk. If it's possible >>to run some kind of minimal system configuration in 120mb, will someone >>please confirm it? BTW, they posted the reply to pc.rt or unix.aix. > >This is one of the most ridiculous things I've ever heard. NO UNIX system >requires 120mb of disk, and especially AIX, since it's broken into dozens >of LPPs which may or may not be installed, depending on the user's available >space and requirements. So don't load the Cobol compiler if you have it and >don't need it. But don't blame IBM if you can't put everything they sell >on a smallish disk. > Great Steve, I'm glad you think it's ridiculous. I want you to know that *I* think it's ridiculous too. Now, if someone would like to post some real numbers, we can decide how ridiculous it is. Only one person has posted so far, and they seemed to think that it requires a lot more than 120mb, but they probably have a test machine with all the LPPs loaded and maybe a lot of debug stuff. I would not be at all surprised if AIX 3.1 with C development tools, NFS, X-windows/Motif, man pages + user friendly junk and swap space pushed a 120 mb hd to the limit if not further. You might be able to save some space by eliminating the man pages and user friendly junk :-). Are the disk space requirements covered by non-disclosure still? ---Jason ----- Jason Martin Levitt P.O. Box 49860 Austin, Texas 78765 (512) 459-0055 Internet: jason@cs.utexas.edu | "The most effective debugging tool is UUCP : cs.utexas.edu!hackbox!jason | still careful thought, coupled with BIX : jlevitt | judiciously placed print statements." | -Brian Kernighan [1978]
drake@sd2.almaden.ibm.com (Sam Drake) (02/22/90)
In article <1990Feb21.012432.22401@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes: > >If we're going to discuss benchmarks on these machines, let's start with >something halfway reasonable, like SPEC. Good idea. At the San Francisco announcement, Phil Hester from IBM said that the RISC System/6000 scored the highest SPEC score that any workstation had ever rated, and gave the numbers. I don't have 'em in front of me, but I believe the SPEC rating was >30. Opinions are my own. Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
khb@chiba.Eng.Sun.COM (Keith Bierman - SPD Advanced Languages) (02/22/90)
In article <1990Feb21.012432.22401@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes:
...
Then there's the business of reporting Dhrystone 1.1 numbers. That's just
a disgrace.
....
Far be it from me to defend IBM; but dhrystone 1.1 is what many of the
chip houses still use for quoting. Thus many marketing departments
feel obligated to use it. Also, the kinder gentler compiler groups
carefully detune the optimizer to make the benchmark not disappear.
IBM was quite good about reporting SPEC numbers, and complete postings
of such can be found in comp.arch.
cheers 'all
--
Keith H. Bierman |*My thoughts are my own. !! kbierman@Eng.Sun.COM
It's Not My Fault | MTS --Only my work belongs to Sun* kbierman%eng@sun.com
I Voted for Bill & | Advanced Languages/Floating Point Group
Opus | "When the going gets Weird .. the Weird turn PRO"
"There is NO defense against the attack of the KILLER MICROS!"
Eugene Brookskhb@chiba.Eng.Sun.COM (Keith Bierman - SPD Advanced Languages) (02/22/90)
In article <415@peyote.cactus.org> woan@peyote.cactus.org (Ronald S. Woan) writes:
you get a system at $650/SPEC or so. Why are SPECs the geometric mean rather
than average by the way?
Because the arithmetic mean (the "average") isn't reasonable when
comparing different rates. I've done a writeup or two, but I've
misplaced my brain. Let it suffice to note that:
time A
.. vs time (A+..+N)
time N
and compute the average(s). With the arithmetic mean, the single
number is jerked up a lot by one big outlier. The harmonic mean
produces the same value, which is argued fair by the original
livermore loops report, and other folks ... but which turns out to
penalize a system for screwing up on one particular code by a lot. The
geometric mean falls somewhere between.
When I do the algebra, I pick the harmonic mean. When I stare at
enough benchmark figures, the geometric mean seems more sensible.
Others get different results. I wasn't there for the SPEC meetings,
but I bet this was debated for a bit and decided along Mashey's lines.
--
Keith H. Bierman |*My thoughts are my own. !! kbierman@Eng.Sun.COM
It's Not My Fault | MTS --Only my work belongs to Sun* kbierman%eng@sun.com
I Voted for Bill & | Advanced Languages/Floating Point Group
Opus | "When the going gets Weird .. the Weird turn PRO"
"There is NO defense against the attack of the KILLER MICROS!"
Eugene Brooksibmjb@garnet.berkeley.edu (02/22/90)
In article <1514@fcs280s.ncifcrf.gov> adam@fcs260c.UUCP (Adam W. Feigin) writes: >In article <10307@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: >>> . . . POWERstations and POWERservers . . . >> >>I like the way they adopted the Sun Marketing naming convention >>(SPARCstations and SPARCservers). > >Well, since they licensed the graphics technology from SGI, and they >call their machines the "POWER Series" and IBM is going after Sun and >their SPARCstations/servers, its natural that IBM should call their >machines in this manner. > >> >>> . . . a complete system starting at $12,995. >> >>Including OS and window system? Manuals? Ethernet? Compilers? In >>the original RT/PC announcement a few years ago, the lowball price >>didn't even include a keyboard or monitor, which cost $thousands! One >>analysis showed that if Sun had charged as much for Unix as IBM, they >>could've thrown in the workstation for free! > >Correct. The base price does NOT include OS & Window system or >ethernet (I assume that you get manuals & compilers with the OS, but I >could be wrong, and I certainly wouldn't put it past IBM NOT to >include them, and charge extra for them, as they have a nasty habit of >doing). I did notice that in the glossy that the 3-button mouse and >keyboard are marked as "optional products".. Nice, very nice, looks >like IBM is up to their old tricks again. They never learn, do they ?? > The OS and window system and ethernet are included. >(BTW the price sheet I have show the OS + X-Window System at $2000) > >>Also, one breakdown I saw was that this system only has four slots -- >>and all of them are in use. No expandability. > >I'm not sure about this, but the base system comes with a 120MB >DBA ("Direct Bus Attached" -- whatever that means) disk; I dont know >if you need a slot for a controller, but if you want to add more disk, >you gotta buy a controller. Lets not forget a slot for ethernet >(optional), a slot for graphics (you really dont want to see anything, >do you ??, you can just watch the blinkin' lights...) > Only one of the expansion slots is used in this configuration (display adapter). >>> Documentation is available on a 5.25-inch compact disc >>> (CD-ROM) that may be accessed from the user's POWERstation or from a >>> network POWERserver. >> >>At extra cost > >Of course !! > >It will be a cold day in hell when the temperature is below freezing >when IBM gets serious about the Unix/Workstation market. > The extra cost is only $115, which is pretty cheap if you ask me... Yes, IBM IS serious about the workstation business. Skeptics beware.
