1 INFO-VAX	Mon, 01 Jul 2002	Volume 2002 : Issue 359       Contents:' Re: Andrew wan'ts the numbers, here ... ' Re: Andrew wan'ts the numbers, here ... 	 DDL Files 
 Re: DDL Files 
 Re: DDL Files 
 Re: DDL Files  Re: Defcon whitepaper  Re: fab=cc$rms_fab Re: fab=cc$rms_fab$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications... Re: KZTSA and AIT1 Re: KZTSA and AIT1P Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]P Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]P Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]P Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]P Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]( Re: Mozilla and the mysterious downloads, Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ... OpenVMS 7.3 Update 200 Re: OpenVMS 7.3 Update 200 Re: OpenVMS 7.3 Update 200 OT: RSTS Question  Re: parsing >255 Re: parsing >255 Re: parsing >255 Re: parsing >255/ Re: Powerstorm 300/350 cards and OpenGL on OVMS  Restricted Login Terminations  RMU?? 	 Re: RMU?? 	 Re: RMU?? 	 Re: RMU?? " Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign
 Re: TCO study 
 Re: TCO study  Re: Trouble with BACKUP/RECORD VMS IO up to the level of PC  Re: VMS IO up to the level of PC  Re: VMS IO up to the level of PC$ Re: VMS Marketing Volunteers, ctd... Re: when is a typo not a typo? RE: wow   F ----------------------------------------------------------------------  % Date: Mon, 01 Jul 2002 14:38:23 +0100 U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> 0 Subject: Re: Andrew wan'ts the numbers, here ...0 Message-ID: <afpluv$jni$1@new-usenet.uk.sun.com>   Rudolf Wingert wrote:    > Hallo, > ? > Andrew did wrote a lot of number, but did not write, where we = > can find this number to check the conditions. Anybody could > > write a lot os arguments. But I would like to see verifyable> > statements. We do have an AlphaStation ES40 with four 600MHz< > EV6 CPUs and 16GB memory and a SunFire 880 with the 900MHz= > UltraSparc III. Now I can compare performance/price. Sorry, = > but we do not have an application running on both. The only ? > one I could test is I/O. There was a small difference between ? > both. Alpha a little bit faster (under OpenVMS!!!!). Also the C > boot sequence was much faster (I know this is meaningless because A > OpenVMS have to be booted only once). But the big difference is 9 > the price. We did pay the double value for the SunFire.  >     < Humm, the list price of an 8 CPU V880 is ~100,000 with 16 GB8 the list price of a 4 CPU 667 Mhz 16 GB ES40 is ~180,000
 both dollars.   ; I would be interested in what you used to benchmark the I/O : performance of the two systems and what kind of disks they
 both have.   Regards  Andrew Harrison    ------------------------------  # Date: Mon, 01 Jul 2002 17:37:21 GMT * From: "Bill Todd" <billtodd@metrocast.net>0 Subject: Re: Andrew wan'ts the numbers, here ...B Message-ID: <lr0U8.502574$%y.34192866@bin4.nnrp.aus1.giganews.com>  # "Andrew Harrison SUNUK Consultancy" > <andrew_nospam.harrison_remove_this@sun#.com> wrote in message* news:afpluv$jni$1@new-usenet.uk.sun.com... >  >  > Rudolf Wingert wrote:  > 
 > > Hallo, > > A > > Andrew did wrote a lot of number, but did not write, where we ? > > can find this number to check the conditions. Anybody could @ > > write a lot os arguments. But I would like to see verifyable@ > > statements. We do have an AlphaStation ES40 with four 600MHz> > > EV6 CPUs and 16GB memory and a SunFire 880 with the 900MHz? > > UltraSparc III. Now I can compare performance/price. Sorry, ? > > but we do not have an application running on both. The only A > > one I could test is I/O. There was a small difference between A > > both. Alpha a little bit faster (under OpenVMS!!!!). Also the E > > boot sequence was much faster (I know this is meaningless because C > > OpenVMS have to be booted only once). But the big difference is ; > > the price. We did pay the double value for the SunFire.  > >  >  > > > Humm, the list price of an 8 CPU V880 is ~100,000 with 16 GB: > the list price of a 4 CPU 667 Mhz 16 GB ES40 is ~180,000 > both dollars.  > = > I would be interested in what you used to benchmark the I/O < > performance of the two systems and what kind of disks they > both have.  K I would be interested in where you found such a ridiculously overstated (by ) a factor of perhaps 3) price for an ES40.    - bill   ------------------------------  $ Date: Mon, 1 Jul 2002 16:51:24 +0530# From: "Vivek Soni" <visoni@bmc.com>  Subject: DDL Files/ Message-ID: <ui0em6g2oabma9@corp.supernews.com>    Hi,   + Can anybody help me with docs on DDL files.    thanks Vivek    ------------------------------  $ Date: Mon, 1 Jul 2002 18:59:28 +0530# From: "Vivek Soni" <visoni@bmc.com>  Subject: Re: DDL Files/ Message-ID: <ui0m69si5thj47@corp.supernews.com>    Hi Michael,   F Data Description Language(DDL) files are used for creating C files and
 header files.   K This method was used in many years back to create .h and .c files, it seems  it is not in use.   ; These are text files. Does any body have more info on this.    Thanks in Advance  Vivek     , Vivek Soni <visoni@bmc.com> wrote in message) news:ui0em6g2oabma9@corp.supernews.com...  > Hi,  > - > Can anybody help me with docs on DDL files.  >  > thanks > Vivek  >  >    ------------------------------  % Date: Mon, 01 Jul 2002 08:22:36 -0400 1 From: Michael Austin <maustin@firstdbasource.com>  Subject: Re: DDL Files1 Message-ID: <3D20498C.AE1B0E6@firstdbasource.com>    Vivek Soni wrote:  >  > Hi,  > - > Can anybody help me with docs on DDL files.  >  > thanks > Vivek   % ddl files?  are they Rdb .sql files?  * what is in these ddl files? text? binary? - Not enough information to give you an answer.  --   Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163 7 Sr. Consultant            http://www.firstdbasource.com                            + http://www.firstdbasource.com/donation.html / 704-947-1089 (Office)     704-236-4377 (Mobile)    ------------------------------  % Date: Mon, 01 Jul 2002 08:54:17 -0600  From: Kevin Handy <kth@srv.net>  Subject: Re: DDL Files& Message-ID: <3D206D19.9070400@srv.net>   Vivek Soni wrote:  > Hi,  > - > Can anybody help me with docs on DDL files.  >   . Look for info on CDD "Common Data Dictionary"., It is the module that handles the DDL files, and defines their format.    ------------------------------   Date: 1 Jul 2002 09:31:32 -0700   From: deltauh@yahoo.com (daniel) Subject: Re: Defcon whitepaper= Message-ID: <e6e1751e.0207010831.7694025e@posting.google.com>   B If you are interested in security I just checked out PointSecure'sA SnapShot.  It lets you know how well your security is set on your F OpenVMS box.  Check it out.  I was quite interested with the results. ) It begs to wonder how much they can help.   ' http://www.pointsecure.com/snapshot.htm    ------------------------------  $ Date: Mon, 1 Jul 2002 11:58:42 +0530# From: "Vivek Soni" <visoni@bmc.com>  Subject: Re: fab=cc$rms_fab / Message-ID: <uhvthca37j8220@corp.supernews.com>   5 I think this would be nice page to start with for RMS   8 http://www.openvms.compaq.com/commercial/c/5492p009.html   Thanks Vivek   J Jim Johnson <Jim.Johnson@software-exploration.nospam.com> wrote in message) news:3d1c28eb.9373358@news.demon.co.uk... H > RMS structures contain a number of fields that need to be initialized.B > Given that a number of them are normally initialized to the sameG > values (e.g. fab$b_bln will always be FAC$C_BLN), C tries to help you E > out by providing 'template' RMS structures that you can start from.  > F > So, the statement 'fab = cc$rms_fab' is copying in the template FAB,G > thus initializing your fab to a valid state.  From there you can make E > the modifications that you need (which are likely to be less than a  > full setup of the FAB).  >  > Jim. >  > C > On Fri, 28 Jun 2002 13:19:42 +0530, "Vivek Soni" <visoni@bmc.com>  > wrote: >  > >Hi, > > < > >struct FAB fab=cc$rms_fab;   /*God knows what this does*/ > >struct NAM nam=cc$rms_nam;  > > B > >I think I have spend enough time to get some info/docs on this. > > ? > >Could anybody help me with the link of page for more info on  > >  > >fab=cc$rms_fab  > >  > >Thanks in Advance > >Vivek > >  > >  > 
 > Jim Johnson  > Software Exploration, Ltd.+ > (remove '.nospam' from the reply address)    ------------------------------  $ Date: Mon, 1 Jul 2002 17:49:14 +0530# From: "Vivek Soni" <visoni@bmc.com>  Subject: Re: fab=cc$rms_fab / Message-ID: <ui0i2jido492a8@corp.supernews.com>    Nice page for RMS   A http://www.openvms.compaq.com/commercial/c/5492p008.html#rms_chap    Thanks Vivek     8 JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message& news:3D1CA554.246424F3@videotron.ca...? > > > struct FAB fab=cc$rms_fab;   /*God knows what this does*/   > > > struct NAM nam=cc$rms_nam; > > >  >  > E > Of course Hoff and Fred know what this does :-) :-) :-) :-) :-) :-)    ------------------------------  % Date: Mon, 01 Jul 2002 11:55:26 +0100 U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> - Subject: Re: Fearless IPF Prognostications... 0 Message-ID: <afpcde$gj9$1@new-usenet.uk.sun.com>   Terry C. Shannon wrote:   M > Probability factor (er, likelihood) that VMS will boot on IPF in 1H02: 0.01 4 > Likelihood that VMS will boot on IPF in 2H02: 0.99M > Likelihood that SRI-supplied Alpha-->IPF VMS binary translator will support  > PL/1 apps: 0N > Likelihood that VMS will end up being more portable (to other architectures) > as a result of IPF port: 0.99  >     5 This all seems reasonable, except the last point. Not 4 that I am actually disagreeing with it, but IMHO VMS1 portability to another platform is a red-herring.   4 It is IA-64 or nothing at all, if IA-64 fails and HP3 have to switch platforms they will be in a state of 8 major disarray, the chances of them sucessfully carrying5 the OpenVMS market to another platform post IA-64 are  almost non existant.   Regards  Andrew Harrison    ------------------------------  % Date: Mon, 01 Jul 2002 14:08:50 +0100 U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> - Subject: Re: Fearless IPF Prognostications... 0 Message-ID: <afpk7i$j5b$1@new-usenet.uk.sun.com>   Nic Clews wrote:  * > Andrew Harrison SUNUK Consultancy wrote: >  >>Terry C. Shannon wrote:  >> >>N >>>Probability factor (er, likelihood) that VMS will boot on IPF in 1H02: 0.015 >>>Likelihood that VMS will boot on IPF in 2H02: 0.99 N >>>Likelihood that SRI-supplied Alpha-->IPF VMS binary translator will support >>>PL/1 apps: 0 O >>>Likelihood that VMS will end up being more portable (to other architectures)   >>>as a result of IPF port: 0.99 >>>  >>> 7 >>This all seems reasonable, except the last point. Not 6 >>that I am actually disagreeing with it, but IMHO VMS3 >>portability to another platform is a red-herring.  >>6 >>It is IA-64 or nothing at all, if IA-64 fails and HP5 >>have to switch platforms they will be in a state of : >>major disarray, the chances of them sucessfully carrying7 >>the OpenVMS market to another platform post IA-64 are  >>almost non existant. >> >       3 You missunderstand my point. I am not talking about 5 the technical process of doing the port. I am talking 4 about migrating the OpenVMS market, thats customers,( ISV's etc from one processor to another.  5 In many respects actually doing the OS port is one of 8 the easiest parts of the process, getting your customers7 to buy into what you are doing and getting the ISV's to  support you is much harder.    Regards  Andrew Harrison     G > Wrong. What has been stated here, and in technical forums is that the I > hardware dependencies in VMS have been [are being] removed. This is not # > a special version of VMS for IPF. F > Therefore, porting to another architecture means working out anotherJ > boot loader, if there isn't one similar to the IPF based one anyway, andF > being able to compile VMS for the other architecture, which arguably- > could even be a cross compilation (target).  > H > I believe it would be worth your while properly familiarizing yourselfI > with what is actually happening in the world of VMS, it may turn out to A > be a lesson for all. And you wouldn't want to end up misleading  > customers would you? > J > As to IA64 failing, well if the investment comparisons between a companyJ > that believes it is the centre of the solar system, and the company thatA > simply just is [more or less] everywhere else, between the chip I > technologies, a 1:10 ratio is something even I'd put a bet on to who'll I > end up top. Even IBM predicts an IA64 dominated world, the 32 bits apps J > will run on the IA64s, and in some cases that will be a necessity as theI > IA32 chips become unavailable. [You mentioned you'd seen the IBM stuff,  > this was in the lot I saw].  >        >    ------------------------------  * Date: Mon, 1 Jul 2002 13:28:02 +0000 (UTC) From: david20@alpha2.mdx.ac.uk- Subject: Re: Fearless IPF Prognostications... + Message-ID: <afpld2$9h1$1@aquila.mdx.ac.uk>   T In article <3D2044EC.6EA41041@127.0.0.1>, Nic Clews <sendspamhere@127.0.0.1> writes:) >Andrew Harrison SUNUK Consultancy wrote:  >>   >> Terry C. Shannon wrote: >>  I >As to IA64 failing, well if the investment comparisons between a company I >that believes it is the centre of the solar system, and the company that @ >simply just is [more or less] everywhere else, between the chipH >technologies, a 1:10 ratio is something even I'd put a bet on to who'llH >end up top. Even IBM predicts an IA64 dominated world, the 32 bits appsI >will run on the IA64s, and in some cases that will be a necessity as the   >IA32 chips become unavailable.   F Only if HAMMER fails. Deciding whether to run your IA32 apps on HammerM or trying to run them on IA64s where they run so much slower is a no brainer.   
