1 INFO-VAX	Tue, 21 Dec 2004	Volume 2004 : Issue 706       Contents:= Re: alpha team and tukwila cancelled ... itanium done for ... = Re: alpha team and tukwila cancelled ... itanium done for ... = Re: alpha team and tukwila cancelled ... itanium done for ... = Re: alpha team and tukwila cancelled ... itanium done for ... ! Re: DS10 on a NETGEAR DHCP Router ! Re: DS10 on a NETGEAR DHCP Router  RE: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ) Re: HP pulls out of IA64 chip development ' Re: Installing and running Audit Server ' Re: Installing and running Audit Server  Re: More on Tru64  Re: More on Tru64  Re: More on Tru64  mti Liberator 2200 Questions9 Re: OT: Deutsche Borse wants to buy London Stock Exchange  Pagefile on a non-system disk ! Re: Pagefile on a non-system disk ! Re: Pagefile on a non-system disk ! Re: Pagefile on a non-system disk ! Re: Pagefile on a non-system disk ! Re: Rack mount hardware standards  Re: reverse address translation  Re: Standard Deviation/ Step By Step Instructions For A Free Photo IPOD  Re: tick, tick, tick Re: tick, tick, tick Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald?   Re: Yet another Inquirer article  F ----------------------------------------------------------------------  % Date: Mon, 20 Dec 2004 17:07:29 -0500 ' From: Dave Froble <davef@tsoft-inc.com> F Subject: Re: alpha team and tukwila cancelled ... itanium done for ..., Message-ID: <41C74D21.8050406@tsoft-inc.com>   bob@instantwhip.com wrote:  C > or its time to hire back the alpha team and restart work on alpha ? > ev8 immediately ... and put ev79 into production immediately! 5 > Or sell vms to someone that will do one of the two!  >  >    It sure as hell isn't HP!    ------------------------------  % Date: Mon, 20 Dec 2004 17:56:46 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> F Subject: Re: alpha team and tukwila cancelled ... itanium done for ..., Message-ID: <41C758A8.E6E6E3F0@teksavvy.com>   Dave Froble wrote:E > > or its time to hire back the alpha team and restart work on alpha A > > ev8 immediately ... and put ev79 into production immediately! 7 > > Or sell vms to someone that will do one of the two!  >  > It sure as hell isn't HP!     L Actually, don't write off HP so fast. When Carly leaves to get her politicalJ appointment job in washington, and Stallard or Winkler take over, they mayN spin off the enterprise business, like was done for HP's true roots (agilent).  H Agilent isn't encoumbered by Hp's relationships with Intel or Microsoft.N A spunned off enterprise business wouldn't be emcoumbered by HP's relationshipN with Intel or Microsoft whsih HP desperately needs in order to compete against% Dell at the low end commodity market.   N And a spunned off entreprise business wouldn't have the same Wall Stree CasisoL evaluations as Hp does. HP is evaluated by its consumer PC market share. SunK is evaluated by its financials and to a lesser extent, server market share.   I A spunned off entreprise business would be much closer to Sun in terms of  visibility and image.   1 HP would then be somewhere between Dell and Sony.    ------------------------------  # Date: Mon, 20 Dec 2004 23:21:59 GMT   From: CJT <abujlehc@prodigy.net>F Subject: Re: alpha team and tukwila cancelled ... itanium done for ...* Message-ID: <41C75E97.3040102@prodigy.net>   JF Mezei wrote:  > Dave Froble wrote: > D >>>or its time to hire back the alpha team and restart work on alpha@ >>>ev8 immediately ... and put ev79 into production immediately!6 >>>Or sell vms to someone that will do one of the two! >> >>It sure as hell isn't HP!  >  >  > N > Actually, don't write off HP so fast. When Carly leaves to get her politicalL > appointment job in washington, and Stallard or Winkler take over, they mayP > spin off the enterprise business, like was done for HP's true roots (agilent). > J > Agilent isn't encoumbered by Hp's relationships with Intel or Microsoft.P > A spunned off enterprise business wouldn't be emcoumbered by HP's relationshipP > with Intel or Microsoft whsih HP desperately needs in order to compete against' > Dell at the low end commodity market.  > P > And a spunned off entreprise business wouldn't have the same Wall Stree CasisoN > evaluations as Hp does. HP is evaluated by its consumer PC market share. SunM > is evaluated by its financials and to a lesser extent, server market share.  > K > A spunned off entreprise business would be much closer to Sun in terms of  > visibility and image.  > 3 > HP would then be somewhere between Dell and Sony.    ... and pigs will fly.   --  D The e-mail address in our reply-to line is reversed in an attempt toC minimize spam.  Our true address is of the form che...@prodigy.net.    ------------------------------  % Date: Mon, 20 Dec 2004 19:59:42 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> F Subject: Re: alpha team and tukwila cancelled ... itanium done for ...9 Message-ID: <5AKxd.7602$Z%3.350253@news20.bellglobal.com>   & <bob@instantwhip.com> wrote in message= news:1103550919.865394.201880@f14g2000cwb.googlegroups.com... & > no, it's time to port vms to amd ... > 9 I OK with this idea. Please see the other current thread:  "Time to revive Emerald?  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Mon, 20 Dec 2004 17:04:48 -0500 ' From: Dave Froble <davef@tsoft-inc.com> * Subject: Re: DS10 on a NETGEAR DHCP Router, Message-ID: <41C74C80.7000502@tsoft-inc.com>   JF Mezei wrote:    > Dave Froble wrote: > P >>>I have a Netopia R9100 router which includes multiple RJ45 ports.  The router >>> Q >>does not (easily) support dial-up access.  I also have a Multi-Tech router that K >>supports dial-up as a backup link.  (I'm always on the backup link.)  :-( M >>So, how can I use the Multi-Tech as a dial-up access device, but force all  N >>traffic through the Netopia?  Is this possible, given that each is connected >> > via  >  >>the integral hub/switch? >> > J > I assume that when you mention "dialup", you mean your site initiating aI > dialup PPP connectioon to an ISP, or some user dialing into your site ?     G What I mean is that I'm one of the 2 people in the world still without  P high-speed access.  (Don't everybody laugh at once, you might cause earthquakes N and such.)  The dial-up to an ISP is my only ISP connection.  If I had DSL or E cable I would use only the Netopia, thus solving some of my problems.     ; > dialup modem--[multitech]---*                +----[host1] 0 >                             |                |= >                             |                +----[host2]   0 > dslmodem--------------------+                |; >                             |                +----[host3] 0 >                             |                |; >                             +---[netopia]----+----[host4]  >  > M > You then work on the netopia to have 2 possible configs: one where it deals K > with your broadband provider, in which case, the netopia will forward all / > packets over to the modem's ethernet address.  > O > Or you configure the netopia to forward all packets over to the multitech, at M > which poinmt the multitech acts as an ISP, gives your netopia some other IP W > address, acts as default gateway to your netopia and initiates the dialup connection.     Q I do have the Multi-Tech as the gateway for the Netopia.  But, the connection is  J via the hub/switch integrated into each product, thus putting both on the @ ethernet network where everything is visable to everything else.  L What I envision is that the rest of the network would NOT be visable to the Q Multi-Tech.  Once I attach to the switch integrated into the Multi-Tech, I don't   see any way to do this.     O > Host1 through 4 don't see the difference since they all talk to netopia only. I > You need to ensure that DNS issues are dealt with (either the multitech N > provides the dialup isp's DNS address upstream to netopia and netopid to theP > host1-4, or you hardcode some DSN server which is accessible from both dsl and > dialup connections.  >  > O > The above config would require you put your hub on the WAN side of netopia to O > link netopia, dslmodem and mulltitech together, and you can use the netopia's C > 4 port built=in switch to connect host1 through 4 to the netopia.  > F > There are other ways to do this, such as putting both dsl and dialupJ > connections on the multitech, and then have the netopia only talk to theK > multitech. However, this introduces some additional latency, since you're ' > going through 2 routers at all times.     N That's what I'm doing now.  However, it does not mask the rest of the network  from the Multi-Tech.    N > Ideally, you need a more sophisticated router that can handle 2 links to theN > internet and automatically switch over to a second one when first one fails. >   B The Multi-Tech does this.  What I lack is 2 links to the ISP.  >-(   ------------------------------  % Date: Mon, 20 Dec 2004 17:50:55 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: DS10 on a NETGEAR DHCP Router, Message-ID: <41C7574A.8245725D@teksavvy.com>   Dave Froble wrote:H > What I mean is that I'm one of the 2 people in the world still without > high-speed access.    L Hey, I lived many years with CMU-IP and a PPP software concucted by WolfgangL Moeller (thanks again!) which I automated using ALLIN1 (yes) to generate theI right config after PPP negotiation to start the CMU stack and then sent a N message to application to tell them to go ahead and connect. So with a router, you're better off than I was.   L (now, of cours,e I have an almost real connection on a fixed IP, router etc,' except it is a pesky PPPoE connection).     R > I do have the Multi-Tech as the gateway for the Netopia.  