emv@math.lsa.umich.edu (Edward Vielmetti) (02/22/90)
I stared at the complete price list that I have and have some questions. First of all the base 320 box comes with 8MB of memory and a 120MB ESDI drive, nothing else, for $7475. List for the 2nd 120MB ESDI in a system is $1950, list for the 8MB is roughly $4500 -- if you can subtract like I can subtract they're selling the box for about $1000 and making their money on the bundled peripherals. I suppose the disk earns its keep for having the OS preloaded, that's a win, though at 27ms access it's way too slow & behind the times in drive technology. A builtin SCSI would have made more sense than builtin ESDI in my opinion. The memory is completely unfair of course -- no one can order a memoryless workstation no matter how hard you try -- the other way to look at it is that memory is currently costing $100/MB in commodity SIMM parts, or 2-3x that in semi-proprietary designs, figure $1200 for the card carrier to hold the simms and the memory is overpriced by a factor of two. Can you add PS/2 memory to the 8MB card in 1MB increments? I would hope that you could get it for less than the $495/MB list they have for it. How much memory fits on one memory card ? Will the same memory card hold 4MB SIMM units or do you have to have the other memory card for that (list $20K for 32MB btw). --Ed
rcd@ico.isc.com (Dick Dunn) (02/22/90)
griefer@adg.almaden.ibm.com (Allan D. Griefer) writes: > There's a $125 distribution charge for GNU EMACS, but, last time I looked, > FSF was charging $150 for distribution. Ummm...yes, but the FSF charge is for the entire release tape, which contains not only emacs but X 10.4, Scheme, Bison, T, hack, GNU Chess, and GDB. Main point is that it takes a 9-track tape, which isn't cheap and costs money to handle and ship. Still, $125 is a little high but not terrible (is it?) for something handled by a large company. After all, it's more than just a floppy (esp. if the source is there, which it better be!) -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...Don't lend your hand to raise no flag atop no ship of fools.
rcd@ico.isc.com (Dick Dunn) (02/22/90)
khb@chiba.Eng.Sun.COM (Keith Bierman - SPD Advanced Languages) writes: [I wrote] > Then there's the business of reporting Dhrystone 1.1 numbers. That's just > a disgrace. ... > Far be it from me to defend IBM; but dhrystone 1.1 is what many of the > chip houses still use for quoting. Thus many marketing departments > feel obligated to use it... "But Mom! Everybody else is doin' it!" I won't retract my flame about this. It's still wrong. It's still either dishonest (if they know better) or mind-bogglingly ignorant (if they don't). What I should have pointed out is that IBM is not alone in this sort of travesty...which does not make it any less a travesty. As long as marketing departments feel they can get away with reporting an over-optimized 1.1 Dhrystone, they're likely to do it. That's why I'm making noise. I don't think we should let them get away with it. Also, there are folks who use 2.1 Dhrystone, and they should not be penalized for doing so. Also remember that the big headline-generating MIPS numbers came directly from this old benchmark-of-ill-repute. Anybody got some 2.1 Dhrystone numbers on the 6000 boxes? Anybody looked at the generated code to see if they're playing by the rules? -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...Don't lend your hand to raise no flag atop no ship of fools.
buck@siswat.UUCP (A. Lester Buck) (02/22/90)
> Why are SPECs the geometric mean rather > than average by the way? If you think about MIPS, etc., on a log scale, then the geometric mean is the average of the logarithms. -- A. Lester Buck buck@siswat.lonestar.org ...!texbell!moray!siswat!buck
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) (02/22/90)
One thing that people might wish to consider is that the UNIX market is no longer limited to scientific/engineering; it is growing quite a bit in the commercial sector. Whereas a minimal system might not work so well for someone developing an application, it might work just great for someone who is running the application. Why should they have to pay for a lot of disk and memory if all that they are doing is running terminals at a video checkout counter? Whether a customer needs more performance and disk space will largely be controlled by what they are going to use the systems for. Marc Stephenson (marc@stingray.austin.ibm.com) Location: F57/992, (79)3-3796, ZIP 2401, 1C-22/992, Austin, Texas Internal: marc@stingray.austin.ibm.com VNET: MARC at AUSVM6 External: uunet!cs.utexas.edu!ibmaus!auschs!stingray.austin.ibm.com!marc
pcg@aber-cs.UUCP (Piercarlo Grandi) (02/22/90)
In article <EMV.90Feb20220637@duby.math.lsa.umich.edu> emv@math.lsa.umich.edu (Edward Vielmetti) writes:
[ disk options on the RS/6000
aha, it's an esdi drive that's built in. I wondered.
23ms, 1.3MB/sec transfer is wimpy for a fast machine.
In this configuration the machine is going to be seriously
i/o bound, without a doubt.
Pah. The bottleneck is the filesystem, unless you do asynch io via a raw
device. You cannot get more than 600KB per second out of the filesystem in
the best of circumstances, and even that is only achieved, as far I know, by
the MIPS UNIX. Others top out at around 300KB per second.
Better seek times improve things a bit. Multiple drives, with overlapped
seek and transfer, improve things much more for a timeshared system. It is
here, and not in higher transfer rates (or even seek times) that SCSI wins
over ESDI. But the advantage is nonexistent if you have only one drive.
If your only worry is single task fast transfer rate (signal/image
processing), be prepared to implement something like the Amoeba or Dartmouth
or Cray file systems.
The problem is software, not hardware.
--
Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth | UUCP: ...!mcvax!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.ukjrg@Apple.COM (John R. Galloway Jr.) (02/23/90)
In article <1395@ks.UUCP> drake@ibmarc.uucp (Sam Drake) writes: >In article <1990Feb21.012432.22401@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes: >> >>If we're going to discuss benchmarks on these machines, let's start with >>something halfway reasonable, like SPEC. > >Good idea. At the San Francisco announcement, Phil Hester from IBM said >that the RISC System/6000 scored the highest SPEC score that any workstation >had ever rated, and gave the numbers. I don't have 'em in front of me, >but I believe the SPEC rating was >30. comp.arch has lots of SPEC numbers (back articles even have a fair variety of other systems to compare against). -jrg -- apple!jrg John R. Galloway, Jr. contract programmer, San Jose, Ca These are my views, NOT Apple's, I am a GUEST here, not an employee!!
drake@sd2.almaden.ibm.com (Sam Drake) (02/23/90)
In article <13531@granite.BBN.COM> mlandau@bbn.com (Matt Landau) writes: >Now given the rumors that the trade press has published quite openly >about the AIX 3.1 kernel size, I'll let others decide whether *they* >think 8 MB is enough on a RIOS...) Matt, don't forget that the AIX 3 kernel is pageable, so "kernel size" doesn't really say much about memory requirements. In other *IX systems, one might take the real memory, subtract the kernel size and then get worried; the effect is more complex (and less dramatic) with a pageable kernel. Opinions are my own! Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
drake@sd2.almaden.ibm.com (Sam Drake) (02/23/90)
In article <1152@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: > Someone responded to my question about disk space requirements and >said that AIX 3.1 will barely fit on a 300mb hard disk. If it's possible >to run some kind of minimal system configuration in 120mb, will someone >please confirm it? BTW, they posted the reply to pc.rt or unix.aix. Sure, fits, no problem. If you have lots of Program Products (compilers, extra cost applications, etc) you would probably want to NFS mount them, along with /u. Come on, IBM's not going to run 3 page ads in every big paper in the US advertising a configuration that won't boot! Heck, if it wouldn't boot with a 120MB disk, we would have priced a configuration with a 10MB disk ... woulda really been cheap then! :-) :-) :-) :-) :-) :-) > But that's not the point. The point is that this is definitly >*not* AIX 2.2.1, it's AIX 3.1. AIX 3.1 is a big, hairy, hulking, monster >of an operating system. I don't think the facts bear this one out, especially given that it will boot in 120MB. > > My gripe is that people are quoting this $12,000 >price for a system that won't run. What is this 120mb disk for if you >can't put the operating system on it? If it's a Sun, then it's local >swap or other interesting things that a node which boots off of a server >might need. Since IBM hasn't introducted any diskless technology yet, you >have no choice, you must buy a SCSI adapter and big hard disk. All refuted above. Not true. Naturally, the local disk is for AIX itself and for local swap (perhaps a small /tmp). Again, you're no worse off (and in fact far BETTER off) than you would be with a diskless configuration. How can you lose? Opinions are my own! Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
guy@auspex.auspex.com (Guy Harris) (02/23/90)
>> > IBM announced industry-leading graphical user interfaces . . . >> > available as separate licensed programs. Customers can choose the >> > interface most applicable to their environments. >> >> At extra cost >Yep, since they allow you to choose Ethernet or Token-Ring adapters, >these are optional items. In addition, you can configure the machine >as an asynch timesharing machine, which a host of asynch adapters. So >I don't think it's unreasonable to do it this way. Uhh, while it may be legitimate to have a GUI be an add-on product for a machine that can be sold as an asynch timesharing machine, I fail to see what the fact that network interfaces are optional has to do with the fact that the GUI is optional; as I'm *sure* IBM and everybody in the audience is aware, you don't need a network interface to run X11....
fred@cdin-1.UUCP (Fred Rump) (02/23/90)
In article <EMV.90Feb22004605@duby.math.lsa.umich.edu> emv@math.lsa.umich.edu (Edward Vielmetti) writes:
<>I stared at the complete price list that I have and have some
<>questions.