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Mon, 01 Jul 2002 07:39:35 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)- Subject: Re: Fearless IPF Prognostications... K Message-ID: <rdeininger-0107020739350001@1cust180.tnt2.nashua.nh.da.uu.net>   F In article <afpcde$gj9$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK@ Consultancy <andrew_nospam.harrison_remove_this@sun#.com> wrote:   >Terry C. Shannon wrote:O >> Likelihood that VMS will end up being more portable (to other architectures)   >> as a result of IPF port: 0.99 > 6 >This all seems reasonable, except the last point. Not5 >that I am actually disagreeing with it, but IMHO VMS 2 >portability to another platform is a red-herring. > 5 >It is IA-64 or nothing at all, if IA-64 fails and HP 4 >have to switch platforms they will be in a state of9 >major disarray, the chances of them sucessfully carrying 6 >the OpenVMS market to another platform post IA-64 are >almost non existant.   D "Portable" and "will be ported" are different concepts.  There is no2 question VMS will be more portable than it is now.   ------------------------------  * Date: Mon, 1 Jul 2002 06:48:07 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br>- Subject: Re: Fearless IPF Prognostications... @ Message-ID: <20020701134807.37633.qmail@web20208.mail.yahoo.com>  & I really would like to know if we will0 run OpenVMS under other Itanium branded machines1 like Dell for example ! I would like to buy a oneu. processor server :-))) but Compaq (i.e. HP) is/ too expensive ....  I would like to know if the - Itanium architecture will be the same betweene/ all the brands. If we can run HP-UX in a Dellsi server.  How is the HAL portability ????s     Regardsr   FC F6 --- "Terry C. Shannon" <terryshannon@attbi.com> wrote:3 > Probability factor (er, likelihood) that VMS wills > boot on IPF in 1H02: 0.01 4 > Likelihood that VMS will boot on IPF in 2H02: 0.995 > Likelihood that SRI-supplied Alpha-->IPF VMS binary  > translator will supporta > PL/1 apps: 05 > Likelihood that VMS will end up being more portable  > (to other architectures) > as a result of IPF port: 0.99r >  >  > -- > Terry C. Shannon > Consultant and Publisher > Shannon Knows HPC $ > PLEASE NOTE MY *NEW* EMAIL ADDRESS > email: terryshannon@attbi.comm6 > Web (info on SKHPC):  www.openvms.org, www.tru64.org >  >      =====( ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazils fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?0 Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com.   ------------------------------  % Date: Mon, 01 Jul 2002 14:50:28 +0100n( From: Nic Clews <sendspamhere@127.0.0.1>- Subject: Re: Fearless IPF Prognostications...e) Message-ID: <3D205E24.77D2465D@127.0.0.1>p  ( Andrew Harrison SUNUK Consultancy wrote: >  > Nic Clews wrote: >  ...  > 5 > You missunderstand my point. I am not talking aboute7 > the technical process of doing the port. I am talking 6 > about migrating the OpenVMS market, thats customers,* > ISV's etc from one processor to another. > 7 > In many respects actually doing the OS port is one ofr: > the easiest parts of the process, getting your customers9 > to buy into what you are doing and getting the ISV's toO > support you is much harder.   C Point understood. I absolutely agree that careful management of thekE marketing and customer information far outweighs any technical issues-E (either way) in any such task, which ultimately leads to the level of  success of that task.1   --  ? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencesr nclews at csc dot comc   ------------------------------  % Date: Mon, 01 Jul 2002 13:02:52 +0100$( From: Nic Clews <sendspamhere@127.0.0.1>- Subject: Re: Fearless IPF Prognostications...e) Message-ID: <3D2044EC.6EA41041@127.0.0.1>w  ( Andrew Harrison SUNUK Consultancy wrote: >  > Terry C. Shannon wrote:  > O > > Probability factor (er, likelihood) that VMS will boot on IPF in 1H02: 0.01l6 > > Likelihood that VMS will boot on IPF in 2H02: 0.99O > > Likelihood that SRI-supplied Alpha-->IPF VMS binary translator will supportr > > PL/1 apps: 0P > > Likelihood that VMS will end up being more portable (to other architectures)! > > as a result of IPF port: 0.99e > >  > 7 > This all seems reasonable, except the last point. Not!6 > that I am actually disagreeing with it, but IMHO VMS3 > portability to another platform is a red-herring.e > 6 > It is IA-64 or nothing at all, if IA-64 fails and HP5 > have to switch platforms they will be in a state of : > major disarray, the chances of them sucessfully carrying7 > the OpenVMS market to another platform post IA-64 areh > almost non existant.  E Wrong. What has been stated here, and in technical forums is that theoG hardware dependencies in VMS have been [are being] removed. This is noti! a special version of VMS for IPF.-D Therefore, porting to another architecture means working out anotherH boot loader, if there isn't one similar to the IPF based one anyway, andD being able to compile VMS for the other architecture, which arguably+ could even be a cross compilation (target).m  F I believe it would be worth your while properly familiarizing yourselfG with what is actually happening in the world of VMS, it may turn out to ? be a lesson for all. And you wouldn't want to end up misleadingw customers would you?  H As to IA64 failing, well if the investment comparisons between a companyH that believes it is the centre of the solar system, and the company that? simply just is [more or less] everywhere else, between the chipSG technologies, a 1:10 ratio is something even I'd put a bet on to who'llaG end up top. Even IBM predicts an IA64 dominated world, the 32 bits apps.H will run on the IA64s, and in some cases that will be a necessity as theG IA32 chips become unavailable. [You mentioned you'd seen the IBM stuff,  this was in the lot I saw].s   -- k? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencesn nclews at csc dot comi   ------------------------------  # Date: Mon, 01 Jul 2002 16:25:55 GMT * From: "Bill Todd" <billtodd@metrocast.net>- Subject: Re: Fearless IPF Prognostications...pB Message-ID: <no%T8.501811$%y.34150805@bin4.nnrp.aus1.giganews.com>  ? "Robert Deininger" <rdeininger@mindspring.com> wrote in message E news:rdeininger-0107020739350001@1cust180.tnt2.nashua.nh.da.uu.net....H > In article <afpcde$gj9$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUKB > Consultancy <andrew_nospam.harrison_remove_this@sun#.com> wrote: >- > >Terry C. Shannon wrote:B > >> Likelihood that VMS will end up being more portable (to other architectures)" > >> as a result of IPF port: 0.99 > >p8 > >This all seems reasonable, except the last point. Not7 > >that I am actually disagreeing with it, but IMHO VMS-4 > >portability to another platform is a red-herring. > >v7 > >It is IA-64 or nothing at all, if IA-64 fails and HPg6 > >have to switch platforms they will be in a state of; > >major disarray, the chances of them sucessfully carryingi8 > >the OpenVMS market to another platform post IA-64 are > >almost non existant.  >aF > "Portable" and "will be ported" are different concepts.  There is no4 > question VMS will be more portable than it is now.  J Since that's *precisely* part of what Andrew just said above, I'm not sureK what you think it added to the discussion - especially as it's phrased lessL7 like a 'me, too' comment than as some kind of rebuttal.-   - bill   ------------------------------  # Date: Mon, 01 Jul 2002 17:05:31 GMTa* From: "Bill Todd" <billtodd@metrocast.net>- Subject: Re: Fearless IPF Prognostications...oC Message-ID: <vZ%T8.558535$Oa1.35591352@bin8.nnrp.aus1.giganews.com>m  5 "Nic Clews" <sendspamhere@127.0.0.1> wrote in message # news:3D2044EC.6EA41041@127.0.0.1...h* > Andrew Harrison SUNUK Consultancy wrote: > >u > > Terry C. Shannon wrote:  > >rL > > > Probability factor (er, likelihood) that VMS will boot on IPF in 1H02: 0.018 > > > Likelihood that VMS will boot on IPF in 2H02: 0.99I > > > Likelihood that SRI-supplied Alpha-->IPF VMS binary translator will  support/ > > > PL/1 apps: 0C > > > Likelihood that VMS will end up being more portable (to other  architectures)# > > > as a result of IPF port: 0.99  > > >: > > 9 > > This all seems reasonable, except the last point. NotF8 > > that I am actually disagreeing with it, but IMHO VMS5 > > portability to another platform is a red-herring.n > >f8 > > It is IA-64 or nothing at all, if IA-64 fails and HP7 > > have to switch platforms they will be in a state ofi< > > major disarray, the chances of them sucessfully carrying9 > > the OpenVMS market to another platform post IA-64 are) > > almost non existant. >pG > Wrong. What has been stated here, and in technical forums is that theyI > hardware dependencies in VMS have been [are being] removed. This is notI# > a special version of VMS for IPF.y  J As Andrew already pointed out (more politely than you chose to above), youA read what he said incorrectly and thus responded inappropriately.w   ...e  J > As to IA64 failing, well if the investment comparisons between a companyJ > that believes it is the centre of the solar system, and the company thatA > simply just is [more or less] everywhere else, between the chiphI > technologies, a 1:10 ratio is something even I'd put a bet on to who'llm