But, the connection isK > via the hub/switch integrated into each product, thus putting both on the B > ethernet network where everything is visable to everything else.  N Ok, normally, the residential routers have 5 ports: a WAN side port, and 4 lan side ports.   S However, conceptually, it has 6 ports, with the 6th port being internal to the box:   /           +--[switch 5]----(router)-------[wan]            |  (switch)--+--[switch 4]-------           +--[switch 3]-------           +--[switch 2]-------           +--[switch 1]-------  L So, anything you plug into switch 1 through 4 is part of the lan and trafficK *conceptually* is seen by all nodes on ports 1 to 5. However, a swicth is a N smart hub, more akin to a bridge, and the (switch) software only sends packets. between the ports that need to get the packet.  L So, a packet from port 1 gets transmitted to port 2 by the switch, but not 3L or 4 because the switch knows that the destination etherenet address residesN on port 2. This way, port 3 and 4 aren't loaded with traffice the switch knowsH isn't destined to them. This allows much greater throughput on your lan.  P A hub however si a simple repeater and all packets get transmitted to all ports.  M > What I envision is that the rest of the network would NOT be visable to the  > Multi-Tech.     < The Wan port of your multitech goes to the modem/phone line.M You take one of the LAN ports on your multitech and attach it to the WAN port L of your netopia. Your home lan is plugged into the LAN side of your netopia.  J However, if the multitech has a switch (instead of a hub), then the routerL portion of the multitech won't really see the lan to lan traffic, except forI broadcasts. And it will pass across the router section only TCPIP packets  destined to the internet    M The advantage of having the Multitech chained to the Netopia is that the tiny G lan segment between the multitech's LAN port and the Netopia's WAN port G constitutes a true demilitarized zone. You can have the firewall on the J netopia, and you could also connect a machine on that tiny DMZ lan segment2 which would have full access to/from the internet.  K So the Multitech wouldn't need much in terms of filters, you'd put them all  into the netopia.   D > The Multi-Tech does this.  What I lack is 2 links to the ISP.  >-(    1 Oh, you want 2 dialup links up at the same time ?   D This gets very complex.Short of having a full BGP feed from your ISPJ (doubtful, especially on dialup), you'd need to configure manual routes toM varous parts of the internet so that when you want to send a packet to node X K on the internet, the routing tables would determine to which line/router to  send the packet.   ------------------------------    Date: 20 Dec 2004 12:43:01 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) Subject: RE: HP exodus to Intel 3 Message-ID: <O0+xbHllZWTt@eisner.encompasserve.org>   | In article <FD827B33AB0D9C4E92EACEEFEE2BA2FB4E9E84@tayexc19.americas.cpqcorp.net>, "Main, Kerry" <kerry.main@hp.com> writes:J > Yeah right - and the first time a kernel or key driver or security patchG > is released, the clustering software breaks and when cluster software + > breaks, it can be a very nasty situation.  > H > Same issue was found with Digital Windows NT clustering software. WhenG > Microsoft released a kernel or key driver or security patch Customers G > would install it on the very first weekend it was available and guess ? > what - the Digital clustering software often had problems.=20  > F > Bottom line is that unless you own or have control of the OS kernel,F > where new patches can be tested on the cluster software before beingH > released, it is extremely difficult to implement an integrated cluster > sub-system.=20  @ And how does this statement fit in with the strategy to dump the2 TruCluster/ADVFS to HPUX port and go with Veritas?  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  G         Timid men prefer the calm of despotism to the boisterous sea of $         liberty. -- Thomas Jefferson   ------------------------------  % Date: Mon, 20 Dec 2004 14:01:18 -0500 # From: "John Smith" <a@nonymous.com>  Subject: Re: HP exodus to Intel , Message-ID: <-oadnSxcmLMevFrcRVn-vw@igs.net>   Bob Kaplow wrote: E > In article <1103324903.666026.266420@f14g2000cwb.googlegroups.com>, " > ranjit_mathews@yahoo.com writes:C >> The potential new customers are eagerly waiting in the wings for G >> Linux on Xen on VMS on IA64, with VMS clustering APIs passed through H >> to Linux ... so that they can have something equivalent to Linux with >> TruClusters:-)  > G > Hmmm, maybe the future is to get TruCluster and ADVFS open-sourced to  > the  > Linux community...    ) Nope. Why make linux a better competitor?   E The best thing for HP to do is give each Tru64 customer who relies on L TruCluster an equivalent set of VMS licences and layered products for *free*J and offer porting assistance at vastly reduced rates, ie. senior people at
 junior rates.    ------------------------------  % Date: Mon, 20 Dec 2004 16:09:24 -0500 ' From: Dave Froble <davef@tsoft-inc.com>  Subject: Re: HP exodus to Intel , Message-ID: <41C73F84.9000600@tsoft-inc.com>   JF Mezei wrote:    > Rob Young wrote: > F >>        distant future.  But you have to admit it is pretty funny toI >>        go back and read your posts 5 or so years ago talking about how B >>        VMS is going away in 5 years, Compaq will kill it , etc. >> > P > Mr Young, I believe you are perfectly aware of how close VMS came to be killedQ > under Compaq's regime. So the above statement is rather ironic coming from you.   / And there are witnesses who can attest to this!    ------------------------------    Date: 20 Dec 2004 15:39:17 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: HP exodus to Intel 3 Message-ID: <ALpHrjqdYZ4I@eisner.encompasserve.org>   \ In article <fnvds018n2cq76vu6ao5fimup5qokpln18@4ax.com>, Nigel Barker <nigel@hp.com> writes: > K > Clearly they don't teach you enough English history in US High Schools:-)   G    They teach us that we beat the pants of some German English king and I    his German mercinaries, then twice went back and saved your souls from     more Germans.  +    Something else happened in England?  ;-)    ------------------------------  % Date: Mon, 20 Dec 2004 20:26:12 -0500 ( From: Bill Todd <billtodd@metrocast.net> Subject: Re: HP exodus to Intel = Message-ID: <FrSdndwW1Y8j5lrcRVn-hA@metrocastcablevision.com>    Chip Coldwell wrote:) > On Mon, 20 Dec 2004, Bob Koehler wrote:  > G >> In article <1103324903.666026.266420@f14g2000cwb.googlegroups.com>,  # >> ranjit_mathews@yahoo.com writes:  >>J >>> The potential new customers are eagerly waiting in the wings for LinuxK >>> on Xen on VMS on IA64, with VMS clustering APIs passed through to Linux @ >>> ... so that they can have something equivalent to Linux with >>> TruClusters:-) >> >>D >>   I believe IBM already contributed a DLM based cluster to Linux.  H They certainly contributed their clone of the VMS DLM, and they've also I supported other open-source clustering efforts (I seem to remember a set  E of basic services whose name currently escapes me but which may have  C grown into an attempt to create an open-source standard clustering  F toolbox).  But I'm not sure they've actually provided anything like a  full clustering package.   >  > ) > http://www.lustre.org/docs/luswhite.pdf  > J > This document even cites Kirby McCoy's book "VMS File System Internals" E > as a reference describing a DLM based cluster filesystem for Linux.   G Peter Braam, the prime mover behind Lustre, is an unabashed admirer of  I VMS clustering technology.  I don't know whether Lustre's DLM comes from  D the IBM donation or from independent work done by (I think) Stephen F Tweedie, but in either case it's very similar in external functioning.   - bill   ------------------------------  % Date: Mon, 20 Dec 2004 20:49:25 -0500 ( From: Bill Todd <billtodd@metrocast.net> Subject: Re: HP exodus to Intel = Message-ID: <rf-dnZsOtLWyHFrcRVn-pw@metrocastcablevision.com>    Keith Parris wrote:  > JF mezei wrote:  > H >> Well, the number is a spin to help control the speed of the downward 	 >> spiral  >> that IA64 was put in. >  > $ > IA64 isn't in a "downward spiral".  G I'll suggest that this depends upon whether one looks at sales figures  & (such as they are) or at expectations.  H In the latter department, Itanic has been in a downward spiral for most I of the past decade.  Only the fact that it has Intel (and until recently  E HP) solidly behind it kept it from being declared a failure long ago.   $   It's just beginning to really takeL > off. Revenue growth was 60% quarter-over-quarter, and 225% year-over-year.  G I'm afraid that until the *absolute* revenue figures start threatening  C to come at least within distant sight of those of its competition,  G suggesting that Itanic is 'really taking off' is likely to continue to  G be met by gales of derisive laughter.  But you must be used to that by  < now, and it doesn't seem to faze you, so keep spinning away.   > J > IA64 is doing very well on benchmarks. In the latest TOP500 list, there H > are 84 Itanium2 entries compared with 54 for POWER and 30 for Opteron.  I The Top500 list is not exactly what I tend to think of when 'benchmarks'  H are mentioned:  it's more of a status symbol which can be bought by any F half-acceptable system whose vendor is willing to give configurations ? away to large HPC institutions in order to achieve that status.   G In terms of real 'benchmarks', about the best that can be said is that  I (power considerations aside) Itanic usually manages to be competitive in  D small-system configurations (its large-system architectures are not @ competitive in commercial use, as POWER5 has so recently and so  devastatingly demonstrated).   > F > HP saw more than 180 SAP customers move to Itanium in 180 days (see = > http://www.hp.com/hpinfo/newsroom/press/2004/041118a.html).   F Thanks, but I'll pass:  there's only so much HP spin I can stomach in  one day.   > G > I visited some VMS customers in NYC a couple of months ago, and they  K > were eagerly starting to get VMS running on rx2600 and rx4640 systems to  K > start their porting efforts. There's a lot of excitement in the VMS base  ! > right now about VMS on Itanium.  > B > While a lot of low-end systems are selling for porting purposes,  I Hmmm.  That would be virtually all of them now, right?  Since HP gave up  I trying to move Itanic 'workstations'?  If you're using full-year figures  G for Itanic revenue below, it seems likely that overall revenue will be  F taking a significant hit right about now if 45% percent of it came in ' large part from such workstation sales.      that'sB > not all that's selling: HW split by revenue is 45% low-end; 29% E > mid-range; 20% high-end; 6% other (not sure what "other" is: maybe   > Blades?).   ? Surely you jest:  Itanic blades?  True, there are now at least  C relatively low-power (and commensurately low-performance) versions  G available, but I haven't heard of anyone actually *using* one for such  	 purposes.   I Then again, 6% of a small number is an even smaller number, so who knows?    > F > There are 2,880 applications available on Itanium now, up from just G > 1,000 only one year ago. (For VMS, the count is now 250 applications  C > ready for 8.2, with over 780 applications from over 360 partners  1 > committed.) So the momentum is really building.   C Right, Keith.  Close your eyes, click your heels three times, wish  . really, really hard, and surely it will be so.   - bill   ------------------------------  % Date: Mon, 20 Dec 2004 14:02:56 -0500 # From: "John Smith" <a@nonymous.com> 2 Subject: Re: HP pulls out of IA64 chip development, Message-ID: <Wt6dndC9HZN4vFrcRVn-3A@igs.net>   Bob Kaplow wrote: 7 > In article <41C38105.109631A5@teksavvy.com>, JF Mezei ( > <jfmezei.spamnot@teksavvy.com> writes: >> icerq4a@spray.se wrote:H >>> The things that the Alpha team really can contribute with is the newF >>> system architecture. The current Itanium cores are doing just fineA >>> in performance, and Montecito will have the best power saving ' >>> features in the industry next year.  >>F >> The question is whether the Alpha knowledge is really applicable to >> an EPIC architecture. >>D >> A lot of Alpha's performance tricks were anti-EPIC in that it was >> the AlphaF >> logic which automatiaclly parralelized stuff, optimized instruction
 >> pipelining C >> etc etc. With EPIC, that is supposed to be done by the compiler.  > F > The real crown jewel that Intel got from DEC/Compaq in the alphacideC > announcement was the compiler technology and people. Without that 
 > compilerH > technology EPIC would be as useful as a first generation Pentium doing > double precision FP divides.    F In other words, everyone gets rich using DEC technology....except DEC.   ------------------------------  # Date: Mon, 20 Dec 2004 19:35:28 GMT & From: John Reagan <john.reagan@hp.com>2 Subject: Re: HP pulls out of IA64 chip development1 Message-ID: <4QFxd.4792$Px1.544@news.cpqcorp.net>    Bob Kaplow wrote:    > F > The real crown jewel that Intel got from DEC/Compaq in the alphacideL > announcement was the compiler technology and people. Without that compilerH > technology EPIC would be as useful as a first generation Pentium doing > double precision FP divides.  I Uh, what compiler technology?  Sure, I believe that Intel got permission  A to look at or use existing GEM technology, but to the best of my  G knowledge they didn't take and use any of it (I'm guessing they didn't  D want any BLISS code inside their compiler system).  Now some of the I ex-GEM developers may have "reimplemented" algorithms or interfaces from  G GEM inside the Intel compiler system, I don't know that.  The existing  @ Intel code generate generates reasonable code and in many cases 3 generates better code that what GEM currently does.      --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Mon, 20 Dec 2004 15:36:24 -0500 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: HP pulls out of IA64 chip development, Message-ID: <41C737C8.5070901@tsoft-inc.com>   icerq4a@spray.se wrote:    > Dave Froble wrote: > 9 >>It's ego.  Intel could have continued producing Alphas.  >> > @ > No, they could have not, that was and is not a Intel decision.    Q Care to go into some more detail?  My understanding was that Intel got rights to  O all the Alpha technology.  That seems to me that producing Alpha CPUs would be  	 included.   P Now if you're saying that HP wouldn't buy the Alpha CPUs, that's another topic,  and one has to wonder about HP.    Dave   ------------------------------  # Date: Mon, 20 Dec 2004 20:55:38 GMT & From: John Reagan <john.reagan@hp.com>2 Subject: Re: HP pulls out of IA64 chip development2 Message-ID: <e%Gxd.4797$0x1.2514@news.cpqcorp.net>   Dave Froble wrote: > icerq4a@spray.se wrote:  >  >> Dave Froble wrote:  >>; >>> It's ego.  Intel could have continued producing Alphas.  >>>  >>A >> No, they could have not, that was and is not a Intel decision.  >  >  > I > Care to go into some more detail?  My understanding was that Intel got  G > rights to all the Alpha technology.  That seems to me that producing   > Alpha CPUs would be included.  > K > Now if you're saying that HP wouldn't buy the Alpha CPUs, that's another  ( > topic, and one has to wonder about HP. >  > Dave >   F Getting the "rights" could include just the rights for patents or for H using concepts from the architecture or implementations without risk of G lawsuit for patent or copyright infringment, but not include the right  2 to enhance it or make their own versions for sale.   --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Mon, 20 Dec 2004 15:32:03 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> 2 Subject: Re: HP pulls out of IA64 chip development9 Message-ID: <bFGxd.6424$GK5.577917@news20.bellglobal.com>   3 "John Reagan" <john.reagan@hp.com> wrote in message + news:4QFxd.4792$Px1.544@news.cpqcorp.net...  > Bob Kaplow wrote:  >  >>G >> The real crown jewel that Intel got from DEC/Compaq in the alphacide D >> announcement was the compiler technology and people. Without that >> compiler I >> technology EPIC would be as useful as a first generation Pentium doing  >> double precision FP divides.  > J > Uh, what compiler technology?  Sure, I believe that Intel got permissionL > to look at or use existing GEM technology, but to the best of my knowledgeG > they didn't take and use any of it (I'm guessing they didn't want any C > BLISS code inside their compiler system).  Now some of the ex-GEM G > developers may have "reimplemented" algorithms or interfaces from GEM J > inside the Intel compiler system, I don't know that.  The existing IntelL > code generate generates reasonable code and in many cases generates better$ > code that what GEM currently does. >  > --  
 > John Reagan 1 > HP Pascal/{A|I}MACRO for OpenVMS Project Leader  > Hewlett-Packard Company   B You may be partly in error because back on DEC-7-2004 Steve Lionel0 made the following statement in this news group:   <quote> H Many of us have played the "if only..." game, but I've given up on that.L More insightful is the latter section of the book detailing how the diasporaJ of DECcies has instilled parts of the DEC culture in many major players inH the computer industry. Heck, here at Intel, I keep running across formerF DECcies in addition to the ones I work with directly. Don Harbert, forF example, is once again in my management chain! Lots of chip design andJ manufacturing people from DEC have found their way to Intel, AMD and otherI influential players. Closer to home, the Intel Fortran compiler (and to a L lesser extent the C++ compiler) have a sizeable "DEC" component. So, indeed,> DEC is dead as a company, but in another sense it is very much still alive, and this is good. </quote>  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Mon, 20 Dec 2004 16:13:38 -0500 # From: "John Smith" <a@nonymous.com> 2 Subject: Re: HP pulls out of IA64 chip development, Message-ID: <ZPqdnUff1uEb3VrcRVn-ow@igs.net>   Dave Froble wrote: > B > Now if you're saying that HP wouldn't buy the Alpha CPUs, that's0 > another topic, and one has to wonder about HP.    ? We all wonder about HP, daily, and come away shaking our heads.    ------------------------------  # Date: Mon, 20 Dec 2004 21:39:32 GMT & From: John Reagan <john.reagan@hp.com>2 Subject: Re: HP pulls out of IA64 chip development1 Message-ID: <oEHxd.4803$iB1.743@news.cpqcorp.net>    Neil Rieck wrote:    >  > D > You may be partly in error because back on DEC-7-2004 Steve Lionel2 > made the following statement in this news group: > 	 > <quote> J > Many of us have played the "if only..." game, but I've given up on that.N > More insightful is the latter section of the book detailing how the diasporaL > of DECcies has instilled parts of the DEC culture in many major players inJ > the computer industry. Heck, here at Intel, I keep running across formerH > DECcies in addition to the ones I work with directly. Don Harbert, forH > example, is once again in my management chain! Lots of chip design andL > manufacturing people from DEC have found their way to Intel, AMD and otherK > influential players. Closer to home, the Intel Fortran compiler (and to a N > lesser extent the C++ compiler) have a sizeable "DEC" component. So, indeed,@ > DEC is dead as a company, but in another sense it is very much  > still alive, and this is good.