<>The memory is completely unfair of course -- no one can order a
<>memoryless workstation no matter how hard you try -- the other way to
<>look at it is that memory is currently costing $100/MB in commodity
<>SIMM parts, or 2-3x that in semi-proprietary designs, figure $1200 for
<>the card carrier to hold the simms and the memory is overpriced by a
<>factor of two.
Isn't that how all of IBM's prices are arranged? How else could they offer
close to 50% off to the market and still make money?
No one has yet spoken off 'real' prices yet. It's been all list. Who buys IBM
at list?
Let's wait for some street prices and see where things fall.
fred
--
Fred Rump | UUCP: {uunet bpa dsinc}!cdin-1!fred
CompuData, Inc. | or ...{allegra killer gatech!uflorida decvax!ucf-cs}
10501 Drummond Rd. | !ki4pv!cdis-1!cdin-1!fred
Philadelphia, Pa. 19154| Internet: fred@cdin-1.uu.net (215-824-3000)markb@denali.sgi.com (Mark Bradley) (02/23/90)
In article <1660@aber-cs.UUCP>, pcg@aber-cs.UUCP (Piercarlo Grandi) writes: > In article <EMV.90Feb20220637@duby.math.lsa.umich.edu> emv@math.lsa.umich.edu (Edward Vielmetti) writes: > > [ disk options on the RS/6000 > > 23ms, 1.3MB/sec transfer is wimpy for a fast machine. > In this configuration the machine is going to be seriously > i/o bound, without a doubt. > > Pah. The bottleneck is the filesystem, unless you do asynch io via a raw > device. You cannot get more than 600KB per second out of the filesystem in > the best of circumstances, and even that is only achieved, as far I know, by > the MIPS UNIX. Others top out at around 300KB per second. > > Better seek times improve things a bit. Multiple drives, with overlapped > seek and transfer, improve things much more for a timeshared system. It is > here, and not in higher transfer rates (or even seek times) that SCSI wins > over ESDI. But the advantage is nonexistent if you have only one drive. > > If your only worry is single task fast transfer rate (signal/image > processing), be prepared to implement something like the Amoeba or Dartmouth > or Cray file systems. > > The problem is software, not hardware. Pah, indeed. I am measuring >6 MB/sec. through our filesystem today, abeit not with SCSI. Our SCSI (synchronous) is only a bit over 2 MB/sec. on a single drive. Striping and other wonderful *software* things do much more. ---Through the filesystem, mind you.--- And ESDI is much, much faster in certain applications in that one can better sort, queue and optimize the performance that is limited by the speed of the drives' mechnaisms. It must be agreed, however, that if the software does not permit full utilization of the raw speed of the hardware, then the speed of that hardware does very little for one. markb -- Mark Bradley "Faster, faster, until the thrill of I/O Subsystems speed overcomes the fear of death." Silicon Graphics Computer Systems Mountain View, CA 94039-7311 ---Hunter S. Thompson ******************************************************************************** * Disclaimer: Anything I say is my opinion. If someone else wants to use it, * * it will cost... * ********************************************************************************
abstine@image.soe.clarkson.edu (Arthur Stine) (02/23/90)
From article <1596@awdprime.UUCP>, by marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22): > One thing that people might wish to consider is that the UNIX market is no > longer limited to scientific/engineering; it is growing quite a bit in the > commercial sector. Whereas a minimal system might not work so well for someone > developing an application, it might work just great for someone who is running > the application. Why should they have to pay for a lot of disk and memory if > all that they are doing is running terminals at a video checkout counter? > Whether a customer needs more performance and disk space will largely be > controlled by what they are going to use the systems for. > Is this to say that IBM thinks that people who need terminals at video checkout counters are going to buy RS/6000's? Come on, lets be real. If you want to be in that market, then fine. But if you want to be in the tech workstation market, then package the systems for it. The price/performance of the RS/6000 is nice, but the configuration is not. Throw away the ESDI and put together a system with a couple of fast SCSI channels (read: SYNC scsi) and make a package that is along the same lines as Sun and DEC. But remember, even if the machine has spectacular performance, if it doesn't come down to the right price for a usable system, it doesn't matter how much faster it is. If the buyer just doesn't have the extra $$, then the extra performance doesn't matter. Hence, Sun/DEC still have a good advantahe by having diskless machines under $8K. And at least with these systems, you can add on a cheap ($) scsi disk, since the controller is included. The techie market is different than the commercial market. Techies aren't as worried about support as they are about flexibility and lowest bottom line. -- Art Stine Sr Network Engineer Clarkson U ABStine@CLVMS.Clarkson.Edu
rcd@ico.isc.com (Dick Dunn) (02/23/90)
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) writes: > One thing that people might wish to consider is that the UNIX market is no > longer limited to scientific/engineering; it is growing quite a bit in the > commercial sector... True...but we're talking about workstations, aren't we? It's going to be a little longer before the commercial world is big on workstations in this class. It will happen, but keep in perspective what sort of a machine you've got here. > ...Why should they have to pay for a lot of disk and memory if > all that they are doing is running terminals at a video checkout counter? Now, let's stop and think about this a moment, and let's admit that you don't use a machine in the 6000 class to control a few glorified cash registers. > Whether a customer needs more performance and disk space will largely be > controlled by what they are going to use the systems for. Yes, but there is a general tendency to want more storage as you increase the speed (and cost) of the core of the machine. There are extremes--places where you need minimal power and lots of storage or vice-versa. But they're unusual. While I think I could use as much crunch as these machines have, and still do useful work on a 120, I've also seen a fair number of moderate-speed 386es (20 MHz and under, no cache, which puts them at 1/8 or less the claimed speed of the low-end RIOS) with 120 Mb or more well-used. So, it's reasonable for people to be asking whether 120 Mb isn't leaving the base configuration a little too carved-down for such a fast machine. -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...Don't lend your hand to raise no flag atop no ship of fools.
