1 INFO-VAX	Sat, 06 Sep 2003	Volume 2003 : Issue 493       Contents: Re: Do we need a DCL debugger? Re: Do we need a DCL debugger? Re: Do we need a DCL debugger? Re: Do we need a DCL debugger? Re: Do we need a DCL debugger? Re: HP *ALMOST* Advertised VMS% Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived % Re: New Itanium sales figures arrived ' Re: OpenVMS Pearl - Enterprise printing 9 Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1 9 Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1 9 Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1  RDB Interactive SQL questions  Re: RdB on EVA ?* Slow keyboard autorepeat rate for new DS10 VMS on the trading floor1 Re: Will OpenVMS I64 run on a Dell PowerEdge 3250 1 Re: Will OpenVMS I64 run on a Dell PowerEdge 3250 $ Re: [Q] Learning to use Queues in C.  F ----------------------------------------------------------------------  % Date: Fri, 05 Sep 2003 15:28:50 -0500 * From: Patrick Spinler <pspinler@yahoo.com>' Subject: Re: Do we need a DCL debugger? ( Message-ID: <3F58F202.3060006@yahoo.com>   david20@alpha1.mdx.ac.uk wrote: Z > In article <bjab0l$65f$1@tribune.mayo.edu>, Patrick Spinler <pspinler@yahoo.com> writes: > I >>At least this particular ucx ftp server doesn't appear to support NLST:  >> >>  $ ucx show version >>C >>    Compaq TCP/IP Services for OpenVMS Alpha Version V5.3 - ECO 2 < >>    on a AlphaServer 4100 5/600 8MB running OpenVMS V7.3-1 >> >>  $ ftp localhost  >>  FTP> quote nlst  > F > You are confusing the FTP client commands such as get, put, ls  with' > the underlying FTP protocol commands. K > LIST and NLST are the underlying protocol commands - EVERY FTP server and 2 > CLIENT has to support those mandated by RFC 959. >   H Um, not really.  The 'quote' command sends the following text as an FTP  protocol command.   G With some FTP clients, the QUOTE command is the only way I can do some   actions on our VMS servers.    FTP> help quote    QUOTE   D       Sends your input directly to the remote host. Lets you use FTPG       commands that are implemented by the remote host but not known to        the local host.    -- Pat   ------------------------------   Date: 5 Sep 2003 16:13:33 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) ' Subject: Re: Do we need a DCL debugger? 3 Message-ID: <3gQukv+E3Iyh@eisner.encompasserve.org>   X In article <bjab0l$65f$1@tribune.mayo.edu>, Patrick Spinler <pspinler@yahoo.com> writes: > I > At least this particular ucx ftp server doesn't appear to support NLST:  >  >    FTP> quote nlst8 >    425-Can't build data connection for 127.0.0.1,54446+ >    425 connect to network object rejected   F    OK, it's been a few months since I got into that RFC.  Are you sure    it isn't case sensitive?   I    Also, does it respond with the same results to the client commands ls  H    and dir?  Typically clients send NLST for one and LIST for the other.  F    I never had UCX fail to respond to both ls and dir, but I never had    the version you have.   ------------------------------  * Date: Fri, 5 Sep 2003 21:23:28 +0000 (UTC) From: david20@alpha1.mdx.ac.uk' Subject: Re: Do we need a DCL debugger? ) Message-ID: <bjausg$p0h$1@news.mdx.ac.uk>   U In article <3F58F202.3060006@yahoo.com>, Patrick Spinler <pspinler@yahoo.com> writes:   >david20@alpha1.mdx.ac.uk wrote:[ >> In article <bjab0l$65f$1@tribune.mayo.edu>, Patrick Spinler <pspinler@yahoo.com> writes:  >>  J >>>At least this particular ucx ftp server doesn't appear to support NLST: >>>  >>>  $ ucx show version  >>> D >>>    Compaq TCP/IP Services for OpenVMS Alpha Version V5.3 - ECO 2= >>>    on a AlphaServer 4100 5/600 8MB running OpenVMS V7.3-1  >>>  >>>  $ ftp localhost >>>  FTP> quote nlst >>  G >> You are confusing the FTP client commands such as get, put, ls  with ( >> the underlying FTP protocol commands.L >> LIST and NLST are the underlying protocol commands - EVERY FTP server and3 >> CLIENT has to support those mandated by RFC 959.  >>   > I >Um, not really.  The 'quote' command sends the following text as an FTP   >protocol command. >   < Sorry you are correct about quote sending protocol commands.  - However all I can say is that not supporting    
 quote nlst or
 quote list   is fairly common  H For instance  from my Suse Linux system to an FTP server on a SUN system   ftp> open gold.mdx.ac.uk Connected to 158.94.5.209.& 220 gold FTP server (SunOS 5.6) ready.$ Name (gold.mdx.ac.uk:david): david20" 331 Password required for david20.	 Password:  230 User david20 logged in.  Remote system type is UNIX. $ Using binary mode to transfer files. ftp> quote nlst 4 425 Can't build data connection: Connection refused. ftp> quote list 4 425 Can't build data connection: Connection refused.  K I think quote is usually implemented so that it can only be used for those  0 protocol commands which are optional in the RFC.    