 > end up top.i  I Did you bet similarly on iAPX432?  Or the i860 (though that wasn't such al9 complete disaster)?  Intel is far from immune to failure.@  , > Even IBM predicts an IA64 dominated world,  K Right. That's why they've put POWER into maintenance mode.  Not.  In markedrI contrast to Alpha, they've got major new core design work publicly mapped F out through 2005, with no indication that it will stop after that date either.a  K And perhaps you missed their decision not to bother porting AIX to Itanic anJ while ago, though they certainly seem to be continuing to pursue AIX salesF on POWER platforms and treating Linux as a parallel, lower-end (thoughF important) market rather than as a successor.  Hardly the actions of aC company that expected Itanic to put the rest of its hardware out oft	 business.H  H IBM's philosophy for a while now has been to sell the customer what s/heH wants (novel concept, that:  too bad Compaq and DEC before it weren't asL bright).  That includes Itanics, should they be desired:  most of the effortK to support them had been put in place before it became obvious just how far E short of expectations they would fall, and prudence would dictate not K throwing such investment down the drain as long as Itanic had any chance at H all of succeeding (which is, however, a far cry from 'predicting an IA64 dominated world').    the 32 bits apps-J > will run on the IA64s, and in some cases that will be a necessity as the  > IA32 chips become unavailable.  L IA32 won't begin to become unavailable until late in this decade at the veryG earliest.  By that point in time, IA64's fate will have been thoroughlyuL sealed one way or the other, so suggesting that IA32's disappearance will inG some way help it out of the hole Intel has dug for it is ludicrous.  In H point of fact, IA32's continuing significant *absolute* performance leadE over its bigger, hotter, more expensive, far less 'industry-standard'dH brother that can't even run its *native* code (let alone IA32 code) withG IA32-level performance is more likely to help sink the good ship ItanicrC (even without Hammer's added influence), since once IA64 is clearlySL relegated to being just another 64-bit player (and not a very impressive oneF at that) the impetus for Intel to continue to invest heavily in such a+ low-volume platform will diminish markedly.d   - bill   ------------------------------  % Date: Mon, 01 Jul 2002 07:30:26 -0400I2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: KZTSA and AIT1hK Message-ID: <rdeininger-0107020730260001@1cust180.tnt2.nashua.nh.da.uu.net>a  E In article <howard-DC8FDE.21532230062002@enews.newsguy.com>, Howard S  Shubs <howard@shubs.net> wrote:t  G >I plugged this TurboChannel card into an Alpha 3000 box.  The console  H >sees the card.  It doesn't see the tape drive.  Booted into VMS 7.2-1, 1 >it didn't see the tape drive either.  Any ideas?-  J Start with the old favorites:  cables, termination, and SCSI ID conflicts.  J The KZTSA is a fast, wide, high-voltage differential SCSI controller.  HVDG is NOT the same as today's low-voltage differential.  If the tape drive@E isn't LVD, you'll need a converter.  Unless the tape or the convertero3 provides termination, you'll need a HVD terminator.   G I'd make sure the DEC 3000 and the KZTSA both have the latest firmware:2  @ The DEC 3000 SRM console should be V7.0, with VMS PALcode V5.56. The KZTSA should be at A11.l  > At the DEC 3000 console, you can use the KZTSA's CNFG utility:  & >>> T TC1 CNFG   (This assumes slot 1)   Does the tape show up here?r  ? >>> T TC1 ?      gives you help for the controller's utilities.s  G I have the KZTSA User's Guide in a PDF file.  268 kB.  E-mail me if youw+ want be to send it as an e-mail attachment.r   ------------------------------  % Date: Mon, 01 Jul 2002 07:43:36 -0400 ' From: Howard S Shubs <howard@shubs.net>t Subject: Re: KZTSA and AIT1>< Message-ID: <howard-BBFEA4.07433101072002@enews.newsguy.com>   In article t@ <rdeininger-0107020730260001@1cust180.tnt2.nashua.nh.da.uu.net>,4  rdeininger@mindspring.com (Robert Deininger) wrote:  G > The KZTSA is a fast, wide, high-voltage differential SCSI controller.r  H Oh.  Well, that's the end of my quest, then.  The device is a fast/wide E single-ended type SCSI device.  It's okay, I just figured to use the  3 Alpha for testing.  Guess that won't happen. <grin>    -- i# "Run in circles, scream and shout!"e I hope you have good backups! ) Are there any more networked SJFs around?    ------------------------------   Date: 1 Jul 2002 04:38:25 -0700h- From: mvsjetti@hss.hns.com (Mahesh V S Jetti)rY Subject: Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]t= Message-ID: <adb15ffe.0207010338.7ba24970@posting.google.com>,   Hi jim,s    Thanks for the info.FC  I had inserted the LIB$FIND_FILE_END and my code was working fine.hL  one thing i want to know is how to see the maximum amount of dynamic memoryB  that can be allocated to a process and how much amount of mmemoryF  my process is using? i had tried "show system" command, in that it isF  showing the total physical memory my process is using.I had tried theI  SYSGEN utility but not able to find the exact parameter which determinese@  the maximum limit of memory that can be allocated to a process.J  can u please respond as soon as possible as this is a critical issues for  mew  e  Thanks and regards,  Maheshh      w Jim.Johnson@software-exploration.nospam.com (Jim Johnson) wrote in message news:<3d1c757f.28977176@news.demon.co.uk>...tH > On 28 Jun 2002 07:09:38 -0700, mvsjetti@hss.hns.com (Mahesh V S Jetti) > wrote: >  > >Hi craig, > >b > > Thanks for the help.G > >In my pascal file i am calling LIB$FIND_FILE at different locations. O > >But i am not calling  LIB$FIND_FILE_END  to deallocate the context allocated( > >when we call LIB$FIND_FILE. > > I > >Is it possible to get the following problem due to the above scenario?q > * > Over time it is almost a dead certainty. >  > >.& > >$ write sys$output f$message(99540)' > >%RMS-F-DME, dynamic memory exhaustedy > >eF > >Are there any means of increasing the dynamic memory by tuning some% > >system parameters or by any means?o > D > Yes, but this will not solve your problem.  It will, at best, move9 > back the point where you fail.  You really need to calle > LIB$FIND_FILE_END. >  > Fwiw,1 > Jim. >  > >i5 > >Thank you once again for the information provided.B > >d > >Regards,d	 > >maheshr > >h@ > >In article <adb15ffe.0206280501.2c791914@posting.google.com>,2 > > mvsjetti@hss.hns.com (Mahesh V S Jetti) wrote: > > J > >>  I had a pascal file which had a call to LIB$FIND_FILE that gives theJ > >>  latest version of the file.But problem is some times it is returningJ > >>  value of 99540 even though file exists . i tried to search what thisE > >>  value means. But i was  not able to get any information on when H > >>  LIB$FIND_FILE returns a value  of 99540 and what this value means.M > >>  Can u please help me to get some  idea about it.Any help will be highly2 > >>  appreciated. > > & > >$ write sys$output f$message(99540)' > >%RMS-F-DME, dynamic memory exhaustedc > >d > >Try > >r > >$ help/message dmec > 
 > Jim Johnsony > Software Exploration, Ltd.+ > (remove '.nospam' from the reply address)(   ------------------------------   Date: 1 Jul 2002 04:46:33 -0700 - From: mvsjetti@hss.hns.com (Mahesh V S Jetti)tY Subject: Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]i< Message-ID: <adb15ffe.0207010346.100a2fd@posting.google.com>   Hi jim,H    Thanks for the info.oE  I had inserted LIB$FIND_FILE_END in my code and it was working fine.cI  one thing i want to know how can we see the maximum dynamic memoery that G  can be allocated to a process and how much dynamic memory is currentlyoH  allocated to my process? I had tried "show system" which is showing theF  total physcial memory my process is using.I had also tried the SYSGENH  utility. But not able to trace the exact parameters that will determine@  maximum and current allocation of dynamic memory to a process.   J  cau please respond to it as soon as possible as this was a critical issue  for mee    Thanks and regards,  Mahesh     w Jim.Johnson@software-exploration.nospam.com (Jim Johnson) wrote in message news:<3d1c757f.28977176@news.demon.co.uk>...lH > On 28 Jun 2002 07:09:38 -0700, mvsjetti@hss.hns.com (Mahesh V S Jetti) > wrote: >  > >Hi craig, > >  > > Thanks for the help.G > >In my pascal file i am calling LIB$FIND_FILE at different locations.oO > >But i am not calling  LIB$FIND_FILE_END  to deallocate the context allocatedS > >when we call LIB$FIND_FILE. > > I > >Is it possible to get the following problem due to the above scenario?- > * > Over time it is almost a dead certainty. >  > >r& > >$ write sys$output f$message(99540)' > >%RMS-F-DME, dynamic memory exhausted  > > F > >Are there any means of increasing the dynamic memory by tuning some% > >system parameters or by any means?  > D > Yes, but this will not solve your problem.  It will, at best, move9 > back the point where you fail.  You really need to call, > LIB$FIND_FILE_END. >  > Fwiw,  > Jim. >  > >=5 > >Thank you once again for the information provided.g > >i > >Regards,-	 > >mahesha > >n@ > >In article <adb15ffe.0206280501.2c791914@posting.google.com>,2 > > mvsjetti@hss.hns.com (Mahesh V S Jetti) wrote: > >-J > >>  I had a pascal file which had a call to LIB$FIND_FILE that gives theJ > >>  latest version of the file.But problem is some times it is returningJ > >>  value of 99540 even though file exists . i tried to search what thisE > >>  value means. But i was  not able to get any information on whenoH > >>  LIB$FIND_FILE returns a value  of 99540 and what this value means.M > >>  Can u please help me to get some  idea about it.Any help will be highlyg > >>  appreciated. > > & > >$ write sys$output f$message(99540)' > >%RMS-F-DME, dynamic memory exhaustede > >t > >Try > >  > >$ help/message dmet > 
 > Jim Johnsonp > Software Exploration, Ltd.+ > (remove '.nospam' from the reply address)    ------------------------------  # Date: Mon, 01 Jul 2002 13:46:33 GMT ? From: Jim.Johnson@software-exploration.nospam.com (Jim Johnson)eY Subject: Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]e0 Message-ID: <3d2056c8.20207376@news.demon.co.uk>   Mahesh,t  E It's not as simple as one would like.  DME is RMS' way of saying thatr5 it couldn't allocate the memory it thought it needed. ? Unfortunatately, there are a number of ways in which this might19 happen, and two different memory pools that are involved.u  C If the request is for a process permanennt file, then the memory is4C coming from a fixed pool within P1 space.  The size of that pool ise1 controlled by a SYSGEN parameter called PIOPAGES.(  A Otherwise, the memory comes from P0 space from memory dynamicallyo> allocated to the process.  The amount that can be allocated isB governed by a collection of quotas and sysgen parameters.  SettingC that up properly is something that tends to get described in tuningtF books, manuals, and articles.  It's not something that I'd want to try to describe briefly here.e  F As for tracking how much you're using, and how much you could possiblyC need, and how much you could possibly allocate -- that story is not 