palowoda@fiver.UUCP (Bob Palowoda) (02/23/90)
From article <1028700001@ux1.cso.uiuc.edu>, by shair@ux1.cso.uiuc.edu: > > Alternatively one can purchase Feature number 2828 > (SCSI adapter) for $1,200 and Feature number 2541 > (Select 320MB disk) for $3,000. These two features ^^^^^^^^^^^^^^^^^^^^^ First drive cost 3000.00 > provide one high-performance (12.5ms avg seek, 2MB/sec xfer) > 320MB disk. > > Once one has the SCSI adapter, any number of other interesting > devices can be installed, including (exclusive of cable costs) > Second 320MB disk $5,500 ^^^^^^^^^^^^^^^^^^^^^^ Why is the second drive 2500.00 more than the first? Is that the price of the cable? I'm sure you have to be mistaken. > External CD ROM $1,695 > External QIC tape $1,995 ^^^^^^^^^^^^^^^^^^^ WoW! Does third party MCA QIC tape drives work in in the 6000. > External 8mm tape $6,650 > (but this is really beginning to look like advertising...) I've notice the price for the AIX is 3.0 software is 1250.00 for the "1-2 USER VERSION" how much is the "UNLIMITED" version? And what is the upgrade cost for a "1-2 USER VERSION" ---> "UNLIMITED USER" version? ---Bob -- Bob Palowoda pacbell!indetech!palowoda *Home of Fiver BBS* login: bbs Home {sun|daisy}!ys2!fiver!palowoda (415)-623-8809 1200/2400 Work {sun|pyramid|decwrl}!megatest!palowoda (415)-623-8806 2400/9600/19200 TB Voice: (415)-623-7495 Public access UNIX XBBS
palowoda@fiver.UUCP (Bob Palowoda) (02/23/90)
From article <1566@awdprime.UUCP>, by marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22): > No FORTRAN compiler is supplied with the base OS. FORTRAN, COBOL, Pascal, > and Ada are listed in the back as separate items. How about a "C" compiler? -- Bob Palowoda pacbell!indetech!palowoda *Home of Fiver BBS* login: bbs Home {sun|daisy}!ys2!fiver!palowoda (415)-623-8809 1200/2400 Work {sun|pyramid|decwrl}!megatest!palowoda (415)-623-8806 2400/9600/19200 TB Voice: (415)-623-7495 Public access UNIX XBBS
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) (02/23/90)
In article <1990Feb23.004045.26109@sun.soe.clarkson.edu> abstine@image.soe.clarkson.edu (Arthur Stine) writes: >From article <1596@awdprime.UUCP>, by marc@stingray..austin.ibm.com (Marc J. Stephenson): >> [My ill-fated video counter example] >> >Is this to say that IBM thinks that people who need terminals at video >checkout counters are going to buy RS/6000's? Come on, lets be real. If >you want to be in that market, then fine. But if you want to be in the >tech workstation market, then package the systems for it. The price/performance >of the RS/6000 is nice, but the configuration is not. Throw away the ESDI and >[diskless and SCSI recommendations deleted in the interest of space] > >-- >Art Stine >Sr Network Engineer >Clarkson U >ABStine@CLVMS.Clarkson.Edu Disclaimer: I'm not in marketing and I can't say what market IBM is heading for. I'm a developer well-shielded from any of those decisions, and any marketing directions which I opine about are truly my own. All of the development machines which my little area here use have SCSI and 16MB, various tape drives, and Token-Ring. Most of them are various forms of the 530 box. I like them. <-- Opinion :-) Anyway, there are "oodles" of configurations available. I would hope that we have good solutions at good prices for the various computing situations where workstations compete. Seriously. By the way, the video checkout example (though I may have the facts a little off, maybe they're doing video checkout and quantum mechanics :-)) is a real-life RT example running COBOL. What I meant was that there is a market aside from the techie market. I'm sure that my upper-level management understands that the techie market is extremely important. But once again, I'm no marketeer. Let me escape back to the technical questions, please... As an aside, a lot of these questions have to do with the hardware rather than AIX. The comp.sys.ibm.pc.rt newgroup is being used along with this one for hardware and system information. I would hope that we will soon have enough interest and questions to merit a new newsgroup for this information. Neither current newsgroup seems the proper place for this particular posting. Anybody want to take responsibility, or is it too soon? start Marc Stephenson (marc@stingray.austin.ibm.com) Location: F57/992, (79)3-3796, ZIP 2401, 1C-22/992, Austin, Texas Internal: marc@stingray.austin.ibm.com VNET: MARC at AUSVM6 External: uunet!cs.utexas.edu!ibmaus!auschs!stingray.austin.ibm.com!marc
m1phm02@mfsws6.fed.frb.gov (Patrick H. McAllister) (02/23/90)
A lot of discussion has been going on about the processor performance of these machines, but I would also be interested in knowing something about I/O performance. Does anybody out in netland have any information that can be posted about the disk options? What types of disk, what types of controllers? Access time, transfer rates (in theory and in practice)? I/O performance would be an important consideration around here. Pat Pat McAllister m1phm02@fed.frb.gov (UUCP: uunet!fed!m1edb00 ATT, etc.: (202) 452-2443 snail mail: Federal Reserve Board, Washington, DC 20551)
ANKGC@CUNYVM (Anil Khullar) (02/24/90)
In article <1990Feb23.074559.8135@ico.isc.com>, rcd@ico.isc.com (Dick Dunn) says: > >marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) writes: >> One thing that people might wish to consider is that the UNIX market is no >> longer limited to scientific/engineering; it is growing quite a bit in the >> commercial sector... > >True...but we're talking about workstations, aren't we? It's going to be a >little longer before the commercial world is big on workstations in this >class. It will happen, but keep in perspective what sort of a machine >you've got here. > The commercial world (i.e Trading houses and Comm banking) in this city (NYC) have had workstations (DEC and SUNS) on their desks for years now. I think some of them have already ordered these machines. According to a friend of mine who has a DECStation and works for a large Wall St. trading house, they have ordered a dozen to test out the performance on some of their applications. I guess we'll have to wait and see what the market pronounces in a year's time. >Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 > ...Don't lend your hand to raise no flag atop no ship of fools. Anil anil@eleni.gc.cuny.edu ----------------------------------------------------------------- All Disclaimers apply.
kaiser@cheese.enet.dec.com (02/24/90)
In article <1565@awdprime.UUCP>, ron@woan.austin.ibm.com (Ronald S. Woan/2100000) writes... >Every major manufacturer charges for their own X-Window product to >cover porting (one heck of an optimization job too) and distribution >costs (don't forget full manual set). So far as I know, Digital doesn't charge separately for DECwindows on any DEC platform. (And GNU Emacs is included gratis on the "Unsupported" tape.) ---Pete kaiser@cheese.enet.dec.com +1 508 480 4345 (machine: +1 617 641 3450)
m1phm02@fed.FRB.GOV (Patrick H. McAllister) (02/24/90)
A lot of discussion has been going on about the processor performance of these machines, but I would also be interested in knowing something about I/O performance. Does anybody out in netland have any information that can be posted about the disk options? What types of disk, what types of controllers? Access time, transfer rates (in theory and in practice)? I/O performance would be an important consideration around here. Pat -- Pat McAllister m1phm02@fed.frb.gov (UUCP: uunet!fed!m1edb00 ATT, etc.: (202) 452-2443 snail mail: Federal Reserve Board, Washington, DC 20551) Pat McAllister m1phm02@fed.frb.gov (UUCP: uunet!fed!m1edb00 ATT, etc.: (202) 452-2443 snail mail: Federal Reserve Board, Washington, DC 20551)
ehrlich@cs.psu.edu (Daniel Ehrlich) (02/24/90)
In article <1004@fiver.UUCP> palowoda@fiver.UUCP (Bob Palowoda) writes: Bob> From article <1028700001@ux1.cso.uiuc.edu>, by shair@ux1.cso.uiuc.edu: > > Alternatively one can purchase Feature number 2828 > (SCSI adapter) for $1,200 and Feature number 2541 > (Select 320MB disk) for $3,000. These two features Bob> ^^^^^^^^^^^^^^^^^^^^^ Bob> First drive cost 3000.00 > provide one high-performance (12.5ms avg seek, 2MB/sec xfer) > 320MB disk. > > Once one has the SCSI adapter, any number of other interesting > devices can be installed, including (exclusive of cable costs) > Second 320MB disk $5,500 Bob> ^^^^^^^^^^^^^^^^^^^^^^ Bob> Why is the second drive 2500.00 more than the first? Bob> Is that the price of the cable? I'm sure you have to Bob> be mistaken. Because the first drive is actually the upgrade cost from the 120Mb drive that comes by default. -- Dan Ehrlich <ehrlich@cs.psu.edu> Voice: +1 814 863 1142 FAX: +1 814 865 3176
dmcanzi@watserv1.waterloo.edu (David Canzi) (02/24/90)
In article <1004@fiver.UUCP> palowoda@fiver.UUCP (Bob Palowoda) writes:
)> Alternatively one can purchase Feature number 2828
)> (SCSI adapter) for $1,200 and Feature number 2541
)> (Select 320MB disk) for $3,000. These two features
) ^^^^^^^^^^^^^^^^^^^^^
) First drive cost 3000.00
)
)> provide one high-performance (12.5ms avg seek, 2MB/sec xfer)
)> 320MB disk.