 > </quote> >   H Not to put words in Steve's mouth (I said hello to him in the cafeteria D just today), but reading his quote he seems to be talking about the G culture and former DEC employees.  I do not read that the software has  I 'a sizeable "DEC" component.' in terms of software acquired from Compaq.  I   Again, they have have taken some and adapted it for their use, I don't  @ know.  And in the Fortran case, there may be a larger amount of 7 non-code-generator code that was acquired/adapted/used.    --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Mon, 20 Dec 2004 20:04:45 -0500 ( From: Bill Todd <billtodd@metrocast.net>2 Subject: Re: HP pulls out of IA64 chip development= Message-ID: <2KKdnScWiMsq61rcRVn-sA@metrocastcablevision.com>    icerq4a@spray.se wrote:    ...   C > IA64 has good performance and obviously better than POWER in some  > areas   H Since that's the second time you've claimed this recently, perhaps it's # time to ask you for some specifics.   I The new 1.6 GHz, 9 MB Itanic has managed to open up a noticeable (though  E hardly devastating) lead in single-processor SPECint - but as far as  F I've seen that's about it, and even then only if you're running HP-UX G with their special compiler (SGI's results for the same model are only  H barely faster than POWER5's score).  The new 1.6 GHz Itanics enjoy only I a few percent single-processor lead over POWER5 in SPECfp_base and are a  I bit behind it in SPECfp_peak - tests where streaming memory bandwidth is  I important but raw memory latency is less so, making your claim that much  E of POWER's advantage accrues from its on-chip memory controller less  # persuasive in this particular area.   D And in 4-processor TPC-C, where HP's specially-designed zx1 chipset D should minimize the system-level advantages that POWER so obviously I enjoys in larger systems, POWER5 still maintains a significant lead over  G the new Itanics.  The same seems likely to be true in other commercial  F benchmarks, but I don't have time to check right now (so by all means  correct me if appropriate).t  A Note, of course, that POWER5 manages all this while supporting 2 aE processor cores on each chip, rather than Itanic's one - and without lF resorting to Herculean cooling efforts.  This suggests that POWER5 is H already operating in performance-per-Watt regions similar to those that G Itanic only hopes to achieve next year with the Montecito power-saving lG improvements that you've touted, and may well continue to lead in this  I area when the new IBM/AMD strained-silicon process debuts (likely before o Montecito).P  I No - in reality, Itanic, even when you restrict your observations to its nD core technology and attempt to factor out system-level differences, F still seems to be struggling just to keep within spitting distance of B the industry-leading products that IBM, AMD, and Intel itself are I fielding.  And with the redirection of the Tukwila effort (whatever that iG may entail), its chances of jumping ahead of that competition any time E soon seem slimmer than ever.   - bill   ------------------------------    Date: 20 Dec 2004 10:53:32 -0800( From: "Shawn" <shawnm1964@sbcglobal.net>0 Subject: Re: Installing and running Audit ServerA Message-ID: <1103568812.653804.8790@z14g2000cwz.googlegroups.com>D   Thanks for the FYI     Ken Robinson wrote:P > Shawn wrote (in part): > > Hi All,i > >iF > > I have an AlphaServer which is running OpenVMS 7.2-1.  I have beenG > > asked to investigate the process to install and run Auditing on theiD > > server in order to track who is connecting to the server and the > timesi > > they are connecting. >iE > Security auditing is built in to OpenVMS, you just have to know how- to* > turn it on and analyze what you collect. >-' > Start by reading the documentation at G > <http://h71000.www7.hp.com/doc/732FINAL/aa-q2hlg-te/aa-q2hlg-te.HTMl>  > Ken Robinson   ------------------------------  % Date: Mon, 20 Dec 2004 17:08:51 -0500>- From: JF Mezei <jfmezei.spamnot@teksavvy.com> 0 Subject: Re: Installing and running Audit Server, Message-ID: <41C74D70.A0F4AA98@teksavvy.com>   Shawn wrote: > 	 > Hi All,s > D > I have an AlphaServer which is running OpenVMS 7.2-1.  I have beenE > asked to investigate the process to install and run Auditing on the4H > server in order to track who is connecting to the server and the times > they are connecting.  J You may also use ACCOUNTING instead of auditing. This is perhaps closer to what you need.  T there is a VMS accounting utility manual I believe. and HELP SET ACCOUNTING as well.  N However, the acocunting utility will only track named processes. For instance,F someone connecting to the web server to fetch a page won't trigger anyN accounting since it is using a process that is already there (the web server).   ------------------------------  % Date: Mon, 20 Dec 2004 19:06:35 -0500y( From: Bill Todd <billtodd@metrocast.net> Subject: Re: More on Tru64= Message-ID: <CqOdnQsWDZKK9FrcRVn-gA@metrocastcablevision.com>p   Bob Willard wrote:   ...   M > Mind your manners, GS.  After all, this NG is devoted to software from DEC,hL > the company which produced the ugliest linker of all time - TKB.  In fact,J > TKB was IMHO one of the major forces driving RSX folks to switch to VMS.  F Now, Bob:  TKB was in fact a thing of beauty when it was needed.  The G main problem was that DEC never created much in the way of simplifying cF wrappers to mask its complexity when said complexity was *not* needed.  C I suppose DEC could have chosen to implement virtual memory on the dI PDP-11 instead, but that would hardly have been suitable for a real-time -G OS like RSX.  With TKB, you could shoe-horn as much code and data into aH 64 KB of process address space (or 128/192 KB if you included I&D space E and/or Supervisor mode) as you might need to, and control how it was zH paged in (or distributed around system physical memory and, optionally, E paged there, in the case of memory-resident libraries or PLAS-mapped n* process memory) at a fine level of detail.  B If another 16-bit OS linker had nearly that degree of flexibility G beneath a much more approachable interface, then by all means point to  I it as a shining example of how things should have been done.  Otherwise, lI just be thankful for all the things that TKB made possible until VAX and o6 VMS matured and became at least relatively affordable.   - bill   ------------------------------  % Date: Mon, 20 Dec 2004 19:46:48 -0500n) From: "Neil Rieck" <n.rieck@sympatico.ca>n Subject: Re: More on Tru649 Message-ID: <%nKxd.7594$Z%3.347981@news20.bellglobal.com>   . "gokrix" <gokrix@hotmail.com> wrote in message< news:1103445300.6176fa529969420bc7278ffc2a2b3496@teranews... > Bob Kaplow wrote:l [...snip...] >sC > Yeah, HPUX is one of the screwier Unix implementations out there.dF > Reputedly the HPUX linker is one of the weirdest on the planet, evenK > weirder than the AIX one, which takes some doing (the AIX linker is weird J > by design but this one is screwy by implementation).  Add those 'just toG > be different from the others' features like .sl extensions for sharedoH > libraries and you got something of a mess.  And then there is the tinyI > matter of having to recompile the kernel each time you add new hardwareoF > because the damn thing does not support loadable driver modules even > now... >RL I remember having to recompile the kernel quite a bit in DUNIX 4.x but don'tL remember doing it in Tru64 5.x (with DUNIX 4.x you always had to have a copyH of the vanilla kernel in the root directory of your boot disk because ifL something happed to the real kernel while it was being recompiled, you could: always come up on the vanilla kernel and rebuild from it).  F Even still, using LSM to manage RAID sets on either DUNIX or Tru64 wasL child's play compared to other operating systems. In Feb of 2004 I watched aJ colleague set up RAID on a Solaris box and I couldn't believe how kludgey  thej process seemed.a  