 David Webb VMS and Unix team leader CCSS Middlesex University      H >With some FTP clients, the QUOTE command is the only way I can do some  >actions on our VMS servers. >  >FTP> help quote >  >QUOTE > E >      Sends your input directly to the remote host. Lets you use FTP H >      commands that are implemented by the remote host but not known to >      the local host. >  >-- Pat  >    ------------------------------  # Date: Fri, 05 Sep 2003 22:11:48 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")' Subject: Re: Do we need a DCL debugger? 6 Message-ID: <00A25748.7DECA7E4@SSRL.SLAC.STANFORD.EDU>  f In article <7K16b.36009$Nc.8317328@news1.news.adelphia.net>, Colin Blake <colin@theblakes.com> writes: >Bob Koehler wrote:  > F >>>Starting with 1.5 beta, Mozilla (on all platforms) will talk to an  >>>OpenVMS FTP server. >>>      >>>  >>1 >>   "an OpenVMS FTP server"?  Which one?  (UCX?)  >>B >I tested with TCP/IP Services for OpenVMS, TCPware, and MultiNet. >   K I want to take this opportunity to express my appreciation for Colin's work K on Mozilla.  It's _great_ to have VMS releases (at mozilla.org) at the same B time as those for other platforms, and to also have the option of K officially-supported SWB releases at the same time.  He's also made a point H of not privileging UCX / TCP/IP Services over other IP products for VMS.  L Finally, if he's responsible for introducing support for OpenVMS FTP serversN on _all Mozilla platforms_, he's done VMS webmasters a great service.  (And ifF he's just responsible for telling us about it, that's not bad either.)   Thanks, Colin!   -- Alan  --  O =============================================================================== 0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025 O ===============================================================================    ------------------------------  % Date: Fri, 05 Sep 2003 17:54:07 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>' Subject: Re: Do we need a DCL debugger? ( Message-ID: <3F5905E8.AEC35F5@istop.com>   > >    FTP> quote nlst: > >    425-Can't build data connection for 127.0.0.1,54446- > >    425 connect to network object rejected   ( For the VMS (TCPIP services) FTP server:  * 214-The following commands are recognized:@    USER    TYPE    RETR    RNFR    NLST    PWD     ALLO    EPSV @    PASS    STRU    STOR    RNTO    CWD     CDUP    SYST    QUIT @    SITE    PORT    STOU    DELE    MKD     NOOP    STAT    HELP 8    MODE    EPRT    APPE    LIST    RMD     ABOR    PASV  214 End of Help.  E Now, in the case of NLST and many others, it requires that the client M cooperate because the outpit isn't sent on the "control channel" (not sure of H the actual terminology) but on a separate link established by the remote server to the client.   J When the client software issues the NLST command, it knows to listen for aG connect request for the "data" connection. But when you issue the quote M command, the client has no idea what you are doing and doesn't know to expect N the remote server to try to connect to it to dump some data, hence, the server compains it cannot connect.    ------------------------------  # Date: Sat, 06 Sep 2003 00:11:41 GMT # From: "John Smith" <a@nonymous.com> ' Subject: Re: HP *ALMOST* Advertised VMS H Message-ID: <1D96b.169555$_V.67196@news04.bloor.is.net.cable.rogers.com>  G "Peter Weaver" <WeaverConsultingServices@sympatico.ca> wrote in message 3 news:bja67t$gsjc4$1@ID-141708.news.uni-berlin.de... ? > This, http://h71000.www7.hp.com/usps.html, had me excited for > > a few minutes, until I checked out every link on the page to@ > find that VMS did not appear on any of them (unless you create? > a cut-out that blocks out everything except the second letter ? > of the first word, the first letter of the ninth word and the 4 > last letter of the ninth word on the two-page ad). > ; > Maybe we should all print a few million stickers that say ? > "Hey, they're talking about VMS here!" and hit every magazine < > stand in our areas inserting the sticker in every magazine > they run the ad in?     $ As seen in my post of August 26th...  