)>
)> Once one has the SCSI adapter, any number of other interesting
)> devices can be installed, including (exclusive of cable costs)
)> Second 320MB disk $5,500
) ^^^^^^^^^^^^^^^^^^^^^^
) Why is the second drive 2500.00 more than the first?
) Is that the price of the cable?
Actually, the $3000 figure is for upgrading from a 120 MB disk to a 320
MB disk. Since the 120 MB disk you give up for this option is worth
$1950, the actual cost of the first 320 MB disk is $4950. So the cable
for the second 320 MB disk is only $550. Quite reasonable.
--
David Canzijason@cs.utexas.edu (Jason Martin Levitt) (02/24/90)
In article <1398@ks.UUCP> drake@ibmarc.uucp (Sam Drake) writes: >In article <1152@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: > >>If it's possible >>to run some kind of minimal system configuration in 120mb, will someone >>please confirm it? > >Sure, fits, no problem. If you have lots of Program Products (compilers, >extra cost applications, etc) you would probably want to NFS mount them, >along with /u. Come on, IBM's not going to run 3 page ads in every >big paper in the US advertising a configuration that won't boot! > [...stuff deleted....] >:-) :-) :-) :-) :-) :-) :-( :-( :-( ;-) You're support and trust of IBM's claims is encouraging. But, I'm interested in some approximate numbers, not unsubstantiated opinions. How much space does AIX 3.1 require? X/Motif? What is the the recommended amount of swap space? [...lots of dead-end material deleted...] ----Jason ----- Jason Martin Levitt P.O. Box 49860 Austin, Texas 78765 (512) 459-0055 Internet: jason@cs.utexas.edu | "The most stupendous system of UUCP : cs.utexas.edu!hackbox!jason | organized robbery known has been that BIX : jlevitt | of the church towards woman." --Gage
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) (02/24/90)
In article <1005@fiver.UUCP> palowoda@fiver.UUCP (Bob Palowoda) writes: >From article <1566@awdprime.UUCP>, by marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22): > >>> No FORTRAN compiler is supplied with the base OS. FORTRAN, COBOL, Pascal, >> and Ada are listed in the back as separate items. > >> How about a "C" compiler? > A C compiler and an assembler come with the base OS. I had hoped that would be clear by the question I was answering. Guess not. :-( Marc Stephenson (marc@stingray.austin.ibm.com) Location: F57/992, (79)3-3796, ZIP 2401, 1C-22/992, Austin, Texas Internal: marc@stingray.austin.ibm.com VNET: MARC at AUSVM6 External: uunet!cs.utexas.edu!ibmaus!auschs!stingray.austin.ibm.com!marc
shawn@jdyx.UUCP (Shawn Hayes) (02/24/90)
According to the documents I've seen AIX 3.0 goes for
$1250-$2000 for 1-2 users
$1250-$5000 for 3-32 users
$2000-$1000 for >32 users
The price range for each group of users reflects the group that the customer
falls into. I guess this refers to the type of institution(educational,
or commercial)
Shawn Hayesbill@bilver.UUCP (Bill Vermillion) (02/25/90)
In article <WINDLEY.90Feb21094048@cheetah.cheetah.ucdavis.edu-> windley@cheetah.ucdavis.edu (Phil Windley/20000000) writes: -> ->In article <5098@brazos.Rice.edu> schafer@brazos.rice.edu (Richard A. Schafer) writes: -> -> I just signed on IBMLINK and ran a configuration on the standard grayscale -> configuration. I don't come up with the $12,995 price, but did produce the -> following configuration: -> -> Powerstation 320 $7475 -> 120MB disk $1950 -> Grayscale adapter $1395 -> keyboard $ 255 -> mouse $ 130 -> mono display $1295 -> ------ Hardware total, $12,500 -> AIX 3.0 $1250 -> AIX Xwindows $ 500 -> ------ Software total, $ 1,750 -> Package total, $14,250 -> -> -> ->An IBM ad in the San Francisco Chronicle yesterday quoted the $12995 price ->for a 320 with gretscale monitor, 8Mb RAM, 120 Mb disk, ethrenet, AIX, and ->AIX Windows. I don't know what kind of pricing arrangement it is, but ->apparently, they're selling a "complete" machine for that price. In the product description of the 320 under the heading of Machine Requirements arr: 1. IBM Supported ASCII terminal or IBM supported display and keyboard. 2. AIX Version 3 for RISC System/6000 (5706-088) That implies (or is it infers) that you can have a fully operation system for the $7475 base + $1250 for AIX + $659 (IBM's most expensive of the ASCII line). That's $9384 for as stripped as you can get. Makes it cheaper than their Model 80 line running Unix or Xenix. In their 6000 overview publication (190-019) I find a few items of interest. These are direct quotes. "BIM intends to provide a FDDI adapater for the RISC System/6000 family. THis interface will allow a RISC System/6000 to attach to a FDDI 100 Mbps optical LAN." "IBM intends to prove a Serial Optical Channel ro RISC System/6000 high speed, inter-processor communications and attachment to future high-bandwidth devices." -- Bill Vermillion - UUCP: uunet!tarpit!bilver!bill : bill@bilver.UUCP
drake@sd2.almaden.ibm.com (Sam Drake) (02/25/90)
In article <1162@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >:-( :-( :-( ;-) You're support and trust of IBM's claims is encouraging. >But, I'm interested in some approximate numbers, not unsubstantiated >opinions. How much space does AIX 3.1 require? X/Motif? What is the >the recommended amount of swap space? My earlier reassurances that 120MB systems worked WERE substantiated. Merriam gives "to establish by competent evidence" as a definition for "substantiated". I have personally run a 320 with 120MB of disk space in use (using pre-release software). It would be unfair for me to comment on exact sizes of things (since I've only seen pre-release code, any more exact sizes I might give you would be somewhat off). And clearly any "recommended amount of swap space" would depend highly on the applications a given machine was running. But rest assured that first-hand evidence shows that a 320 with 120MB of DASD is a bootable, operable, configuration. Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
jason@cs.utexas.edu (Jason Martin Levitt) (02/27/90)
In article <1418@ks.UUCP> drake@ibmarc.uucp (Sam Drake) writes: >In article <1162@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >>:-( :-( :-( ;-) You're support and trust of IBM's claims is encouraging. >>But, I'm interested in some approximate numbers, not unsubstantiated >>opinions. How much space does AIX 3.1 require? X/Motif? What is the >>the recommended amount of swap space? > >It would be unfair for me to comment on exact sizes of things (since I've only >seen pre-release code, any more exact sizes I might give you would be >somewhat off). That's the temporary answer to my question. Thanks. ----- Jason Martin Levitt P.O. Box 49860 Austin, Texas 78765 (512) 459-0055 Internet: jason@cs.utexas.edu | "The most effective debugging tool is UUCP : cs.utexas.edu!hackbox!jason | still careful thought, coupled with BIX : jlevitt | judiciously placed print statements." | -Brian Kernighan [1978]
nghiem@ut-emx.UUCP (Alex Nghiem) (02/27/90)