 Neil Rieck Kitchener/Waterloo/Cambridge,0 Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  # Date: Tue, 21 Dec 2004 01:35:06 GMTr2 From: Bob Willard <BobwBSGS@TrashThis.comcast.net> Subject: Re: More on Tru64- Message-ID: <e5Lxd.535756$wV.90916@attbi_s54>.   Bill Todd wrote: > Bob Willard wrote: >  > ...a > J >> Mind your manners, GS.  After all, this NG is devoted to software from  >> DEC, H >> the company which produced the ugliest linker of all time - TKB.  In  >> fact,K >> TKB was IMHO one of the major forces driving RSX folks to switch to VMS.  >  > H > Now, Bob:  TKB was in fact a thing of beauty when it was needed.  The I > main problem was that DEC never created much in the way of simplifying rH > wrappers to mask its complexity when said complexity was *not* needed. > E > I suppose DEC could have chosen to implement virtual memory on the PK > PDP-11 instead, but that would hardly have been suitable for a real-time iI > OS like RSX.  With TKB, you could shoe-horn as much code and data into  J > 64 KB of process address space (or 128/192 KB if you included I&D space G > and/or Supervisor mode) as you might need to, and control how it was -J > paged in (or distributed around system physical memory and, optionally, G > paged there, in the case of memory-resident libraries or PLAS-mapped  , > process memory) at a fine level of detail. > D > If another 16-bit OS linker had nearly that degree of flexibility I > beneath a much more approachable interface, then by all means point to tK > it as a shining example of how things should have been done.  Otherwise,  K > just be thankful for all the things that TKB made possible until VAX and a8 > VMS matured and became at least relatively affordable. >  > - bill  H I suppose wrappers could have fixed some of the ugliness of TKB - maybe.I The problem with TKB was that, while everything was possible, nothing washH easy.  Any overlay definition required hand-to-hand combat with a really ugly and non-obvious syntax.  H Before joining DEC, I had spent some time using the link editor that wasH part of the Fox-1 software development environment; for most cases, thatI link editor automatically created overlays that worked without any effortcH on the part of the user -- the defaults produced good working .EXEs, andH the linker followed the trail of calls to automatically produce overlaysG that worked.  That was a far different experience from TKB.  And, Fox-1u8 was a non-VM OS on a memory-starved CPU - just like RSX.  I VM had nothing to do with the horrors of TKB, other than VM made overlaysl unnecessary. -- e Cheers, Bob    ------------------------------    Date: 20 Dec 2004 12:41:56 -0800+ From: williams.dan@gmail.com (Dan Williams)m% Subject: mti Liberator 2200 Questions = Message-ID: <26c11a64.0412201241.4eff5432@posting.google.com>   C I am now the proud owner of 7 of these.  I have a few questions forg anyone who has used them.n1 Does anyone know how to program the front panel ? 5 How quick are they compared to standard dssi drives ? E Can I replace the scsi drives in them with bigger ones (ie. 4 or 9gb)l   Thanks   Dant   ------------------------------  % Date: Mon, 20 Dec 2004 15:06:34 -0600 ? From: Swedish Chef <thisisafakeaddress@staffan.tjernstrom.name>dB Subject: Re: OT: Deutsche Borse wants to buy London Stock ExchangeD Message-ID: <pan.2004.12.20.21.06.34.196003@staffan.tjernstrom.name>  M Paul Sture scribbled something like this, on Sun, 19 Dec 2004 15:24:38 +0100:-   > JF Mezei wrote:2 > M >> Does Deutsche Borse use VMS ? Any chance they may influence the LSE's next:- >> technological decisions in favour of VMS ?  > ( > Anyone know what systems Euronext use?  J Deutsche Borse's amalgamated feed is produced using Suns, although some of: the regional exchanges (eg Xetra), runs on OVMS I believe.  D Euronext at least used to run OVMS, not sure if they still do (their? soon-to-be-replaced feed structure certainly is very VMS-like).f  F Bearing in mind that both of these have had major glitches in the lastJ couple of weeks (as anyone who follows the European markets sae), my guessC is that it's the feed systems inate fear of having to deal with thec) exasperatingly cryptic LSE trading rules.t   -- a ---y Swedish Chef+ The only thing I know, is that I don't know    ------------------------------    Date: 20 Dec 2004 11:13:54 -0800, From: "DerekB" <derek.boczenowski@gmail.com>& Subject: Pagefile on a non-system diskB Message-ID: <1103570034.876547.94040@z14g2000cwz.googlegroups.com>  C I have a Alpha running VMS that has a really small system disk, andoF autogen is demanding a much larger pagefile than it can handle. I wantD to create a new pagefile on a different disk, but I'm getting reallyC confused and I'm hoping you can set me straight. Here's the message-' from the SYS$SYSTEM:AGEN$PARAMS.REPORT:   9 PAGEFILE1_SIZE information (for SYS$SYSTEM:PAGEFILE.SYS): 6 PAGEFILE1_SIZE will be modified to hold 2105300 blocks  > ** Note **  Free space on SYS$SYSTEM is insufficient to create a / PAGE file of 2105300 blocks.   The file will bew  extended to hold 1190578 blocks.   ** WARNING ** - Error creating! SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.    ** WARNING ** - Error creating! SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.c  ? ** WARNING ** - Error creating SYS$SYSROOT:[SYSEXE]SYSDUMP.DMP.T  < ** ERROR ** - error creating SYSDUMP.DMP.  SYSDUMP.DMP needs) to be created manually with 453602 blocksh  G Can you give me a quick and dirty set of commands to correctly create aaD pagefile on a different disk than the system disk? Any help would be greatly appreciated! Thanks!O -Derek Boczenowski  St. Mary's Credit Union   ------------------------------  # Date: Mon, 20 Dec 2004 19:27:44 GMT & From: "B Hall" <bill03060@comcast.net>* Subject: Re: Pagefile on a non-system disk. Message-ID: <QIFxd.533248$wV.144565@attbi_s54>  7 "DerekB" <derek.boczenowski@gmail.com> wrote in message < news:1103570034.876547.94040@z14g2000cwz.googlegroups.com...E > I have a Alpha running VMS that has a really small system disk, andsH > autogen is demanding a much larger pagefile than it can handle. I wantF > to create a new pagefile on a different disk, but I'm getting reallyE > confused and I'm hoping you can set me straight. Here's the messagea) > from the SYS$SYSTEM:AGEN$PARAMS.REPORT:z >z; > PAGEFILE1_SIZE information (for SYS$SYSTEM:PAGEFILE.SYS):d8 > PAGEFILE1_SIZE will be modified to hold 2105300 blocks >o@ > ** Note **  Free space on SYS$SYSTEM is insufficient to create > a 1 > PAGE file of 2105300 blocks.   The file will be " > extended to hold 1190578 blocks. >d  > ** WARNING ** - Error creating# > SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.d >u  > ** WARNING ** - Error creating# > SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.g >fA > ** WARNING ** - Error creating SYS$SYSROOT:[SYSEXE]SYSDUMP.DMP.  >d> > ** ERROR ** - error creating SYSDUMP.DMP.  SYSDUMP.DMP needs+ > to be created manually with 453602 blocks  >nI > Can you give me a quick and dirty set of commands to correctly create a F > pagefile on a different disk than the system disk? Any help would be > greatly appreciated!	 > Thanks!m > -Derek Boczenowski >  St. Mary's Credit Union >d $ RUN SYS$SYSTEM:SYSGEN 4 CREATE disk:[dir]pagefile.sys/siz=2105300/contiguous' INSTALL disk:[dir]pagefile.sys/pagefilet  4 You also have to remember to add the SYSGEN  INSTALL< disk:[dir]pagefile.sys/pagefile to your system startup file.  = If you need more information, type HELP when you're in SYSGENt   Bill   ------------------------------  % Date: Mon, 20 Dec 2004 14:58:00 -0500 2 From: "Stanley F. Quayle" <squayle@insight.rr.com>* Subject: Re: Pagefile on a non-system disk/ Message-ID: <41C6E878.30849.3EE8A184@localhost>t  & On 20 Dec 2004 at 11:13, DerekB wrote:E > I have a Alpha running VMS that has a really small system disk, and H > autogen is demanding a much larger pagefile than it can handle. I wantF > to create a new pagefile on a different disk, but I'm getting really2 > confused and I'm hoping you can set me straight.  6 1.	Put the following line in SYS$SYSTEM:MODPARAMS.DAT:   		PAGEFILE=0  4 	This will prevent the pagefile from being adjusted.    . 2.	Create a new pagefile "by hand" as follows:   		$ MCR SYSGEN$ 		CREATE <filespec> /SIZE=<whatever>  A 	You can also specify /CONTIG, if continguous space is available.f    @ 3.	In your SYSTARTUP start, install the new pagefile as follows:   		$ MCR SYSGEN 		$ DECK, 		  INSTALL <directory>PAGEFILE.SYS/PAGEFILE 		$ EOD   # 	You can also do this on-the-fly...             Here's the messagee) > from the SYS$SYSTEM:AGEN$PARAMS.REPORT:- > ; > PAGEFILE1_SIZE information (for SYS$SYSTEM:PAGEFILE.SYS):J8 > PAGEFILE1_SIZE will be modified to hold 2105300 blocks > @ > ** Note **  Free space on SYS$SYSTEM is insufficient to create > a 1 > PAGE file of 2105300 blocks.   The file will beo" > extended to hold 1190578 blocks. >   > ** WARNING ** - Error creating# > SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.  >   > ** WARNING ** - Error creating# > SYS$SYSROOT:[SYSEXE]PAGEFILE.SYS.  > A > ** WARNING ** - Error creating SYS$SYSROOT:[SYSEXE]SYSDUMP.DMP.  > > > ** ERROR ** - error creating SYSDUMP.DMP.  SYSDUMP.DMP needs+ > to be created manually with 453602 blocksh > G > Can you give me a quick and dirty set of commands to correctly createhH > a pagefile on a different disk than the system disk? Any help would be1 > greatly appreciated! Thanks! -Derek Boczenowskid >  St. Mary's Credit Union >   
 --Stan Quaylee Quayle Consulting Inc.  