C I was watching a program on CNBC this evening (6:30-7:00pm Eastern)    CNBC's CheckPoint = http://moneycentral.msn.com/content/CNBCTV/TV_Info/P37413.asp   C "As the war on terrorism continues, the office of Homeland Security > evolves and the geopolitical situation remains in flux, CNBC's? CheckPoint" examines the business side of security and war, and A analyzes the impact of increased national, corporate and personal D security on the economy, corporate America and individual investors.  F "CNBC's CheckPoint" is a limited-run program that features insight andF perspective from military and security experts. The experts shed lightF on the intelligence community's efforts against terrorism, at home and@ overseas. The program also features interviews with business andE government leaders as they face the challenge of increasing security.  "     * and they had a story on disaster recovery.  B It talked of how experts in the light of terrorism and natural andE man-made disasters (ie. electricity grid failures, etc..) recommended E that corporations maintain data centers at least 400 miles apart. And < that they spread their operations across multiple buildings.  E They had a talking head on from Kroll, the risk management company to C the Global 1000, http://www.krollworldwide.com/,  talking about the @ pressing need for corporations of all sizes to do these sorts of things.   C The piece also talked of how IBM has built nearly somewhere between E 100-200 multi-vendor recovery sites around the world. And they showed C video of racks of Compaq-branded Alphaservers inside these IBM data  centers.    F Of course the ads for VMS clusters were nowhere to be seen on TV or in@ print since the great blackout of 2003....or before that either.  F I can understand the 'queasiness' HP might feel about using the WTC in> an ad about disaster recovery, but to not have had VMS clusterF advertising in the business media by now using the blackout as part ofD the storyline is just **massively ignorant** on the part of everyone2 associated with HP's management and marketing, and2 **counterproductive** to HP shareholder interests.    F A simple ad would suffice in the Wall Street Journal, Forbes, Business
 Week etc.....   F Show a map of the USA and Canada, and/or Europe, to scale. Label major cities. ( Show the blackout area as a shaded area.1 Have a triangle with sides to scale of 500 miles.    The instructions read:  / 1) Cut out the triangle and overlay on the map. / 2) Locate your primary data center at one apex. < 3) Locate your backup and quorum sites at the other corners.% 4) Compare what other vendors can do. < 5) Install HP Alphaservers running OpenVMS at each location. 6) Stay in business.  F OpenVMS clusters fit almost every budget, from two small machines withF one processor each through to clusters of thousands of processors with" dynamic partitioning capabilities.  > No matter which business you are in, disaster-tolerant OpenVMS@ clusters ensure that the lights don't permanently go out on your company.  D We wrote the book on clusters. Everyone else is just learning how to read.   * (c) 2003, John Smith. ALL RIGHTS RESERVED.    A HP - contact me off-line for the BIC code for my Caymans account.    ------------------------------  $ Date: Fri, 5 Sep 2003 12:21:54 -0700* From: "Jack Peacock" <peacock@simconv.com>. Subject: Re: New Itanium sales figures arrived2 Message-ID: <I9CdnUCqw7vif8WiU-KYgw@mpowercom.net>  4 "David Svensson" <icerq4a@spray.se> wrote in message7 news:734da31c.0309050819.7853ee98@posting.google.com...  > 7 > HP most likely advertise Itanium more than the others  > D A good point.  However, I don't recall seeing any Itanium ads, HP orJ otherwise.  Where is HP advertising these systems?  Perhaps the same venue as all those VMS ads?     Jack Peacock    ------------------------------  $ Date: Fri, 5 Sep 2003 16:57:46 -0400* From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: New Itanium sales figures arrived2 Message-ID: <pzudnU0BJt5KZcWiXTWJiw@metrocast.net>  4 "David Svensson" <icerq4a@spray.se> wrote in message7 news:734da31c.0309050819.7853ee98@posting.google.com... ; > mistdragon@zdnetonebox.com (mist dragon) wrote in message 9 news:<7500353b.0309050156.4509294a@posting.google.com>...    ...   0 > > As for other amusing things : This slideshowH > > http://www.openvms.org/stories.php?story=03/08/29/3410320 shows thatG > > HP owns 80+ percentage of the Itanium market. In the perspective of G > > this thread this means either that 80+ means nearly 100% or that if J > > its 80% and those 74 boxes are 20%, meaning that the actual sales were > > 296 boxes :) > D > I don't see how we can relate those figures to the IDC Q2 figures.  