In article <1167@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >In article <1418@ks.UUCP> drake@ibmarc.uucp (Sam Drake) writes: >> >>It would be unfair for me to comment on exact sizes of things (since I've only >>seen pre-release code, any more exact sizes I might give you would be >>somewhat off). > > That's the temporary answer to my question. Thanks. > Usually, how "off" are the release sizes from the actual sizes? On another issue: Are there any academic programs available for these new machines? How can an academic entity "upgrade" to a new machine with out losing the scarce funds that were invested keeping the old machine up to date? How does one integrate the new and the old in a cluster if the new operating system is not available on the old machine? __________________________________________________________________________ This article is posted for factual information only. Any misrepresentation, if any, is purely unintentional. Any opinion expressed or implicit in these remarks are solely my own. nghiem@emx.utexas.edu !cs.utexas.edu!ut-emx!nghiem nghiem@walt.cc.utexas.edu !cs.utexas.edu!ut-emx!walt!nghiem
pcg@rupert.cs.aber.ac.uk (Piercarlo Grandi) (03/03/90)
In article <51507@sgi.sgi.com> markb@denali.sgi.com (Mark Bradley) writes: In article <1660@aber-cs.UUCP>, pcg@aber-cs.UUCP (Piercarlo Grandi) writes: > Pah. The bottleneck is the filesystem, unless you do asynch io via a raw > device. You cannot get more than 600KB per second out of the filesystem in > the best of circumstances, and even that is only achieved, as far I know, by > the MIPS UNIX. Others top out at around 300KB per second. [ ... ] Andrew Koenig in another message reports that some 88K Tek machine also gets up to 600KB second (cheers!) and some other machines get to 450KB/sec. Actually, even some SUNs get to that mark. Your average workstation will only do 150-200KB/sec. (which is horrid, considering that I get that much out of a System V filesystem structure, when clean, on my home 386 with an RLL controller), and at most around 300KB. > If your only worry is single task fast transfer rate (signal/image > processing), be prepared to implement something like the Amoeba or Dartmouth > or Cray file systems. > > The problem is software, not hardware. Pah, indeed. I am measuring >6 MB/sec. through our filesystem today, abeit not with SCSI. Our SCSI (synchronous) is only a bit over 2 MB/sec. on a single drive. Striping and other wonderful *software* things do much more. ---Through the filesystem, mind you.--- Oh yeah. Thanks for supporting my contention/complaint. Now, if only other people took heed from the likes of you and reimplemented the filesystem software. The FFS paper is very clear about what are the limits of the BSD design (But I think it has others, actually). I would not go as far as the Amoeba filesystem (files as *contiguous* lumps of disc space, transferred in one IO operation from disc to memory or viceversa, damn external fragmentation, and this works because Unix files are on average minuscule). The Dartmouth flexible extent based filesystem with daemonic compaction looks good enough to me. And ESDI is much, much faster in certain applications in that one can better sort, queue and optimize the performance that is limited by the speed of the drives' mechnaisms. This is the notorious problem that SCSI will hide from the OS the drive geometry (down to sector remapping, wich can be really nasty), which of course pays put to many nice optimizations. On the other hand, ESDI (on PCs, where it is most popular) has the non trivial problem that controllers are on average not multithreaded. Pah again. -- Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcvax!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk
madd@world.std.com (jim frost) (03/09/90)
schafer@brazos.rice.edu (Richard A. Schafer) writes: >my calculations seem to suggest that the >(much?)-less-powerful SparcStation 1 actually costs a little bit *more* >then the equivalently configured PowerStation 320. Yep, but my sparcstation 1 worked right out of the box and hasn't crashed yet. Our 6000 took about a month of IBM technical support help to get running at all and we've already had to do an OS upgrade (which fixed a lot of stuff -- ksh no longer seems to crash the thing -- but there's a lot to be fixed). So it comes down to this: I can get my compiles done in twice the time all the time or half the time every couple of weeks. Which do you prefer? jim frost jimf@saber.com [Standard disclaimers]
madd@world.std.com (jim frost) (03/09/90)
drake@sd2.almaden.ibm.com (Sam Drake) writes: >In article <1152@gort.cs.utexas.edu> jason@cs.utexas.edu (Jason Martin Levitt) writes: >> But that's not the point. The point is that this is definitly >>*not* AIX 2.2.1, it's AIX 3.1. AIX 3.1 is a big, hairy, hulking, monster >>of an operating system. Here I've got to stand up for IBM a little. SunOS is a bug, hairy, hulking monster of an operating system. And I *like* SunOS for the most part. My only gripe about AIX 3.0 is that in my experience it doesn't stay up during simple usage; I haven't been able to *try* anything complicated yet. jim frost jimf@saber.com [standard disclaimers]
madd@world.std.com (jim frost) (03/09/90)
pcg@rupert.cs.aber.ac.uk (Piercarlo Grandi) writes: >This is the notorious problem that SCSI will hide from the OS the drive >geometry (down to sector remapping, wich can be really nasty), which of >course pays put to many nice optimizations. It's not all that difficult to determine the geometry of a SCSI drive. During the last USENIX a BSD person who's been researching FS optimizations which take into account rotational latency hinted that the BSD people have done so. What you do with SCSI, then, is run a geometry analyzer which dumps out the configuration of the SCSI drive so that the filesystem can do the apropriate optimizations. No big trick there. jim frost saber software jimf@saber.com
drake@sd2.almaden.ibm.com (Sam Drake) (03/10/90)
In article <1990Mar9.015943.4351@world.std.com> madd@world.std.com (jim frost) writes: >Yep, but my sparcstation 1 worked right out of the box and hasn't >crashed yet. Hey, that one's hardly fair ... if you have a 6000 at all, today, you by definition have a pre-release, early-ship machine with pre-release, early-ship, not-done-by-definition software. Complaining very publically because it's not as stable as another company's shipped production level system isn't exactly kosher...? Opinions are my own. Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
mjacob@wonky.Sun.COM (Matt Jacob) (03/10/90)
In article <1990Mar9.022931.4674@world.std.com> madd@world.std.com (jim frost) writes: >pcg@rupert.cs.aber.ac.uk (Piercarlo Grandi) writes: >>This is the notorious problem that SCSI will hide from the OS the drive >>geometry (down to sector remapping, wich can be really nasty), which of >>course pays put to many nice optimizations. > >It's not all that difficult to determine the geometry of a SCSI drive. >During the last USENIX a BSD person who's been researching FS >optimizations which take into account rotational latency hinted that >the BSD people have done so. What you do with SCSI, then, is run a >geometry analyzer which dumps out the configuration of the SCSI drive >so that the filesystem can do the apropriate optimizations. No big >trick there. > >jim frost >saber software >jimf@saber.com Umm- it would be interesting to see whether they claim to be able to make sense out of variable geometry (bit-zone recording method) drives. My own personal opinion is that geometry based filesystems are getting to be a bad microoptimization. With the coming of SCSI-2 multiple command targets, it seems to me that one should just concentrate on getting requests out to the target as quickly as possible and let the microprocessor on the drive figure out the best order do them in. -matt jacob
abstine@image.soe.clarkson.edu (Arthur Stine) (03/11/90)