 ----------- Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363-3 8572 North Spring Ct., Pickerington, OH  43147  USA 0 stan-at-stanq-dot-com       http://www.stanq.com   ------------------------------  + Date: Mon, 20 Dec 2004 22:19:23 +0000 (UTC)i7 From: moroney@world.std.spaamtrap.com (Michael Moroney)u* Subject: Re: Pagefile on a non-system disk( Message-ID: <cq7j5b$53q$1@pcls4.std.com>  4 "Stanley F. Quayle" <squayle@insight.rr.com> writes:  A >3.	In your SYSTARTUP start, install the new pagefile as follows:g  F Best to put it in SYS$MANAGER:SYPAGSWPFILES.COM, which exists for thatD very purpose, and is executed early in the startup.  Don't forget toD put a $ MOUNT command for the drive in the file since it is executed before SYSTARTUP.- -- - -Mikee   ------------------------------  % Date: Mon, 20 Dec 2004 17:19:11 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: Pagefile on a non-system disk+ Message-ID: <41C74FDB.B596062@teksavvy.com>5  
 B Hall wrote:x > $ RUN SYS$SYSTEM:SYSGENh6 > CREATE disk:[dir]pagefile.sys/siz=2105300/contiguous) > INSTALL disk:[dir]pagefile.sys/pagefile  > 6 > You also have to remember to add the SYSGEN  INSTALL> > disk:[dir]pagefile.sys/pagefile to your system startup file.  , Further to that, in your MODPARAMS.DAT file:   SWAPFILE = 0 PAGEFILE = 0 DUMPFILE = 0  = This prevents AUTOGEN from messing with your swap/page files.1  J SHOW MEMORY on a live system gives you a good idea of yor page file usage.  I I have a very small page and swap files on the system drive (I think 5000p4 blocks each), and a huge page file on another drive.   ------------------------------    Date: 20 Dec 2004 12:38:54 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)* Subject: Re: Rack mount hardware standards3 Message-ID: <2RFu9nFRJAsW@eisner.encompasserve.org>   c In article <VE$YLcH1cMrG@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:iC > Is there a significant trend in rack mount hole standards for thea1 > type of computer gear generally used with VMS ?  > = > I notice BA356 rack mount kits are available in two styles.a > F > If I were forced to guess, I would say that RETMA was on the way outJ > in the US and Metric was on the way in, but that is entirely conjecture. > D > If that is true, what is the newest VMS-relevant gear that is onlyC > RETMA-mountable ?   And what is the oldest VMS-relevant gear that  > is only Metric-mountable ?  I IMHO, while metric may have made sense, for DEC it was a flash in the pantJ that never took off. The only metric rack I can recall is the SW800 seriesL of cabinets. And it never caught on with other vendors. Since the Compaq andJ now HP mergers, I don't recall seeing any new metric racks, everything has< been RETMA. Likewise for the stuff that mounts in the racks.  F It all goes back to Jimmy Carter. The USA had been slowly heading fromJ English to Metric. But with Jimmy being an environmentalist, he decided weJ needed to clean up the country. He had signs placed all across the countryG to accomplish this goal. Alas, as a result of a typo, they said "Do Not  Liter".t   :-)R  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"e& 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdftL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  G         Timid men prefer the calm of despotism to the boisterous sea oft$         liberty. -- Thomas Jefferson   ------------------------------    Date: 20 Dec 2004 15:45:25 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) ( Subject: Re: reverse address translation3 Message-ID: <dgy+DtSiU1$x@eisner.encompasserve.org>s  v In article <Pine.LNX.4.61.0412201223360.17418@frank.harvard.edu>, Chip Coldwell <coldwell@physics.harvard.edu> writes: > K > Really, I phrased that question very badly.  What I've been trying to do iM > is to take a peek inside the memory set aside for console use.  I can find lM > out which physical pagess it's using with the $GETSYI system service (with  K > the SYI$_PFN_MEMORY_MAP flag).  It turns out to be possible to map those kH > pages into process address space using $CRMPSC (with the SEC$M_PFNMAP K > flag).  I was surprised by that, since the physical pages are flagged to WI > be for use by the console, not openvms.  But nothing prevented me from n > doing it.n  G    Either you have PFNMAP or you don't.  There's no "you can map _some_t    PFNs" privilege.a   ------------------------------  + Date: Mon, 20 Dec 2004 19:35:19 +0000 (UTC)t3 From: "Richard Maher" <maher_rj@hotspamnotmail.com>s Subject: Re: Standard Deviationb/ Message-ID: <cq79hn$hg6$1@titan.btinternet.com>i   Hi,d  . Not sure about VMS RTLs but COBOL has this : -  3          ACOS               LENGTH         ORD-MIN*u9          ANNUITY            LOG            PRESENT-VALUE*s1          ARGCOUNT           LOG10          RANDOM 1          ASIN               LOWER-CASE     RANGE*a.          ATAN               MAX*           REM2          CHAR               MEAN*          REVERSE.          COS                MEDIAN*        SIN/          CURRENT-DATE       MIDRANGE*      SQRTS>          DATE-OF-INTEGER    MIN*           STANDARD-DEVIATION*/          DAY-OF-INTEGER     MOD            SUM*4.          FACTORIAL          NUMVAL         TAN5          INTEGER            NUMVAL-C       UPPER-CASE 4          INTEGER-OF-DATE    ORD            VARIANCE*8          INTEGER-OF-DAY     ORD-MAX*       WHEN-COMPILED          INTEGER-PARTl  I          *Functions which permit a variable number of arguments which may'H          include occurring items subscripted by "ALL" for one or more of          its dimensions.    7 COBOL PROCEDURE_DIVISION miscellaneous_topics Subtopic?    Regards Richard Mahera  9 "Michael D. Ober" <mdo.@.wakeassoc..com> wrote in message ( news:1kDxd.92$Oo1.937@news.uswest.net...J > Is there a function in VMS that will take a vector (or array) of numbers andy > compute the nth STDev? >y	 > Thanks,t > Mike Ober. >e >h   ------------------------------    Date: 20 Dec 2004 19:58:05 -0800& From: "Buck" <srxman2001@peoplepc.com>8 Subject: Step By Step Instructions For A Free Photo IPODC Message-ID: <1103601485.501236.291470@f14g2000cwb.googlegroups.com>b  > This article will give you step-by-step instructions on how toG successfully obtain your free iPod using freeipods.com for as little ascF $1. Many think that this site is a scam or a pyrimid sceam. A scam no,G a pyrimid yes, a sceam no. You do have to put forth some sort of effortl@ to get a free iPod. You must get five friends or relatives to do exactly the following...    / Step 1: Click here to go to the free iPod site.-) http://www.freephotoiPods.com/?r=13187969e: To participate in this great offer you must live in the US    C Step 2: Use an email address at the bottom of the page and choose aCG password. (they will send you email, so please use a secondary account)9    C Step 3: Fill out all required shipping information. This tells them  where to send your iPod.    E Step 4: (Do not exit at this point) This is the part that over 90% of G the people start thinking: "Maybe this isn't for real." They present to>A you 10 different "optional" offers. These are not any of the real>/ offers needed to complete the getting the iPod.e    @ Step 5: Skip all of the previous offers and you get to the referG friends page. You can start sending emails to your friends here or skipi this step and do it later.    F Step 6: Complete an offer. This is where real companies like AOL, RealE Networks and more are advertising through freeIpods.com. This is whatb@ makes them successfull. Choose an offer and purchase or sign up.G (Choose the RealNetworks Real Rapsody. This cost $1...yes $1 dollar fordF 30 days. Cancel anytime. Even if you forget its only $10 a month afterB the first month. For signing up they also give you 5 free songs toG burn. This is the best offer. Nothing is free, but an iPod for a dollarEF and a little work is worth it. Some offers change. There is always oneG offer where you pay next to nothing or just sign up for a 30 day trial.f    @ Step 7: Get your friends to do the same thing. Even if you don't+ complete it you get to download five songs.c    F Note: Completing an offer and not signing up your friends is what theyG are counting on. It is more likely you get intreagued and then give up.yG This doesn't mean that it is a scam. All you must do it finish. If thiseE still sounds to good to be true then go pay $400 for a new iPod. Havea patience and good luck.     1 If you haven't started yet. Get your iPod here!!!e) http://www.freephotoiPods.com/?r=13187969u   ------------------------------    Date: 20 Dec 2004 13:13:14 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)M Subject: Re: tick, tick, tickn3 Message-ID: <Jg2uY2K2vEBQ@eisner.encompasserve.org>I  q In article <gM3W1EHLBon8@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:a_ > In article <41c6ddc5@usenet01.boi.hp.com>, "Guy Peleg" <guy.peleg@remove_this_hp.com> writes:I= >> OpenVMS V8.2 is scheduled to be submitted to manufacturing : >> early-mid Jan 2005, and will hit the streets early Feb. > G >    So the roadmap is out of date.  Well a 1 month slip is better than 