 much simpler.n  E I don't know of any way to easily track the amount of PIO space used.lE As for the P0 space usage, I'm sure that there are tools out there toaE help you track that.  I've tended to use SHOW PROC/CONT and SHOW PROCoD /QUOTA, but these are both very basic tools, and I'm sure that there must be better ones out there.  D As for planning your RMS space requirements, I know that roughly youC should plan on an open file overhead block for every open file, andeE for every wildcard context that you hold (this latter is why you needt@ the LIB$FIND_FILE_END.  It closes down the open wildcard contextE started by the matching LIB$FIND_FILE).  You also need to account forrD a record stream with every $CONNECT that you do.  You then also needD space for each of your data buffers.  For each file the buffers will= each be sized to the maximum bucket size in use on that file.nB Finally, there are lock blocks for shared files, and stack context! blocks for each active operation.n  C What I do not have anymore is the exact sizes for these structures. F I'd guess that a file open context is somewhere around 8-16 KB apiece,B an open record stream ~1KB, and a stack block may be equal to EXEC mode stack pages.  At a guess.  E Oh yes, note also that each open file overhead block may well consumeA a file channel.c   Hope this helps, Jim.    E On 1 Jul 2002 04:46:33 -0700, mvsjetti@hss.hns.com (Mahesh V S Jetti)M wrote:   >Hi jim, >h > Thanks for the info.F > I had inserted LIB$FIND_FILE_END in my code and it was working fine.J > one thing i want to know how can we see the maximum dynamic memoery thatH > can be allocated to a process and how much dynamic memory is currentlyI > allocated to my process? I had tried "show system" which is showing the4G > total physcial memory my process is using.I had also tried the SYSGEN>I > utility. But not able to trace the exact parameters that will determinehA > maximum and current allocation of dynamic memory to a process. A >eK > cau please respond to it as soon as possible as this was a critical issue  > for me >  > Thanks and regards,: > Mahesh >h > x >Jim.Johnson@software-exploration.nospam.com (Jim Johnson) wrote in message news:<3d1c757f.28977176@news.demon.co.uk>...I >> On 28 Jun 2002 07:09:38 -0700, mvsjetti@hss.hns.com (Mahesh V S Jetti)e	 >> wrote:, >> I
 >> >Hi craig,r >> > >> > Thanks for the help. H >> >In my pascal file i am calling LIB$FIND_FILE at different locations.P >> >But i am not calling  LIB$FIND_FILE_END  to deallocate the context allocated >> >when we call LIB$FIND_FILE.y >> >J >> >Is it possible to get the following problem due to the above scenario? >>  + >> Over time it is almost a dead certainty.i >> n >> >' >> >$ write sys$output f$message(99540)e( >> >%RMS-F-DME, dynamic memory exhausted >> >G >> >Are there any means of increasing the dynamic memory by tuning somei& >> >system parameters or by any means? >>  E >> Yes, but this will not solve your problem.  It will, at best, movet: >> back the point where you fail.  You really need to call >> LIB$FIND_FILE_END.  >>   >> Fwiw, >> Jim.b >> d >> >6 >> >Thank you once again for the information provided. >> > >> >Regards,
 >> >mahesh >> >A >> >In article <adb15ffe.0206280501.2c791914@posting.google.com>,.3 >> > mvsjetti@hss.hns.com (Mahesh V S Jetti) wrote:0 >> >K >> >>  I had a pascal file which had a call to LIB$FIND_FILE that gives theoK >> >>  latest version of the file.But problem is some times it is returningPK >> >>  value of 99540 even though file exists . i tried to search what this F >> >>  value means. But i was  not able to get any information on whenI >> >>  LIB$FIND_FILE returns a value  of 99540 and what this value means. N >> >>  Can u please help me to get some  idea about it.Any help will be highly >> >>  appreciated.s >> >' >> >$ write sys$output f$message(99540)o( >> >%RMS-F-DME, dynamic memory exhausted >> > >> >TryV >> > >> >$ help/message dme >> f >> Jim Johnson >> Software Exploration, Ltd.r, >> (remove '.nospam' from the reply address)   Jim Johnsona Software Exploration, Ltd.) (remove '.nospam' from the reply address)t   ------------------------------  # Date: Mon, 01 Jul 2002 14:00:11 GMT $ From: "labadie" <labadie_g@decus.fr>Y Subject: Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]p1 Message-ID: <LfZT8.8$mL1.315491@news.cpqcorp.net>s  : "Mahesh V S Jetti" <mvsjetti@hss.hns.com> wrote in message6 news:adb15ffe.0207010346.100a2fd@posting.google.com...	 > Hi jim,d >a >  Thanks for the info. G >  I had inserted LIB$FIND_FILE_END in my code and it was working fine. K >  one thing i want to know how can we see the maximum dynamic memoery thateI >  can be allocated to a process and how much dynamic memory is currently J >  allocated to my process? I had tried "show system" which is showing theH >  total physcial memory my process is using.I had also tried the SYSGENJ >  utility. But not able to trace the exact parameters that will determineA >  maximum and current allocation of dynamic memory to a process.6 >o   Helloo   for piopages, see 
 begin  urlL http://www.compaq.com/support/asktima/operating_systems/CHAMP_SRC99030800167 1.html end urlr   for ctlpages, seey	 begin urluL http://www.compaq.com/support/asktima/operating_systems/CY-1021490401-1.html end urle   Regardsi   Grard   ------------------------------  # Date: Mon, 01 Jul 2002 14:10:21 GMTc$ From: "labadie" <labadie_g@decus.fr>Y Subject: Re: LIB$FIND_FILE returns 99540( dynamic menory exhausted)[Urgent help required]t1 Message-ID: <hpZT8.9$gK1.286874@news.cpqcorp.net>     Hello   sorry for my previous post.c  - To check the free Piopages for a process, uses	 begin url L http://www.compaq.com/support/asktima/operating_systems/CHAMP_SRC99030800167 1.html end url   1 and to check the free Ctlpages for a process, use.	 begin urluL http://www.compaq.com/support/asktima/operating_systems/CY-1021490401-1.html end urli   regardso   Grard   ------------------------------  % Date: Mon, 01 Jul 2002 07:50:48 -0400c2 From: Atlant Schmidt <atlantnospam@mindspring.com>1 Subject: Re: Mozilla and the mysterious downloads . Message-ID: <3D204218.BA5C1F62@mindspring.com>   Dirk Munk wrote:  / > Mozilla 1.1 is the latest version by the way.r  , Only if you like Alpha-testing stuff. Here's) what the Mozilla web site says about 1.1:v   > Mozilla 1.1 Alpha Released >o. > Mozilla 1.1 Alpha is our bleeding edge alpha, > all the changes that happened on the trunk, > since we branched for Mozilla 1.0. See the' > release notes to find out what's new.s                 Atlant   ------------------------------  # Date: Mon, 01 Jul 2002 12:38:27 GMT 2 From: Mike Rechtman <michael.rechtman@digital.com>5 Subject: Re: My conversation with Linus about VMS ...e+ Message-ID: <3D20751A.6FF16144@digital.com>    Atlant Schmidt wrote:n >  > Jeff Campbell wrote: > F > > Read what I wrote. The NT-only AlphaServers, alias Digital Servers > > (330x, 530x, 730x).a > >s7 > > The Digital Personal WorkStations will boot VMS if:  > >iE > >   1) there is a supported SCSI controller and disk in the machinet9 > >   2) there is a supported SCSI CD-ROM in the machine.a > >-; > > > Even a half-flash system that's currently loaded with- > > > the WIN/NT flash?  > > >o > >  > > Install the SRM firmware.  >  > Ahh, so if I:.) >   1. Choose from among certain systems,a >   2. Add hardware to some, >   3, Add firmware to some,% >   4. And don't expect support, thene7 > *ALL* systems (in the sub-rosa subset) will boot VMS.  > > > Yes, that's certainly the logical equivalent of the original9 > statement that one could take a VMS disk to "any Alpha"d > and boot it there. > $ > Thanks for clarifying that for me! >  > Atlant  F Well if you're set on picking nits, why not add "Will it boot an AlphaD with no memory installed?" - Surely that is included in the class of
 "All Alphas."rF My understanding of the original query was "will it boot from a systemF disk built on any other Alpha, with no special configuration" with theF understood addendum "If such an Alpha can boot VMS." I realise that isC only my understanding of an unstated conditional, but then IANAL...e   Mike  E ---------------------------------------------------------------------5E Usual disclaimer: All opinions are mine alone, perhaps not even that.u? Mike Rechtman                            *rechtman@tzora.co.il*oF Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------i -----BEGIN GEEK CODE BLOCK------ Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------0   ------------------------------  % Date: Mon, 01 Jul 2002 08:03:29 -0400u2 From: Atlant Schmidt <atlantnospam@mindspring.com>5 Subject: Re: My conversation with Linus about VMS ...f. Message-ID: <3D204511.E9516FBE@mindspring.com>   Jeff Campbell wrote:  D > Read what I wrote. The NT-only AlphaServers, alias Digital Servers > (330x, 530x, 730x).s >a5 > The Digital Personal WorkStations will boot VMS if:s >hC >   1) there is a supported SCSI controller and disk in the machinee7 >   2) there is a supported SCSI CD-ROM in the machine.o >d9 > > Even a half-flash system that's currently loaded with  > > the WIN/NT flash?h > >m >v > Install the SRM firmware.S  
 Ahh, so if I::'   1. Choose from among certain systems,    2. Add hardware to some,   3, Add firmware to some,#   4. And don't expect support, theno5 *ALL* systems (in the sub-rosa subset) will boot VMS.d  < Yes, that's certainly the logical equivalent of the original7 statement that one could take a VMS disk to "any Alpha"t and boot it there.  " Thanks for clarifying that for me!   Atlant   ------------------------------   Date: 1 Jul 2002 14:10:14 GMT 1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon) 5 Subject: Re: My conversation with Linus about VMS ...l+ Message-ID: <afpns6$ak7$1@info.cs.uofs.edu>w  - In article <87sn37h3y6.fsf@prep.synonet.com>,a/  Paul Repacholi <prep@prep.synonet.com> writes:e |> |> s> |> I still consider that M-PLUS is the best of the lot. Modulo |> 16 bit, etc...M  ? Nope, RSTS is (not was) the best.  Too bad there's no chance ofs> something like that ever getting open-sourced.  I could really, get into ports for other architectures.  :-)   bill   -- hJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   e   ------------------------------  % Date: Mon, 01 Jul 2002 08:42:34 -0600d% From: Dan O'Reilly <dano@process.com>t5 Subject: Re: My conversation with Linus about VMS ...eB Message-ID: <5.1.0.14.2.20020701084043.01bee1c8@raptor.psccos.com>  , At 08:10 AM 7/1/2002, Bill Gunshannon wrote:. >In article <87sn37h3y6.fsf@prep.synonet.com>,1 >  Paul Repacholi <prep@prep.synonet.com> writes:n >|>l >|>u? >|> I still consider that M-PLUS is the best of the lot. Moduloi >|> 16 bit, etc... >c@ >Nope, RSTS is (not was) the best.  Too bad there's no chance of? >something like that ever getting open-sourced.  I could reallys- >get into ports for other architectures.  :-)2  F RSTS was the best TIMESHARE in wide use (although, IAS would give it aE run for its money).  RSX was far and away the best multiuser realtimeIH system.  RT was the best single-user realtime system.  Used all of them,: starting with V4 for RSTS, V3.1 for RSX and V2.0 for RT...   ------I +-------------------------------+---------------------------------------+tI | Dan O'Reilly                  |                                       |4I | Principal Engineer            |  "Why should I care about posterity?  |aI | Process Software              |   What's posterity ever done for me?" |oI | http://www.process.com        |                    -- Groucho Marx    |TI +-------------------------------+---------------------------------------+n   ------------------------------  % Date: Mon, 01 Jul 2002 10:14:49 -0500-& From: jlsue <jlsuexxxz@screaminet.com>5 Subject: Re: My conversation with Linus about VMS ... 8 Message-ID: <v0gphuobhahri1val6u2un4mcluj0fk0em@4ax.com>  2 On Fri, 28 Jun 2002 08:14:21 -0400, Atlant Schmidt$ <atlantnospam@mindspring.com> wrote:  