From article <1443@ks.UUCP>, by drake@sd2.almaden.ibm.com (Sam Drake): > In article <1990Mar9.015943.4351@world.std.com> madd@world.std.com (jim frost) writes: > >>Yep, but my sparcstation 1 worked right out of the box and hasn't >>crashed yet. > > Hey, that one's hardly fair ... if you have a 6000 at all, today, you > by definition have a pre-release, early-ship machine with pre-release, > early-ship, not-done-by-definition software. Complaining very publically > because it's not as stable as another company's shipped production level > system isn't exactly kosher...? > Well, then why did IBM announce it sooo early, if they aren't prepared to ship 'production' quality machines? Advice: if your product isn't ready then don't announce it and ship some 'pre-release' machines and then turn around and complain when users criticize it for not working right. It seems like a typical IBM tactic to pre-annonounce their machines in order to try to get some market share. Well, in the workstation market, I would predict that if a customer is looking at some machines, and he looks at DEC, Sun, HP, and IBM, the IBM wouldn't be the one he would pick right now because the machines that he can actually get his hands on are 'flaky'. Making bold promises like 'oh, it is a pre-release machine. it will be much better when it is ready', don't really hold water. If your machine isn't ready to ship NOW, then don't announce it. Data General did the same thing. The DG machine machine I saw shortly after they brought them out was flaky. All the IBM RS/6000's I've heard about are flaky. First impressions are VERY lasting impressions. If IBM wanted to make a big splash in the workstation market, they should do it with something that floats once it hits, not something that sinks once it hits and has to be held up with a life preserver until it can float... -- Art Stine Sr Network Engineer Clarkson U ABStine@CLVMS.Clarkson.Edu
rcd@ico.isc.com (Dick Dunn) (03/11/90)
madd@world.std.com (jim frost) writes: ...[deleted: someone else's shot at AIX] > Here I've got to stand up for IBM a little. SunOS is a bug, hairy, > hulking monster of an operating system... (I particularly enjoy the typo big->bug...or, hmmm, WAS it a typo?) >...And I *like* SunOS for the > most part. My only gripe about AIX 3.0 is that in my experience it > doesn't stay up during simple usage; I haven't been able to *try* > anything complicated yet. OK, if I interpret this aright, AIX 3.0 is found to be no more b{iu}g, hairy, and hulking than SunOS. I'm not sure that's much of a compliment, Jim! (As an aside to the rest of you: Do you really think humongous kernels are OK, or at least not a problem to worry about? If so, have you heard about the problems Sun had when things got so big they wouldn't fit on an older machine--I think it was the 3/50?) Or, perhaps Jim didn't really mean a damning-by-faint-praise comparison with SunOS on size...but that leaves us with the mere complaint that his system won't stay up long enough to do much! (Jim - I don't mean to make fun of you, but where *were* you standing up for IBM? I really don't think "not much worse than SunOS" is all that kind!) -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...Relax...don't worry...have a homebrew.
shawn@jdyx.UUCP (Shawn Hayes) (03/11/90)
In article <1990Mar10.174015.16644@sun.soe.clarkson.edu> abstine@image.soe.clarkson.edu (Arthur Stine) writes: >From article <1443@ks.UUCP>, by drake@sd2.almaden.ibm.com (Sam Drake): >> In article <1990Mar9.015943.4351@world.std.com> madd@world.std.com (jim frost) writes: >> >>>Yep, but my sparcstation 1 worked right out of the box and hasn't >>>crashed yet. >> >> Hey, that one's hardly fair ... if you have a 6000 at all, today, you >> > It seems >like a typical IBM tactic to pre-annonounce their machines in order to try >to get some market share. Well, in the workstation market, I would predict that >if a customer is looking at some machines, and he looks at DEC, Sun, HP, and >IBM, the IBM wouldn't be the one he would pick right now because the >machines that he can actually get his hands on are 'flaky'. Making bold >promises like 'oh, it is a pre-release machine. it will be much better when > > >-- >Art Stine >Sr Network Engineer >Clarkson U >ABStine@CLVMS.Clarkson.Edu You would be wrong in at least one case if you think people won't buy a pre-release machine. My department is looking very hard at the RS 6000 for our next generation product. If we didn't have a machine to look at now, it would be very unlikely that we would use it. However, since IBM did as many companies do, and let us see a pre-release version, we can make some hard decisions based on the performance that we see now, and the price that they've listed. This has been standard practice with HP and other companies. After all if you're going to spend enough money any company would be foolish not to let you see the latest hardware, even if it isn't completely finished. I think we're all smart enough to understand that bugs may still exist if the product is over a month from going on sale. NOTICE!! It's not available for purchase until April 29!! IBM is giving people a chance to see the machine to generate interest, but it also helps the people who need a machine with the power/price ratio that the RS 6000 family represents. (Yes, I am pleased with what I've seen so far. I understand the beta-test concept) Shawn Hayes
pete@romed.UUCP (Pete Rourke) (03/12/90)
In article <1443@ks.UUCP> drake@ibmarc.uucp (Sam Drake) writes: >In article <1990Mar9.015943.4351@world.std.com> madd@world.std.com (jim frost) writes: > >>Yep, but my sparcstation 1 worked right out of the box and hasn't >>crashed yet. > >Hey, that one's hardly fair ... if you have a 6000 at all, today, you >by definition have a pre-release, early-ship machine with pre-release, >early-ship, not-done-by-definition software. Complaining very publically >because it's not as stable as another company's shipped production level >system isn't exactly kosher...? Hear Hear!! -- Pete Rourke ..texbell!romed!pete MicroAge Technical Marketing Group ..sun!sunburn!romed!pete Tempe, AZ ..ucsd!romed!pete (602) 968-3168 Ext 2273 ..asuvax!romed!pete
hedrick@athos.rutgers.edu (Charles Hedrick) (03/12/90)
>My own personal opinion is that geometry based filesystems are >getting to be a bad microoptimization. You might want to separate issues of placement and command sorting. If the disk controller is prepared to reorder transactions, and does so well, then I agree it's a mistake for the kernel to do so. It should just get transfer requests to the controller as quickly as possible. The controller is in a better position to know what the heads are doing. However it probably still makes sense for the kernel to try to place blocks of files in positions that require minimal effort to read. I don't know of any controllers that are prepared to take over management of the file system. (In fact even the capability to reorder transactions doesn't seem to be present in most SCSI controllers that are actually available.) It's not clear how much this requires the kernel to know about the disk geometry. My suspicion is that the standard BSD file placement code gains something even if it doesn't know where the exact track boundaries are. At least it will tend to keep files reasonably compact. This assumes that SCSI controllers will map logical to physical addresses in a monotonic fashion, even if it can't be exactly linear. (Apparently some do a better job of this than others. I take this into account when buying disk drives.)