 >    EV8 got.p  G A one month slip is better than releasing before the software is ready..   ------------------------------  % Date: Mon, 20 Dec 2004 16:42:15 -0500n- From: JF Mezei <jfmezei.spamnot@teksavvy.com>d Subject: Re: tick, tick, ticka+ Message-ID: <41C74735.42D8A3B@teksavvy.com>r   Guy Peleg wrote: > < > OpenVMS V8.2 is scheduled to be submitted to manufacturing9 > early-mid Jan 2005, and will hit the streets early Feb.l  J Does this mean that VMS engineers don't get any Christmas vacation so that$ they can tie up all the loose ends ?   ------------------------------  % Date: Mon, 20 Dec 2004 14:00:11 -05004) From: "Neil Rieck" <n.rieck@sympatico.ca>g$ Subject: Re: Time to revive Emerald?9 Message-ID: <3jFxd.6376$GK5.559994@news20.bellglobal.com>n  3 "Roger Ivie" <rivie@ridgenet.net> wrote in message .- news:slrncsdtcv.jf2.rivie@Stench.no.domain...  > One more thing:u >h  M I started off my DEC hardware life with PDP-11/04 then migrated up to larger aF systems including 11/44, 11/70 and 11/84. My VAX carreer started with L VAX-11/730 and 11/70 then migrated up to VAX-8550 and VAX-6430.  All my DEC F training came from the customer classrooms and labs at DEC in Kanata, I Beford, and Maynard where we had to write and run our own diagnostics to aK prove to the instructor that we understood how the hardware worked. What I  $ remeber from that training was this:   Memory Mapping: G Consisted of adding (via an ALU chip) a program's logical address to a kL mapping register which would "translate" to a physical address. The CPU had I mapping registers as did the bus interface (eg. Unibus Map) and this was  I necessary so DMA to any memory location would work. Every OS (virtual or  ) not) employs some kind of memory mapping.o   Swapping/Overlaying:F Although the OS could swap in/out whole processes, or processes could L request the OS to swap in/out program segments (overlays), your application H software could never "see" any more memory than was actually installed. L (BTW, your program had to be build with overlays in mind for this to work). K The OS knew the memory was there and could access it by wrting values into cJ the mapping registers and then translating, but the OS couldn't get there L directly and was limited in the same way as appliction software. So from an L OS point of view, swapping made it seem that more memory was available then I there really was. Application software could never "see" any more memory , than was actually installed.   Virtual Memory:aM In this technique the application s/w didn't need to be built with swappable hL segments (overlays) because the OS saw every application as a collection of G 512 bytes (on the first VAX machines anyway) and would keep track page yJ usage. When another process required memory, the OS would scan the system J looking for pages that hadn't been used for a while. If unused pages were J dirty (had been written to since last loaded in from disk) then they were M first flushed back to disk before the physical RAM was handed to the process bE requiring it. If they were not dirty, the physical memory was simply wJ logically disconnected from one process and assigned to another. This was M called "paging". The cool thing about virtual memory is that the application  K software can now see much more memory than is actually installed (provided aM you have enough  disk space and the processes is allowed to see it) but only s8 up to the address size of the machine (VAX was 32 bits).   DejaVu:aK I thought about my PDP days when I was first exposed to the XMI bus on the nL VAX. The address width of the XMI bus was 40-bits and VMS could only get to L it by first writing values to XMI mapping registers which were then used to B translate a 32-bit logical address into a 40 bit physical address.    
 Neil Rieck Kitchener/Waterloo/Cambridge,o Ontario, Canada.9 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html R   ------------------------------  # Date: Mon, 20 Dec 2004 20:28:56 GMTt% From: Roger Ivie <rivie@ridgenet.net> $ Subject: Re: Time to revive Emerald?3 Message-ID: <slrncsedir.kq6.rivie@Stench.no.domain>t  7 On 2004-12-20, Neil Rieck <n.rieck@sympatico.ca> wrote:p > Memory Mapping:g > Every OS (virtual or n+ > not) employs some kind of memory mapping.    RT-11SJ? CP/M-80?    > Swapping/Overlaying:  G And in the case of OS/8 and RT-11, part of the system is overlayed (thew USR and the device drivers). >w > Virtual Memory:sO > In this technique the application s/w didn't need to be built with swappable mN > segments (overlays) because the OS saw every application as a collection of I > 512 bytes (on the first VAX machines anyway) and would keep track page oL > usage. When another process required memory, the OS would scan the system L > looking for pages that hadn't been used for a while. If unused pages were L > dirty (had been written to since last loaded in from disk) then they were O > first flushed back to disk before the physical RAM was handed to the process DG > requiring it. If they were not dirty, the physical memory was simply  L > logically disconnected from one process and assigned to another. This was O > called "paging". The cool thing about virtual memory is that the application iM > software can now see much more memory than is actually installed (provided eO > you have enough  disk space and the processes is allowed to see it) but only .: > up to the address size of the machine (VAX was 32 bits).  E And that is one of several possible ways to use virtual memory. There, are others.a  A Consider, for example, an embedded diskless system. It might likeaD to use separate virtual address spaces to keep the various processesF from stepping on each others' toes. In this case, the applications seeF less memory than is available on the system (because other portions of? the memory are allocated to other tasks) and no paging is done.nH Essentially, the working set is the entire address space of the process.  	 > DejaVu:dM > I thought about my PDP days when I was first exposed to the XMI bus on the  N > VAX. The address width of the XMI bus was 40-bits and VMS could only get to N > it by first writing values to XMI mapping registers which were then used to D > translate a 32-bit logical address into a 40 bit physical address.  A VAX physical addresses were originally 30 bits; the remaining twouG bits were typically used to encode the size of a transfer. Since bit 29sF was reserved to signal I/O space, a VAX could not have more than 512MB@ of physical memory installed. At some point this was extended toF (where'd I put that ARM? Oh yes, here it is) 34 bits. This means that G EVEN ON A VAX, it is possible to have more physical memory than can be .' supported by the virtual address space.   G You're not going to argue that a VAX with more than 4GB of memory usingt= 34-bit physical addresses isn't doing virtual memory are you?l -- y
 Roger Ivie rivie@ridgenet.net http://anachronda.webhop.org/s -----BEGIN GEEK CODE BLOCK-----a
 Version: 3.12nH GCS/P d- s:+++ a+ C++ UB--(++++) !P L- !E W++ N++ o-- K w O- M+ V+++ PS+? PE++ Y+ PGP t+ 5+ X-- R tv++ b++ DI+++ D+ G e++ h--- r+++ z+++ i ------END GEEK CODE BLOCK------e   ------------------------------  % Date: Mon, 20 Dec 2004 15:42:50 -0500s' From: Dave Froble <davef@tsoft-inc.com>R$ Subject: Re: Time to revive Emerald?, Message-ID: <41C7394A.6000908@tsoft-inc.com>   David J Dachtera wrote:4    ! > Where can we drop off BG & Co.?d     Is he bringing the CASH????m    H >>Where have you been.  For quite a few years Intel has presented a CICSM >>environment using cores that are very far from what you see.  Let me repeataN >>that.  THEY PRESENT AN ENVIRONMENT!  Maybe you should read up on some of the >>technical data on the N-VAX. >> > G > "x86" is typiclaly used to refer to the 80x86 family (JF usualy dropslE > the "x", thus raising confusion between 8/16 bit CPUs and 16/32 bittE > CPUs), the 80186 and its successors up to and including the currenti@ > Pentium and Celeron CPUs. The "environment" is incidental. The. > presentation is what is seen on the outside.    ( You're saying the same thing I'm saying.  Q >>>>To the people building CPUs today, 'NATIVE' no longer has the same meaning assU >>>>when CPUS were a collection of boards, and a MOVL truly was a hardware operation.t >>>> >>>>D >>>Well, yes and no. At some level, even the GEM back-end eventually0 >>>produces code specific to the underlying CPU. >>>l8 >>No!  It produces code specific to the >>ENVIRONMENT<<! >> > 4 > ...which is facilitated by what? (think carefully) >  > I >>I'm not the best to explain this.  I do understand at some level what'siO >>happening.  I'd suggest that you find someone who can explain the concepts sop  >>you can understand the issues. >> > I > The bottom line is this: whether the "environment" runs on a System 3x05F > family, a z80, or even ENIAC is of no specific importance. What *IS*F > important is what is out there in the thousands and millions of rack > spaces around the world. > D > Call it a "CPU", call it an "environment", call it a "George Bush"J > nonsense whatever, call it a "gribbleniff" running at 182 "quibbledunk",? > call it what you will, it still executes a pre-defined set ofn' > instructions. *THAT* is what matters.e    ? Same thing I said, in different, and maybe not the best, words.y   Dave   ------------------------------  % Date: Mon, 20 Dec 2004 15:41:37 -0500k) From: "Neil Rieck" <n.rieck@sympatico.ca> $ Subject: Re: Time to revive Emerald?9 Message-ID: <9OGxd.6447$GK5.580080@news20.bellglobal.com>I  2 "Roger Ivie" <rivie@ridgenet.net> wrote in message- news:slrncsedir.kq6.rivie@Stench.no.domain... 9 > On 2004-12-20, Neil Rieck <n.rieck@sympatico.ca> wrote:e >vI > You're not going to argue that a VAX with more than 4GB of memory usingC? > 34-bit physical addresses isn't doing virtual memory are you?   H No. I think all this started off when someone said that PDP did virtual B addressing which it doesn't. But if anyone quotes an official DEC , publication on this matter then I'll recant.  