 >jlsue wrote:  >a5 >> On Thu, 27 Jun 2002 08:15:14 -0400, Atlant Schmidtn' >> <atlantnospam@mindspring.com> wrote:t >> >> >Lyndon Bartels wrote:s >> >M >> >> I said that no matter what, as long as it's alpha, I can take my systemaM >> >> disk and move it from one box to another, and it'll work. I may have tohA >> >> retune, but I can get a basic OS running with no changes...d >> >>r= >> >> The person's response.... "Wow,  that's *REALLY* nice."e >> >* >> >It'd be even better if it were *TRUE*. >> >'Shame it's not, ehh?o >> >
 >> >Atlant >> > >>" >> Why do you make this statement?D >> I KNOW that I can boot my system disk on any VMS-supported Alpha.C >> I've done it and even performed off-line VMS upgrades for entire ( >> VMSclusters by using this capability. >.7 >Note the change from the original statement "no matter.: >what, as long as it's an Alpha" to your re-cast "any VMS-8 >supported Alpha". There's a world of difference between >those two.F >    So what?E First, there's only a very small proportion of Alpha systems that VMS  will not run on.  B Second, The range of systems on which it *is* true is so wide thatC most people wouldn't notice the difference.  So for all *practical*i purposes, it is true.t  D Third, there's still no other OS that can do it as flawlessly as VMS' across such a wide spectrum of systems.e  1 Not speaking for anyone, certainly not DEC/Compaq - (get rid of the xxxz in my address to e-mail)    ------------------------------  % Date: Mon, 01 Jul 2002 11:43:37 -0400a2 From: Atlant Schmidt <atlantnospam@mindspring.com>5 Subject: Re: My conversation with Linus about VMS ...=. Message-ID: <3D2078A9.6EEF2237@mindspring.com>   Bill Gunshannon wrote:  # > Nope, RSTS is (not was) the best.b   (Shameless self-promotion...).  2 Yes, especially the Space Invaders Run-Time-System3 (INVADE.RTS). Using it, our dual-RK06 11/60 used to 5 be able to support a suprising number of simultaneousr( players of my VT-52 Space Invaders game.  4  I wonder if I still have the sources kicking around
 somewhere...?d   Atlant   ------------------------------  * Date: Mon, 1 Jul 2002 06:50:12 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br> Subject: OpenVMS 7.3 Update 200>@ Message-ID: <20020701135012.46251.qmail@web20207.mail.yahoo.com>  5 Any idea when the  OVMS 7.3 new update 200 (or better 6 Service Pack  2 ) will be launched. There are tons of 2 7.3 patches after Update 100.... I will install an ES-40 with OVMS 7.3 .... E   Regards    FC     =====. ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - BrazilM fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?0 Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.como   ------------------------------  $ Date: Mon, 1 Jul 2002 09:18:30 -05001 From: "Dave Gudewicz" <david.gudewicz@abbott.com>t# Subject: Re: OpenVMS 7.3 Update 200,1 Message-ID: <afpoet$ro3$1@fizban.pprd.abbott.com>a  J Congratulations on the World Cup championship.  Didn't you get the day off because of this?  ;-)t   -- Dave...   G It is noble to teach oneself, but still nobler to teach others-and lessl trouble. -----Mark TwainN  ; "Fabio Cardoso" <fabiopenvms@yahoo.com.br> wrote in messagem: news:20020701135012.46251.qmail@web20207.mail.yahoo.com...7 > Any idea when the  OVMS 7.3 new update 200 (or betterl7 > Service Pack  2 ) will be launched. There are tons of 4 > 7.3 patches after Update 100.... I will install an > ES-40 with OVMS 7.3 .... >i	 > Regards  >  > FC >  > =====w > ========================== > Fbio dos Santos Cardoso > OpenVMS System Manager > Rio de Janeiro - Brazily > fabiopenvms@yahoo.com.br > ========================== >y4 > __________________________________________________ > Do You Yahoo!?2 > Yahoo! - Official partner of 2002 FIFA World Cup > http://fifaworldcup.yahoo.como   ------------------------------  * Date: Mon, 1 Jul 2002 08:41:04 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br># Subject: Re: OpenVMS 7.3 Update 200e@ Message-ID: <20020701154104.53931.qmail@web20209.mail.yahoo.com>   Hi Daveh  ( Thanks !!!! We have 5 Stars now ! :-))))    - Get the day off ? No way ! I work for a state  company.... :-))   Regardsd   FC l      4 --- Dave Gudewicz <david.gudewicz@abbott.com> wrote:1 > Congratulations on the World Cup championship. n > Didn't you get the day off > because of this?  ;-)o >  > --	 > Dave...$ > 3 > It is noble to teach oneself, but still nobler too > teach others-and less 
 > trouble. > -----Mark Twainh > 5 > "Fabio Cardoso" <fabiopenvms@yahoo.com.br> wrote in 	 > messagee >a: news:20020701135012.46251.qmail@web20207.mail.yahoo.com...2 > > Any idea when the  OVMS 7.3 new update 200 (or > better6 > > Service Pack  2 ) will be launched. There are tons > of6 > > 7.3 patches after Update 100.... I will install an > > ES-40 with OVMS 7.3 .... > >c > > Regardsl > >X > > FC > >k	 > > =====  > > ========================== > > Fbio dos Santos Cardoso > > OpenVMS System Manager > > Rio de Janeiro - Brazil, > > fabiopenvms@yahoo.com.br > > ========================== > >e6 > > __________________________________________________ > > Do You Yahoo!?4 > > Yahoo! - Official partner of 2002 FIFA World Cup! > > http://fifaworldcup.yahoo.comw >  >      =====i ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazill fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?0 Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.comw   ------------------------------   Date: 1 Jul 2002 14:50:44 GMTt1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)o Subject: OT: RSTS Question+ Message-ID: <afpq84$ak7$2@info.cs.uofs.edu>r  A I admit this is off-topic, but this is the greatest collection ofc9 serious DECies and ex-Decies I know of, so here goes.....e  @ Any chance anyone here has a copy of RSTS Documentation for someA version at the level of 9.7 or newer that they want to get rid of A or that they could be convinced to at least provide a copy of oneF particular section??  > I am badly in need of the Programmers Manual.  I would like to= devote some time this summer to trying to port the Kent Small>@ Systems TCPIP Stack to RSTS and realy need this manual to do it.  : Of course, I would be even happier to get a full set.  :-)   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Mon, 01 Jul 2002 10:10:18 +0200Y' From: JOUKJ <joukj@hrem.stm.tudelft.nl>E Subject: Re: parsing >255m2 Message-ID: <3D200E6A.3080908@hrem.stm.tudelft.nl>   Guy Peleg wrote: > Aaron, > G > The current DCL limit is 255 bytes per command. It can be extended toi > 1024 using the hyphen sign.@ > J > We are currently working on a project to increase this limit. New limitsH > should be 4096 bytes per command, using the hyphen 8192 bytes. We have. > also increased the symbol size to 1024bytes. > 7 > The new DCL is currently scheduled to ship in H1CY03.aH Good that will certainly help. But why cannot the command string length I not be allocated dynamically upon its need? Probably I will go over this vG 8192 bytes easily when link complicated programs using many objects in hH many different directories. Now I split up because the 1024 limitation, ' but sometimes I need more than 8 parts.e                     Jouk   ------------------------------  % Date: Mon, 01 Jul 2002 19:39:30 +0010r% From: paddy.o'brien@zzz.tg.nsw.gov.au  Subject: Re: parsing >255o5 Message-ID: <01KJLBUFP93M00034I@tgmail.tg.nsw.gov.au>r  N Jouk wrote (in response to Guy Peleg's saying that DCL command limit is being  increased):   I >Good that will certainly help. But why cannot the command string length 5J >not be allocated dynamically upon its need? Probably I will go over this H >8192 bytes easily when link complicated programs using many objects in I >many different directories. Now I split up because the 1024 limitation,  ( >but sometimes I need more than 8 parts.  M Just a thought, why not use OLBs.  I doubt that you have each object file in i a different directory.  N This should also speed up linking as an .OLB is only one file to open, versus 
 each .OBJ.  K In our applications, we group our applications and utilities into separate  K directorys (under CMS/MMS control).  No application references more than 5  J directories, hence 5 .OLBs maximum, but a heck of a lot of .OBJs would be  referenced.i   Regards, Paddy   ------------------------------   Date: 1 Jul 2002 06:33:55 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen)a Subject: Re: parsing >255t3 Message-ID: <vAgza3z7Sb3q@eisner.encompasserve.org>   \ In article <3D200E6A.3080908@hrem.stm.tudelft.nl>, JOUKJ <joukj@hrem.stm.tudelft.nl> writes:  J > Good that will certainly help. But why cannot the command string length - > not be allocated dynamically upon its need?   E DCL storage is in P1 space, and below it are the stacks.  I don't see G anyplace to dynamically allocate supervisor mode memory, unless you set1F up a special area in P1 space above the stacks.  But if that is all itI is to be used for, why not dedicate it to the command buffer in the firsta, place?  That is what they seem to be doing !   ------------------------------  % Date: Mon, 01 Jul 2002 07:34:58 -0400o2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: parsing >255/K Message-ID: <rdeininger-0107020734580001@1cust180.tnt2.nashua.nh.da.uu.net>-  8 In article <3D200E6A.3080908@hrem.stm.tudelft.nl>, JOUKJ" <joukj@hrem.stm.tudelft.nl> wrote:  I >Good that will certainly help. But why cannot the command string length 4J >not be allocated dynamically upon its need? Probably I will go over this H >8192 bytes easily when link complicated programs using many objects in I >many different directories. Now I split up because the 1024 limitation,  ( >but sometimes I need more than 8 parts.  H Do you know about linker option files?  You don't ever need long command lines for the linker.e   ------------------------------  % Date: Mon, 01 Jul 2002 09:51:59 -0700-' From: David Mathog <mathog@caltech.edu> 8 Subject: Re: Powerstorm 300/350 cards and OpenGL on OVMS+ Message-ID: <3D2088AF.2DCC8B53@caltech.edu>n   Hiroyuki Tanaka wrote: >  > Dear Readers,1 > D > I have a DEC Alpha 500AU with an Elsa Gloria card and I wishing toG > upgrade to use OpenGL.  I have looked at the SPD for Open3D and thereSD > is no mention of support for the Powerstorm 300 and Powerstorm 350 > cards. > E > Will the Powerstorm 300 and 350 series of card work with OpenGL andi > OVMS?  > 7 > What is the minimum version of VMS I need to install?c > @ > Which card do I need to get if I to use OpenGL overlays?  (The8 > Powerstorm 4d20 supported overlays, but not the 3d30.)  8 The options for OpenGL on VMS are not at all attractive,5 either for price, performance, or support.  Given the 5 extremely limited state of VMS graphics these days itu: really makes very little sense to continue work on that OS> for graphics programs.  Unless your X11 code is linked to some exceedingly VMS , specific internals it shouldn't be very hard2 to move it to Unix (more to go to Windows though.)9 Have a look at the cards Xi supports on Linux or the manyo; OpenGL cards supported on Windows. We've been experimentinge9 with a Radeon 8500 using the Xi driver on Linux and it is 7 _incredibly_ fast and the whole system was about $1600.s6 As for support, there was a major glitch when we tried9 to set up the system and Xi fixed it in a couple of weeksd: (very fast for a board level problem with the Via AGP chip8 set). The new Wildcat VP cards which are just coming out1 look very good too, although for OpenGL they will-3 be Windows only until Xi writes a driver for them. m   Regards,   David Mathog mathog@caltech.edu   ------------------------------   Date: 1 Jul 2002 08:34:17 -0700.% From: whohe@whoever.com (DL Phillips) & Subject: Restricted Login Terminations= Message-ID: <af0dc2ea.0207010734.18a4d866@posting.google.com>"   VMS v6.2 & up:  -  When a user's login has /noaccess times set,-