drake@sd2.almaden.ibm.com (Sam Drake) (03/12/90)
In article <1990Mar10.174015.16644@sun.soe.clarkson.edu> abstine@image.soe.clarkson.edu (Arthur Stine) writes: >Well, then why did IBM announce it sooo early, if they aren't prepared to >ship 'production' quality machines? Advice: if your product isn't ready >then don't announce it and ship some 'pre-release' machines and then turn >around and complain when users criticize it for not working right. The RISC System/6000 machines were announced on 15 Feb, with availability dates generally listed as second quarter 1990. I think if you look at what most vendors of computer hardware are doing today, you'll see that a lag time between announcement and availability of only 1 quarter (3 months or so) is not "sooo early". Many products are being announced by many companies many Years before they actually ship. Your sentiment is well taken ... but I think you're aiming at an innocent party. Every customer that has a '6000 today got it knowing full well that it was an early ship machine. If your organization didn't have a relationship with IBM, you wouldn't have the machine at all today; you'd be waiting until 2Q90 along with everyone else. If your organization wanted a fully tested machine, it should have said, "no thanks, we'll wait for the 2Q90 date". Since your organization has a 6000, I assume you agreed to take a pre-release machine; and so I really still think it's unfair for folks to be saying, in effect, "hey, the early ship machine I asked IBM to send me isn't fully tested". Naturally it is not, and both you and IBM knew that when the machine was delivered. To complain about it now is just not right. The bottom line is that pre-release software and hardware ... by definition ... has bugs. Every pre-release machine by every manufacturer has had bugs. This says NOTHING about the quality of the final, generally-available product; you can't extrapolate at all about what 2Q90 machines will be like from what early machines are like. If you order and receive a machine after the general availability date in 2Q90, and it has bugs then, by all means flame away. Opinions are entirely my own; I do NOT speak for my employer. Really! Sam Drake / IBM Almaden Research Center Internet: drake@ibm.com BITNET: DRAKE at ALMADEN Usenet: ...!uunet!ibmarc!drake Phone: (408) 927-1861
madd@world.std.com (jim frost) (03/13/90)
mjacob@wonky.Sun.COM (Matt Jacob) writes: >My own personal opinion is that geometry based filesystems are >getting to be a bad microoptimization. With the coming of SCSI-2 >multiple command targets, it seems to me that one should just >concentrate on getting requests out to the target as quickly >as possible and let the microprocessor on the drive figure out >the best order do them in. You'd be wrong no matter what you did. Let's face it, there's no way you can know what someone wants to do with your drive. IBM has had hardware keying in its drives for years and years, and I know people who absolutely swear it's faster than software keying could EVER be. What's the problem with this? For every access technique I've ever seen, there's an optimal and a suboptimal series of requests. There are a number of techniques which boast near-even access times all the time, and a number of them which produce access times which are near optimal for the hardware for specific sequences. None that I've seen can offer near optimal ALL the time, for ALL given sequences. You'd have to do that to get hardware to perform as well as software would when software can know ahead of time how the accesses are going to be done. The application writer has the ability to examine how accesses are going to be done and optimize the data layout based on that knowledge. The drive manufacturer does not. Therefore, if the software just happens to use the worst-case access sequence, it gets terrible performance. There are a hell of a lot of algorithm books out there which bend over backwards trying to prove that you can't fool all of the people all of the time, which is what you're wishing for if you think the hardware designer can predict what everyone is going to want to do with the hardware. Happy hacking, jim frost saber software jimf@saber.com
madd@world.std.com (jim frost) (03/13/90)
rcd@ico.isc.com (Dick Dunn) writes: >madd@world.std.com (jim frost) writes: >...[deleted: someone else's shot at AIX] >> Here I've got to stand up for IBM a little. SunOS is a bug, hairy, >> hulking monster of an operating system... >(I particularly enjoy the typo big->bug...or, hmmm, WAS it a typo?) Pretty amusing typo, I will admit :-). >>...And I *like* SunOS for the >> most part. >OK, if I interpret this aright, AIX 3.0 is found to be no more b{iu}g, >hairy, and hulking than SunOS. I'm not sure that's much of a compliment, >Jim! (As an aside to the rest of you: Do you really think humongous >kernels are OK, or at least not a problem to worry about? Nope, I hate 'em. If you ask me the kernel should be minimal and everything should be a process. Mach is on the right track, although there are some things that I just don't like about it. Once you've gone and paged the kernel, you're asking for trouble. If it's got to be paged, what's the reason for leaving it in the kernel? Gag. It lacks cleanliness. >(Jim - I don't mean to make fun of you, but where *were* you standing up >for IBM? I really don't think "not much worse than SunOS" is all that >kind!) Nope, I wasn't particularly standing up for IBM, I just don't want people damning IBM for a huge OS and praising Sun when the latter is about the same size, and when Sun has put some pretty interesting things in the kernel to make them more usable than they might be as a user-level process. I don't like the SunOS monolithic kernel, but I do like the fact that the system is very usable. jim frost saber software jimf@saber.com
madd@world.std.com (jim frost) (03/13/90)
shawn@jdyx.UUCP (Shawn Hayes) writes: >I >think we're all smart enough to understand that bugs may still exist if the >product is over a month from going on sale. NOTICE!! It's not available >for purchase until April 29!! Eek. It's *that* late in the beta period? I was unaware. jim
henry@utzoo.uucp (Henry Spencer) (03/14/90)
In article <132788@sun.Eng.Sun.COM> mjacob@sun.UUCP (Matt Jacob) writes: >... With the coming of SCSI-2 >multiple command targets, it seems to me that one should just >concentrate on getting requests out to the target as quickly >as possible and let the microprocessor on the drive figure out >the best order do them in. This is reasonable, provided that (a) one can impose constraints on the ordering to meet filesystem-integrity requirements, and (b) the micro on the drive has enough queue space for (potentially) hundreds of requests. I'm not holding my breath. -- MSDOS, abbrev: Maybe SomeDay | Henry Spencer at U of Toronto Zoology an Operating System. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
michael@xanadu.com (Michael McClary) (03/19/90)
In article <1990Mar13.190317.17846@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: >In article <132788@sun.Eng.Sun.COM> mjacob@sun.UUCP (Matt Jacob) writes: >>... With the coming of SCSI-2 >>multiple command targets, it seems to me that one should just >>concentrate on getting requests out to the target as quickly >>as possible and let the microprocessor on the drive figure out >>the best order do them in. > >This is reasonable, provided that (a) one can impose constraints on the >ordering to meet filesystem-integrity requirements, and (b) the micro >on the drive has enough queue space for (potentially) hundreds of >requests. I'm not holding my breath. I hereby spend a little of the net's bandwith to point out, especially to the authors of drive firmware, that (a) is VERY important. VERY VERY VERY VERY VERY VERY VERY VERY VERY VERY VERY VERY important. A drive that does no write-order optimization whatsoever is usable for a high-reliability database. A drive that buffers and re-orders writes is UNusable UNLESS its write order can be constrained (or neither it nor the computer it is connected to EVER suffer any failures or unexpected shutdowns). The smallest simple-to-implement constraint I know is to be able to tell the drive "Be sure everything you got before >NOW< is written and power-fail safe before writing anything you get after >NOW<." If you can't give me at least that, write the data in the order you got it. The constraint "Be sure everything you got before >NOW< is written and power-fail safe before allowing another operation to start." is sufficient, but causes an unnecessary performance hit for some applications. Thank you for your attention. ========================================================================= I normally have the option of turning opinions expressed in my postings into 1/5 of 1% of the opinion of Xanadu Operating Company. On this issue, my opinion >IS< the opinion of Xanadu Operating Company. =========================================================================
marc@stingray..austin.ibm.com (Marc J. Stephenson/140000;1C-22) (03/21/90)
In article <1990Mar13.025647.19878@world.std.com> madd@world.std.com (jim frost) writes: >shawn@jdyx.UUCP (Shawn Hayes) writes: >>NOTICE!! It's not available for purchase until April 29!! >Eek. It's *that* late in the beta period? I was unaware. >jim As far as I know, no specific date for general availability has been set. At announce time, February 15, I believe that we set GA as second quarter, 1990. No refinement of that date has been announced, to the best of my knowledge. Marc Stephenson (marc@stingray.austin.ibm.com) DISCLAIMER: The content of this posting is independent of official IBM position. External: uunet!cs.utexas.edu!ibmaus!auschs!stingray.austin.ibm.com!marc Internal: marc@stingray.austin.ibm.com VNET: MARC at AUSVM6 T/L: 793-3796