 Neil Rieck Kitchener/Waterloo/Cambridge,n Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  # Date: Mon, 20 Dec 2004 21:00:12 GMTe% From: Roger Ivie <rivie@ridgenet.net>o$ Subject: Re: Time to revive Emerald?3 Message-ID: <slrncsefde.kq6.rivie@Stench.no.domain>c  7 On 2004-12-20, Neil Rieck <n.rieck@sympatico.ca> wrote:c > Swapping/Overlaying:H > Although the OS could swap in/out whole processes, or processes could N > request the OS to swap in/out program segments (overlays), your application J > software could never "see" any more memory than was actually installed. N > (BTW, your program had to be build with overlays in mind for this to work).   E The point of overlays is to cram more than is possible into a limitedoF address space. Usually your application could not "see" all the memory+ that was installed, let alone more than it.R  4 Here's the most outrageous example I could think of:  G MP/M-80 is a multi-user version of CP/M-80, capable of supporting up toeH 8 users on systems that approach 512K of memory space. This memory spaceE is managed by bank switching; there's a common chunk of memory at theoG top of the address space, and the lower address regions can be switchedh in and out.e  F However, MP/M-80 doesn't actually know how to switch banks. To do this@ job, it relies on routines in the hardware-specific XIOS (MP/M's equivalent of the BIOS).  D It is *possible*, although probably not useful, to construct an XIOSE that swaps the banked area of memory out to disk whenever the bank is D switched. Each job believes it has the machine to itself (within theF model provided by MP/M). When the OS switches banks, the old bank goes% out to disk and the new bank come in.e  7 Virtual memory without any hardware support whatsoever.e -- n
 Roger Ivie rivie@ridgenet.net http://anachronda.webhop.org/  -----BEGIN GEEK CODE BLOCK-----,
 Version: 3.12nH GCS/P d- s:+++ a+ C++ UB--(++++) !P L- !E W++ N++ o-- K w O- M+ V+++ PS+? PE++ Y+ PGP t+ 5+ X-- R tv++ b++ DI+++ D+ G e++ h--- r+++ z+++ - ------END GEEK CODE BLOCK------o   ------------------------------  % Date: Mon, 20 Dec 2004 16:03:37 -0500 ' From: Dave Froble <davef@tsoft-inc.com>%$ Subject: Re: Time to revive Emerald?, Message-ID: <41C73E29.2050701@tsoft-inc.com>   John Smith wrote:4   > Neil Rieck wrote:7 >  >>Time to revive Emerald?o >> >> > M > If anybody ever wanted to make Alpha 'industry standard' all they'd have tos= > do is give China the unrestricted rights to manufacture it.t > 5 > 1 billion chips before you could say 'Peking Duck'.e >  >  >   5 Since when has China required 'manufacturing rights'?c   ------------------------------  % Date: Mon, 20 Dec 2004 16:11:58 -0500s# From: "John Smith" <a@nonymous.com>a$ Subject: Re: Time to revive Emerald?, Message-ID: <x5SdnaeJeca_3VrcRVn-iQ@igs.net>   Dave Froble wrote: > John Smith wrote:  >e >> Neil Rieck wrote: >> >>> Time to revive Emerald?o >>>p >>>o >>F >> If anybody ever wanted to make Alpha 'industry standard' all they'dF >> have to do is give China the unrestricted rights to manufacture it. >>6 >> 1 billion chips before you could say 'Peking Duck'. >> >> >> >t7 > Since when has China required 'manufacturing rights'?   + Somebody would have to give them the masks.c   ------------------------------  % Date: Mon, 20 Dec 2004 16:37:56 -0500t- From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Time to revive Emerald?, Message-ID: <41C74633.FC75B7A4@teksavvy.com>  ( > > As a normal UNIX, except clustering,@ > > Tru64 was not in any way better than the other UNIX systems.  B Based on what I heard, Tru64 had its strengt in its documentation.   ------------------------------  % Date: Mon, 20 Dec 2004 17:10:21 -0500-) From: "Neil Rieck" <n.rieck@sympatico.ca> $ Subject: Re: Time to revive Emerald?9 Message-ID: <j5Ixd.7408$Z%3.313924@news20.bellglobal.com>4  3 "Roger Ivie" <rivie@ridgenet.net> wrote in message l- news:slrncsefde.kq6.rivie@Stench.no.domain... 9 > On 2004-12-20, Neil Rieck <n.rieck@sympatico.ca> wrote:d [...snip...] > I > MP/M-80 is a multi-user version of CP/M-80, capable of supporting up toTJ > 8 users on systems that approach 512K of memory space. This memory spaceG > is managed by bank switching; there's a common chunk of memory at theuI > top of the address space, and the lower address regions can be switchedh
 > in and out.s >lH > However, MP/M-80 doesn't actually know how to switch banks. To do thisB > job, it relies on routines in the hardware-specific XIOS (MP/M's > equivalent of the BIOS). >eF > It is *possible*, although probably not useful, to construct an XIOSG > that swaps the banked area of memory out to disk whenever the bank isKF > switched. Each job believes it has the machine to itself (within theH > model provided by MP/M). When the OS switches banks, the old bank goes' > out to disk and the new bank come in.y >a9 > Virtual memory without any hardware support whatsoever.e > -- s  K I'm familiar with bank switching and it is not a form of virtual memory. I  I first saw it back in 1977 on my Apple ][ to "map out" the "Integer BASIC  L ROMS" and "map in" the language card which usually contained UCSD Pascal or J anything else you wanted. That product line ended with the Apple //e with K 128KB of memory which was a cool trick since it only had 16 address lines.  J But it would have been next to impossible to have software in a "switched ; out" bank handle a external event like an interrupt or DMA.   I I think we're just arguing semantics here. I still maintain that virtual  L memory is a mapping between RAM and disk to make your system appear to have M more physical memory than it really does. Through a process called "paging", nJ fixed sizes of memory are moved between RAM and disk as determined by the  OS.o  C http://developer.apple.com/documentation/mac/Memory/Memory-152.htmlkD http://www.freescale.com/files/archives/doc/ref_manual/M68000PRM.pdfH http://www.sasktelwebsite.net/jbayko/cpu3.html#Sec3Part4 see section on  68851l8 http://www.sasktelwebsite.net/jbayko/cpuAppendA.html#VAX    
 Neil Rieck Kitchener/Waterloo/Cambridge,M Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Mon, 20 Dec 2004 17:03:22 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ) Subject: Re: Yet another Inquirer articled, Message-ID: <41C74C28.6516D6FD@teksavvy.com>   Karsten Nyblad wrote:e > 6 > The Inquirer continues it attacks on HP and Itanic. + > http://www.theinquirer.net/?article=20304I  N I think that it makes way too much from legal wranglings between HP and Intel.K Once both agree that IA64 isn't viable (which I think was done some time in N early 2004), then it is a question of plotting an exit strategy, and they willM use the contracts to their own personal advantage, hoping the other gets more3 egg on face.  M Intel needs to thread very carefully: Right now, I assume HP and SGI are busyuE negotiating between AMD and Intel on who can provide an 8086 with the2< enterprise features needed to build the very large systems.   M AMD only has one architecture to focus on. Intel has 2. So, by shifting AlphaoL designers onto the 8086, Intel stands a chance to deliver an enterprise 8086L sooner, allowing HP and SGI to migrate to an Intel 8086, versus migrating toJ an ADM 8086 because Intel's efforts were diluted between IA64 and 8086 and- couldn't deliver enterprise 8086 soon enough.a  9 I think that HP learned from the premature Alpha murder.    H HP isn't about to announce the death of IA64 until the 8086's enterpriseK features are sufficiently close to be credible. And it can use its contracteK with Intel to ensure that both continue to talk positively about IA64 untilt the end of 2007.  L But the minute the 8086 has the features needed to build superdomes etc, theF transition will be very quick. For VMS, I expect much quicker than theF Alpah->IA64 migration where work began only after the public was told.   ------------------------------   End of INFO-VAX 2004.706 ************************