  the FM says:N  ,  "Users still logged in when the access time+  has expired... their process is terminated+  by the job controller"   )  How does this termination occur? Are allE)  buffers flushed, and such, or is there a-  chance of file corruption?u      TIA3   DL Phillips ,   is Who He at Whoever in the kingdom of com   ------------------------------  * Date: Mon, 1 Jul 2002 11:08:04 +0100 (BST)F From: =?iso-8859-1?q?Tadimeti=20Keshav?= <keshav_tadimeti@yahoo.co.uk> Subject: RMU??@ Message-ID: <20020701100804.93161.qmail@web21007.mail.yahoo.com>  
 Hello all,0 Can someone tell me what RMU (on OVMS) is? Is it similar to RMS?e   TIAc Keshav  2 __________________________________________________ Do You Yahoo!?+ Everything you'll ever need on one web paget- from News and Sport to Email and Music ChartsT http://uk.my.yahoo.com   ------------------------------   Date: 1 Jul 2002 06:35:14 -0600r- From: Kilgallen@SpamCop.net (Larry Kilgallen), Subject: Re: RMU??3 Message-ID: <u36Hy2sL$Bxy@eisner.encompasserve.org>m   In article <20020701100804.93161.qmail@web21007.mail.yahoo.com>, =?iso-8859-1?q?Tadimeti=20Keshav?= <keshav_tadimeti@yahoo.co.uk> writes:M > Hello all,2 > Can someone tell me what RMU (on OVMS) is? Is it > similar to RMS?t  L I believe RMU is a management utility for the Oracle (nee DEC) Rdb database.   ------------------------------  % Date: Mon, 01 Jul 2002 08:19:03 -0400w1 From: Michael Austin <maustin@firstdbasource.com>@ Subject: Re: RMU??2 Message-ID: <3D2048B7.62AE1D69@firstdbasource.com>   Larry Kilgallen wrote: >  > In article <20020701100804.93161.qmail@web21007.mail.yahoo.com>, =?iso-8859-1?q?Tadimeti=20Keshav?= <keshav_tadimeti@yahoo.co.uk> writes:  > > Hello all,4 > > Can someone tell me what RMU (on OVMS) is? Is it > > similar to RMS?l > N > I believe RMU is a management utility for the Oracle (nee DEC) Rdb database.  > That would be correct. open/close databases, backup databases,> stats - in real-time - something Oracle still can't do... even% after 15+ years of seeing Rdb do it. n  > Rdb is one of the most powerful databases on the planet - buth  then again I am somewhat biased.   -- 0 Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163e7 Sr. Consultant            http://www.firstdbasource.como                          0+ http://www.firstdbasource.com/donation.html3/ 704-947-1089 (Office)     704-236-4377 (Mobile)p   ------------------------------  $ Date: Mon, 1 Jul 2002 14:50:44 +0200$ From: "Peter Flunger" <p-i-b@gmx.at> Subject: Re: RMU??0 Message-ID: <afpj75$up8$1@newsreader1.netway.at>  5 "Tadimeti Keshav" <keshav_tadimeti@yahoo.co.uk> wrotes > Hello all,2 > Can someone tell me what RMU (on OVMS) is? Is it > similar to RMS?t  B      The Rdb Management Utility (RMU) lets database administratorsE      display useful information about DEC Rdb databases. RMU commands B      are executed at the OpenVMS system prompt. RMU command syntaxF      follows the rules and conventions of the DIGITAL Command Language      (DCL).-         See the RMUALTER help also.  A      RMU commands display the contents of database files, control B      the DEC Rdb monitor process, verify data structures, and listC      information about current database users and database activity       statistics.  >      RMU commands consist of words, generally verbs, that have9      parameters and qualifiers to define the action to be-      performed.-  
 Regards Peter- PS:-D I am 'still' working with RMU every now and then, although 99,99% ofB our production databases have been migrated to Oracle 7 years ago,@ just shortly before Oracle purchased RDB - what an irony....  :)   ------------------------------  % Date: Mon, 01 Jul 2002 13:45:42 +0100aU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>u+ Subject: Re: Sun benchmarketeering campaignt0 Message-ID: <afpis6$ilg$2@new-usenet.uk.sun.com>   Rob Young wrote:  c > In article <0ljq11rbuKMv@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:  >  >  >>	Reference?  Sure... >>/ >>http://www.digit-life.com/articles/ibmpower4/e >> >>4. POWER4 diew >>E >>The POWER4 houses 2 processors each having an L1 cache for data andRL >>instructions. The die has a single L2 cache of 1450 KBytes controlled by 3N >>separate controllers connected to the cores via a CIU (Core Interface Unit).M >>The controllers work independently and can process 32 bytes per clock. EachoP >>processor uses two separate 256-bit buses to connect the CIU for data fetchingP >>and data loading, as well as a separate 64-bit bus to save the results; the L2I >>cache has a bandwidth of 100 GBytes/s. The L2 cache's system looks wellaJ >>balanced and very powerful. Each processor has a special unit to supportP >>noncachable operations (Noncacheable Unit). The L3 controller and the memory'sQ >>one are located on die as well. For connection with the L3 cache working at 1/3lJ >>of the processor's speed and with the memory there are two 128-bit busesM >>operating at 1/3 of the processor's frequency. The throughput of the memory Q >>interface is about 11 GBytes/s. Data flows coming from the memory and L2 and L3DM >>caches and the buses of the chips are controlled by the Fabric Controller: - >> >> >> >  > 	Big mistake.  > D > 	The L2 cache is 1.5 MBytes, not "more than 5 MBytes" , an earlierA > 	link discovered is combining all caches to lead me to believe   > 	5.6 MBytes was/is the total.> >     ! Ahh the dangers of cut and paste.i  0 So I guess thats the end of the argument thanks.     Regardse   Andrew Harrison    ------------------------------  % Date: Mon, 01 Jul 2002 11:41:54 +0100tU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>n+ Subject: Re: Sun benchmarketeering campaign 0 Message-ID: <afpbk2$gb8$1@new-usenet.uk.sun.com>   Paul Repacholi wrote:-  Y > Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes:c >  > A >>Is this against the rules of SPEC, no, they were drafted before.A >>this kind of configuration was possible on micorprocessor basedp
 >>systems. >> >    > E >>But it means that it is now virtually impossible to use SPECint andpF >>SPECfp as a usefully measure ot even raw CPU int and fp performance. >> >   C > Look at the Spec rules again Andrew. There is a seperate category @ > for `depopulated' or partial systems. However, IBM ships a 1/2D > minded config *as standard*, so it is not a partial system. A 8x00D > could be configured with 1 CPU per board, but it is NOT a standard
 > config.  >     ? Yes they do its called the P690 HPC and their reasons for doingo= this are also well documented, having 2x the L1 and 2x the L2w; cache plus more throughput per CPU and less contention is ag7 good thing. That of course does not help you if you area7 buying the P690 Turbo which is marketed squarely at thed: commercial server space and for which the SPEC results are? not at all helpfull. IBM have also chosen to run the SPECint/fpr$ number on the Turbo and not the HPC.  ? And as for the SPEC rules, the P690 HPC counts as a depopulated5A system (it has half the CPU's disabled). The P690 Turbo does not.M     Regardsa Andrew Harrisonl   ------------------------------  % Date: Mon, 01 Jul 2002 13:43:41 +0100fU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> + Subject: Re: Sun benchmarketeering campaign=0 Message-ID: <afpiol$ilg$1@new-usenet.uk.sun.com>   Rob Young wrote:   > In article <afi38a$8f2$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes: >  >> >>Rob Young wrote: >> >>  = >>Well since there are currently no CPU's that I know of with-? >>a 5 MB onchip cache with a published SPECint or SPECfp numberD@ >>for them I guess your point isn't a terribly clever one is it. >> >> > ? > 	Well yes.  But in your haste to reply you assume I am making.@ > 	a big mistake.  You wouldn't be correct.  By running just oneC > 	processor of a two processor core IBM has posted SpecInt/Specfp.)F > 	All 3 L2s are accessed by the single processor.  Details everywhereG > 	to be found and yes, according to nomenclature considered on-die anduB > 	I will agree with that.  The reason is simple.  L2 runs at coreC > 	speed.  If it ran at anything less than core speed, I don't cares9 > 	where it is located, it isn't very interesting, is it?r >     < No sorry, the L2 cache on a Power 4 is only sharable between9 2 CPU's. Your interpretation of the article it incorrect.d  8 Each power 4 chip contains 2 CPU's which share a 1440 KB4 L2 cache. 4 Power 4 chips are packaged onto a single6 MCM module and the L3 cache is external to the MCM and< is usually 32 MB per Power4 chip but all 128 MB is available to all 8 CPU's.   9 The fact that the MCM module contains another 3 L2 cacheso7 is not relevant in this case, they are not available to, other CPU's.  > If you need confirmation of this then you should reference the= IBM SPECint and SPECfp disclosures which list the L2/L3 cache?= used for the benchmarks, this clearly shows 1 x L2 not sharedp= hence 1440 and 8 x L3 or 128 MB.Or to the Power 4 white paper=  M http://www-1.ibm.com/servers/eserver/pseries/hardware/whitepapers/power4.htmla  ; This is a better source of information than the articel you_ posted._  < Incedentally if your thesis was true then it would of course< even more effectively distort the SPECint/SPECfp numbers for the Power4.   ; The internet is a very usefull place to do research, havingt< a basic understanding of a subject however is a very usefull pre-requesite.    @ >>None do any I/O of any note, none get any significant benefitsD >>from running on a 64 bit system, none of them make any significant >>use of any OS facilities.o >> >> > ? > 	So?  TPC-H does little to exercise the floating point units. E > 	Not enough system stuff for you sure.  But it certainly is a good tD > 	indicator of just how well hardware runs given applications that  > 	*MOST* people are using.i >     ; So if this is true why is SPECint on Alphas such a terriblet< predictor of the TPC-H performance of a GS320 and why is the< same true for TPC-C, why does this also apply to SAP, Oracle	 apps etc.   6 Lets examine TPC-C, TPC-H and Oracle apps SAP has been done to death already.  3 This table shows the model, the number of CPU's the.5 clock speed, the SPECint per CPU, the actual relativen9 performance on the benchmark and the expected performance| based on SPECint.s    : Benchmark   Model  No   Mhz   SPECint Actual Expected perf  2 TPC-C       GS320  32   1001  621     +5%    +267%"          IBM P680   24   600   310  1 TPC-H       GS320  32   731   397     +5%    +36%t"          Sun F6800  24   750   390  1 Oracle apps GS320  32   731   397     -28%   +36%m"          Sun F6800  24   750   390  8 So you can see that for SPECint is a spectacular failure  7 at predicting the relative performance of the GS320 and0. the system thats closest to it in performance.  7 It isn't much better for TPC-H and Oracle apps. Now why 4 is this, could it be that there are other factors at: work here which are limitters on performance which SPECint does not measure.      >  >  >>Filesystem - I/O no.# >>Virtual Memory and scheduling no.s >>Threading no.s >>Context switching no.  >>Real time no.n >> >  > > > 	I never claimed it was an OS test.  It is a test to see howH > 	well hardware is able to be pushed.  I contend hardware is important.+ > 	I contend hardware encompasses a system.a >     @ But how well hardware can be pushed and you only mean CPU, cacheA and a little memory in this context is to a very large extent foraE real applications limitted by OS, VM, Threads, I/O, Memory contentiont7 etc none of which are modeled at all by SPECint/SPECfp.w  ? Very few people that I know of run general purpose systems thatt? do not have an OS, but SPECint and SPECfp are not influenced tot4 any significant extent by what the OS is capable of.     >  > - >>They all fit in less than 200 MB of RAM andI. >>can given the right configuration run almost+ >>entirely in cache as demonstrated by IBM.t >> > B > 	And yet their benchmarks don't reflect dominance, so your pointB > 	is?  Besides, the issue isn't IBM , I stated that earlier.  The? > 	issue is Mr. Yen's statement.  It is a false statement.  Hisn; > 	attempts to downplay Itanium 2 because his company's CPUi0 > 	is under-powered in comparison are laughable. >     C They have the fastest number for a 64bit CPU for SPECint and SPECfp E and the second fastest for SPECint it you include IA-32. So currently-E if you are building a largish or even smallish server and you beleivepF that SPECint is a good systems benchmark then Power 4 is the processor to beat.     > 6 >>>	I'm only getting started.  There a number of holesJ >>>	in your support of Mr. Yen that require detailed technical response.   >>>c' >>>	Something you seldom supply, right?  >>>g >>>t >>/ >>Ohh come on Rob are you really that deluded ? / >>I would really welcome a technical discussions/ >>with you, it hasn't happend yet but if you upo$ >>your game then it may be possible. >> >> > ? > 	Sure.  And watch how fast the IBM notes above get trimmed ort > 	spun.  Spin on Andrew.t >     5 As you can see I didn't trim or spin on the IBM notesi5 there was no need, your reading of them was incorrect-3 as you will now realise having read the IBM Power 4r white paperr     Regardso   Andrew Harrison)   ------------------------------   Date: 1 Jul 2002 11:17:38 -06001+ From: young_r@encompasserve.org (Rob Young)t+ Subject: Re: Sun benchmarketeering campaignJ3 Message-ID: <nWVkoybLCUiQ@eisner.encompasserve.org>t   In article <afpis6$ilg$2@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes: >  >  > Rob Young wrote: > d >> In article <0ljq11rbuKMv@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: >> m >> o >>>	Reference?  Sure...e >>>e0 >>>http://www.digit-life.com/articles/ibmpower4/ >>>? >>>4. POWER4 die >>>oF >>>The POWER4 houses 2 processors each having an L1 cache for data andM >>>instructions. The die has a single L2 cache of 1450 KBytes controlled by 3 O >>>separate controllers connected to the cores via a CIU (Core Interface Unit)..N >>>The controllers work independently and can process 32 bytes per clock. EachQ >>>processor uses two separate 256-bit buses to connect the CIU for data fetchingOQ >>>and data loading, as well as a separate 64-bit bus to save the results; the L2fJ >>>cache has a bandwidth of 100 GBytes/s. The L2 cache's system looks wellK >>>balanced and very powerful. Each processor has a special unit to supporthQ >>>noncachable operations (Noncacheable Unit). The L3 controller and the memory'sIR >>>one are located on die as well. For connection with the L3 cache working at 1/3K >>>of the processor's speed and with the memory there are two 128-bit busespN >>>operating at 1/3 of the processor's frequency. The throughput of the memoryR >>>interface is about 11 GBytes/s. Data flows coming from the memory and L2 and L3N >>>caches and the buses of the chips are controlled by the Fabric Controller:  >>>c >>>r >>>i >> u >> 	Big mistake. >> mE >> 	The L2 cache is 1.5 MBytes, not "more than 5 MBytes" , an earliersB >> 	link discovered is combining all caches to lead me to believe   >> 	5.6 MBytes was/is the total. >> d >  > # > Ahh the dangers of cut and paste.i > 2 > So I guess thats the end of the argument thanks. >    	No.   			Rob   ------------------------------  % Date: Mon, 01 Jul 2002 17:47:30 +0100tU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>)+ Subject: Re: Sun benchmarketeering campaigni0 Message-ID: <afq11i$n61$2@new-usenet.uk.sun.com>   Rob Young wrote:   > In article <afpis6$ilg$2@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes: >  >> >>Rob Young wrote: >> >>d >>>In article <0ljq11rbuKMv@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: >>>o >>>. >>>e >>>>	Reference?  Sure... >>>>1 >>>>http://www.digit-life.com/articles/ibmpower4/  >>>> >>>>4. POWER4 dieo >>>>G >>>>The POWER4 houses 2 processors each having an L1 cache for data and N >>>>instructions. The die has a single L2 cache of 1450 KBytes controlled by 3P >>>>separate controllers connected to the cores via a CIU (Core Interface Unit).O >>>>The controllers work independently and can process 32 bytes per clock. Each-R >>>>processor uses two separate 256-bit buses to connect the CIU for data fetchingR >>>>and data loading, as well as a separate 64-bit bus to save the results; the L2K >>>>cache has a bandwidth of 100 GBytes/s. The L2 cache's system looks wellRL >>>>balanced and very powerful. Each processor has a special unit to supportR >>>>noncachable operations (Noncacheable Unit). The L3 controller and the memory'sS >>>>one are located on die as well. For connection with the L3 cache working at 1/3 L >>>>of the processor's speed and with the memory there are two 128-bit busesO >>>>operating at 1/3 of the processor's frequency. The throughput of the memoryxS >>>>interface is about 11 GBytes/s. Data flows coming from the memory and L2 and L3yO >>>>caches and the buses of the chips are controlled by the Fabric Controller: s >>>> >>>> >>>> >>>> >>>	Big mistake. >>>UE >>>	The L2 cache is 1.5 MBytes, not "more than 5 MBytes" , an earliernB >>>	link discovered is combining all caches to lead me to believe   >>>	5.6 MBytes was/is the total. >>>2 >>>9 >># >>Ahh the dangers of cut and paste.@ >>2 >>So I guess thats the end of the argument thanks. >> >> >  > 	No. >     * So what tack are you going to try next ???   Regards0 Andrew Harrison1   ------------------------------  % Date: Mon, 01 Jul 2002 17:45:40 +0100_U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>M Subject: Re: TCO study0 Message-ID: <afq0u5$n61$1@new-usenet.uk.sun.com>   sasadmin wrote:e   > sms@antinode.org writes: >  > W >>From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>a >>, >>>His loss he might have learn't something. >>>dH >>   From Mr. Harrison, he's about as likely to learn how to construct a- >>sentence, or the proper use of apostrophes,e >> >>...i >> > M > Please (re-)read Bill The Quill. You might want to reconsider your Americanr > English chauvinism.  > Q > I agree Andrew may sometimes be over the top, but I welcome a dissenting P.O.V.t > K > That said, Andrew, do you have any remarks on the var Business editorial?7 >       8 The editor seemed to have forgotten what VAR stands for.   Value Added Reseller.P   1. u: Sun has actually been remarkably consistent in its channelB strategy, almost diametrically opposite to Digital, Compaq, HP andB IBM. If we have changed our channel strategy over the years it has? been as nothing when compared with the upheavals in other major @ vendors channels, a point picked up on by one of the responders.   2. o= The VAR channel is being squeezed not because Sun has changedtA its policy but because there is a recession and margins are beings? squeezed across the board. Blaming Sun for a worldwide downturn + in IT spending is hardly fair or plausible.c   3.  ? Sun has always expected its Value Added Resellers to add value,bF when times were good many resellers were able to just ship tin withoutE adding any value at all. Some resellers have not adjusted to the fact B that margins are down and shipping vanilla tin isn't adding value.  @ When you are competing against a vendor who is prepared to offer? systems at 70% or more discount directly in order to try to wino? business shipping through a channel and having to protect theirn? margins particularly if they arn't offering value add is hardlyh good business.   4. y< Sun has been very keen to allow resellers to have a peice ofE the action when it comes to services, something that HP, IBM etc have_? not. The downside of this is that Sun requires the resellers tokB demonstrate that they are competent to deliver the services, its aB fair exchange, complaining that certification is a problem ignores@ the downstream issues which Sun will ultimately have to pick up.  @ Paradoxically the editorial claims that resellers are laying off< staff that deliver these services and suggests a strategy of= cosying up to IBM, HP etc none of which allow their VAR's theh7 the same level of choice to do this in the first place.3  9 This point is well refuted by the owner of one of the Sunr# VAR's in a response to the article.      5. p3 Despite the article and claims to the contrary made>9 by certain vendors Sun's market share is improving, so ifU% you are a VAR what would you choose ?   8 None of this should be taken as a criticism of the VAR's8 who are struggling and who have exited the market, it is) tough out there and people are suffering.e   Regardsp Andrew Harrisone   ------------------------------  # Date: Mon, 01 Jul 2002 16:58:47 GMT31 From: "Terry C. Shannon" <terryshannon@attbi.com>d Subject: Re: TCO study, Message-ID: <bT%T8.56250$Uu2.9437@sccrnsc03>  # "Andrew Harrison SUNUK Consultancy"t> <andrew_nospam.harrison_remove_this@sun#.com> wrote in message* news:afq0u5$n61$1@new-usenet.uk.sun.com... >n >f > sasadmin wrote:t >s > > sms@antinode.org writes: > >  > >2+ > >>From: Andrew Harrison SUNUK Consultancy - <andrew_nospam.harrison_remove_this@sun#.com>  > >>. > >>>His loss he might have learn't something. > >>>eJ > >>   From Mr. Harrison, he's about as likely to learn how to construct a/ > >>sentence, or the proper use of apostrophes,a > >> > >>...> > >> > >hF > > Please (re-)read Bill The Quill. You might want to reconsider your American > > English chauvinism.w > >iL > > I agree Andrew may sometimes be over the top, but I welcome a dissenting P.O.V. > >rB > > That said, Andrew, do you have any remarks on the var Business
 editorial? > >: >: >n >n: > The editor seemed to have forgotten what VAR stands for. >u > Value Added Reseller.o >w > 1.< > Sun has actually been remarkably consistent in its channelD > strategy, almost diametrically opposite to Digital, Compaq, HP andD > IBM. If we have changed our channel strategy over the years it hasA > been as nothing when compared with the upheavals in other majoruB > vendors channels, a point picked up on by one of the responders.  G I've only kept an eye on Sun and its channels strategy since ~1993, but J based on what I've seen over the last decade Andrew speaks the truth here.H Contrast this with DEC's ~1994-5 efforts to pursue an indirect strategy.   ------------------------------  # Date: Mon, 01 Jul 2002 13:40:51 GMTc+ From: Roland Barmettler <rob@bbp.ch.remove> ' Subject: Re: Trouble with BACKUP/RECORDd7 Message-ID: <20020701154048.5fe0e687.rob@bbp.ch.remove>h   Hi all   Thanks to all for your help!  , Apparently, it is not a bug, it's a feature:R http://www.compaq.com/support/asktima/operating_systems/CHAMP_SRC000308003657.html   ;-)    Greetings, Roland   F --------------- bbp - Biveroni Batschelet Partners AG ----------------:              Bahnhofstrasse 28, CH-5401 Baden, SwitzerlandF ----------------------------------------------------------------------   ------------------------------  * Date: Mon, 1 Jul 2002 08:25:05 +0000 (UTC)* From: Osmo Kujala <kujala@tukki.cc.jyu.fi>% Subject: VMS IO up to the level of PC , Message-ID: <afp3l1$djd$1@mordred.cc.jyu.fi>  E When searching net for info about faster (affordable) SCSI adapters Ie2 found this by Patrick Young (P.Young@unsw.EDU.AU): ...nB > At my disposal I had an Adaptec AIC-7892 (29160 64 bit Ultra 160D > controller using SYS$PKADRIVER) and the Storageworks RAID software! > V2.4A for OpenVMS/Alpha (V7.3).K: > The drives do 41MB/Sec (tested reading) under OpenVMS...  F Does the card do 160MB/sec with VMS? Is there a speed penalty for not C being supported? I see that AIC-7895 and AIC-7899 are mentioned in iH V73-SYS$CONFIG. How about Compaq Smart Array 431? Could it work for VMS? It has AIC-7899G.iH There is also Adaptec 39160 with 7899G and I've seen mentioned that SRM 9 sees 39160 as KZPEA-DB. These have very much in common...n  G Our Alphas, which need faster SCSI are PWS433a(u), Alphaserver 4000 andp AlphaStation 5000.  H About net cards: Intel based DE600-cards have chips i82559,i82558. WhichF                  commodity intel-cards would work for VMS and what are@                  their names? (NC3123,NC3131... for Compaq PC:s)   Osmo   ------------------------------  % Date: Mon, 01 Jul 2002 07:58:43 -0500tC From: "Craig A. Berry" <craig.berry@nospam.SignalTreeSolutions.com> ) Subject: Re: VMS IO up to the level of PCYH Message-ID: <craig.berry-4FDB31.07584301072002@news.directvinternet.com>  , In article <afp3l1$djd$1@mordred.cc.jyu.fi>,,  Osmo Kujala <kujala@tukki.cc.jyu.fi> wrote:  J > There is also Adaptec 39160 with 7899G and I've seen mentioned that SRM ; > sees 39160 as KZPEA-DB. These have very much in common...a > I > Our Alphas, which need faster SCSI are PWS433a(u), Alphaserver 4000 anda > AlphaStation 5000.  H Careful.  The KZPEA is only supported on EV6 systems.  Dunno whether it : works despite the lack of support on older systems or not.   ------------------------------  # Date: Mon, 01 Jul 2002 14:36:40 GMTs From: lbohan@dbc..spamless..com1) Subject: Re: VMS IO up to the level of PCp8 Message-ID: <0vp0iukvmtjp0k4vt610rkrs3m8te57j2a@4ax.com>  4 On Mon, 1 Jul 2002 08:25:05 +0000 (UTC), Osmo Kujala <kujala@tukki.cc.jyu.fi> wrote:_  F >When searching net for info about faster (affordable) SCSI adapters I3 >found this by Patrick Young (P.Young@unsw.EDU.AU):  >...C >> At my disposal I had an Adaptec AIC-7892 (29160 64 bit Ultra 160tE >> controller using SYS$PKADRIVER) and the Storageworks RAID softwareu" >> V2.4A for OpenVMS/Alpha (V7.3).; >> The drives do 41MB/Sec (tested reading) under OpenVMS...  >bG >Does the card do 160MB/sec with VMS? Is there a speed penalty for not oD >being supported? I see that AIC-7895 and AIC-7899 are mentioned in I >V73-SYS$CONFIG. How about Compaq Smart Array 431? Could it work for VMS?o >It has AIC-7899G.I >There is also Adaptec 39160 with 7899G and I've seen mentioned that SRM e: >sees 39160 as KZPEA-DB. These have very much in common... >uH >Our Alphas, which need faster SCSI are PWS433a(u), Alphaserver 4000 and >AlphaStation 5000.- >wI >About net cards: Intel based DE600-cards have chips i82559,i82558. WhichPG >                 commodity intel-cards would work for VMS and what areoA >                 their names? (NC3123,NC3131... for Compaq PC:s)r >U >Osmoc  B I've used the following in DS20/DS20E'S, and a PC164 based system.= (DE600-AA).  if you do network boots, you'll probably need toC) double check your systems; firmware rev. 2  P http://www.cc-inc.com/macmall/shop/detail.asp?action=techspec&dpno=805059#scroll  ) NC3123 Fast Ethernet NIC PCI 10/100 WOL	   Platform	Universal	  Part No.	805059	 a Manufacturer	Compaq	 9 Mfr. Part No.	174830-B21	  S Your Price:	  $100.00	   ------------------------------  # Date: Mon, 01 Jul 2002 07:44:18 GMTt1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie) - Subject: Re: VMS Marketing Volunteers, ctd...t: Message-ID: <mLTT8.20448$p85.583728@twister.austin.rr.com>  ' Jay E. Morris (jem@epsilon3.com) wrote:r : J : Back when Network Solutions was the only game in town (US side anyway), B : they would not sign up a .org unless you could prove you were a F : non-profit (on purpose).  Since the loosening of the reins that has  : dropped by the wayside.a : - : Yes, I tried to get a .org and was told no.  : ? This domain was registered back in 1995 to an individual whose .E cat's name  is Kamikaze, and who was not a non-profit organization...r  3    whois -h whois.networksolutions.com kamikaze.org       Registrant:.    Kamikaze Feline Enterprises (KAMIKAZE2-DOM)    c/o MTD Media LLC    4423 Lehigh Road 113     College Park, MD 20740e      Domain Name: KAMIKAZE.ORG      Administrative Contact:8       MTD Media Administrator  (MMA6-ORG)		admin@MTD.COM       MTD Mediab       4431 Lehigh Road 113       College Park, MD 20740	       USAb       301 854 9635       Fax- {junk}     Technical Contact:s:       MTD Media Hostmaster  (MMH5-ORG)		hostmaster@MTD.COM       MTD Median       4431 Lehigh Road 113       College Park, MD 20740       US       301 854 9635       Fax- {junk}o  !    Record expires on 21-Nov-2002.i!    Record created on 20-Nov-1995.T4    Database last updated on 1-Jul-2002 03:32:03 EDT.  "    Domain servers in listed order:  ,    NSF.ALGX.NET                 64.245.20.14,    NSE.ALGX.NET                 64.245.43.14  < WARNING: Going to this web site may cause problems with your          local Net-Nazis.s      http://kamikaze.org/l    A  2 --Jerry Leslie   (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for emaild   ------------------------------  % Date: Mon, 01 Jul 2002 12:32:53 +0100 ( From: Nic Clews <sendspamhere@127.0.0.1>' Subject: Re: when is a typo not a typo? ( Message-ID: <3D203DE5.CDFF1CC@127.0.0.1>   Paul Sture wrote:  > b > In article <afi7b6$2nmc$1@info.cs.uofs.edu>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:7 > > In article <oNjUYmkWHxSX@eisner.encompasserve.org>, 4 > >  koehler@encompasserve.org (Bob Koehler) writes: > > |>- > > |>    Stuff this is you favorite browser:e > > |> > > |>    www.openvms.copaq.com. > > |> > > |>    That's right: copaq. > > |> > >r/ > > I got nothing.  Couldn't connect to server.y > >eP > I got a connection timeout. The server certainly exists and is registered to a > private individual.a  9 I got some cycle thing on Friday. This morning - nothing.   @ It is probably the "copaq.com" domain, and the [a] DNS server is5 matching any domains below it in hierarchical order. J   -- g? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencesi nclews at csc dot coms   ------------------------------  $ Date: Mon, 1 Jul 2002 09:57:11 -0400* From: WILLIAM WEBB <WWEBB1@email.usps.gov> Subject: RE: wow- Message-ID: <0033000070855078000002L082*@MHS>,  ) =0AI *told* you that HP was listening....:   :^)    WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNET>% Sent: Tuesday, June 25, 2002 10:01 AMRB To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET Subject: RE: wow     "John N." wrote: >4H > I am going to go have another cup of coffee, and come back and read t= hisaH > whole article.  I am not sure I read it right and I don't want to get=  myc > hopes up:p >c > From yahoo news: >rE > HP Unveils Enhanced OpenVMS Operating System that Is e-Business and  > Internet-Ready. > http://biz.yahoo.com/bw/020625/250215_1.html   no --- WWWOOWWWW!!!!!e   -- Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163t7 Sr. Consultant            http://www.firstdbasource.comtE                           http://www.firstdbasource.com/donation.htmll0 704-947-1089 (Office)     704-236-4377 (Mobile)=   ------------------------------   End of INFO-VAX 2002.359 ************************