J I believe that was his point:  that the '80+%' figure appears inconsistentL with the IDC ship numbers.  They might be reconcilable if the 80+% refers toF processor count (the IDC numbers are for systems):  HP didn't ship anyL system with more than 4 processors during the period in question, while SGI,J NEC, and Bull were shipping larger systems (even though very few of them).   > G > > One may question why Itanium sells +3000 boxes in HP and only 74 in 4 > > other vendors. Is it because HP is much better ?   No apparent reason to think so.     This is standard  > > iron, so it shouldnt be.  E As has already been observed, this rather stretches the definition of H 'standard'.  However, even if one accepts that, the surrounding chipsets; differ - let alone other aspects of the systems as a whole.   +  Or is it because HP is giving them free to H > > developers (figure for shipped units != units sold to customers). Or > > is it because of VMS ?  L Rather clearly not:  not even the earliest Itanic VMS release (that aimed atB the most aggressive early developers) appeared in time to make any difference in the numbers.  /  It would be interested to get official figures I > > from HP directly, but I doubt they'll give them and for good reasons.  > 8 > HP most likely advertise Itanium more than the others,  I Perhaps.  And HP's products have nominally been available for over a year I now, while NEC's appeared only late last year, SGI's early this year, and K Dell's only a couple of months ago.  Even Intel's own Itanic chipset didn't E seem to become available until sometime this year (perhaps only quite G recently).  So people have had considerably more time to come around to D deciding to buy HP's hardware than anyone else's, and that more thanL anything else may account for most of HP's relatively higher ship rate (such
 as it is).    and HP's boxes  > also have better chip-sets.   F In the period in question, HP's boxes have only one chipset:  the zx1.D While that received good initial reviews, they were in part based onL erroneous memory latency figures (the original claim of 112 ns. was based onK a flawed measurement, and the actual number turned out to be 154 ns. - plus C another 25 ns. in both cases if you use more than 16 GB of memory).   L When using the same compilers (certainly Intel, and perhaps gcc - though I'mH not sure direct comparisons with the latter are available), SGI's ItanicJ products appear to match the zx1 chipset in performance (despite SGI's far greater scalability).   $  I think customers today buy Itanium> > systems to run HP-UX in the first place, then Windows/Linux.  G That would be kind of interesting to know.  I'm not acquainted with any I low-end (4-processor and under) PA-RISC products:  if they actually don't H exist, then Itanic would fill a hole in that space; otherwise, I'd guessA people might be somewhat slow to switch (just as seems to be true I elsewhere).  My impression is that some interest exists in the Linux area L (though whether it would if Alpha weren't on the way out is questionable, asL is the question of whether it will persist given the increasing availabilityI of far more cost-effective AMD64 platforms).  And I guess there are a few 8 people who want to get their feet wet in 64-bit Windows.  I But at the current ship rates, the distribution probably doesn't say much ; about whatever potential market may eventually materialize.    - bill   ------------------------------  # Date: Fri, 05 Sep 2003 21:24:23 GMT & From: Rick Jones <foo@bar.baz.invalid>. Subject: Re: New Itanium sales figures arrived2 Message-ID: <ba76b.4144$Xd1.3039@news.cpqcorp.net>  ) Bill Todd <billtodd@metrocast.net> wrote: B > In the period in question, HP's boxes have only one chipset: theC > zx1.  While that received good initial reviews, they were in part F > based on erroneous memory latency figures (the original claim of 112E > ns. was based on a flawed measurement, and the actual number turned B > out to be 154 ns. - plus another 25 ns. in both cases if you use > more than 16 GB of memory).   B The additional latency is when you use memory extenders (ie in theD rx5670) and is not _necessarily_ a function of memory quantity.  ForF example, with the 2GB DIMMs these days, an rx2600 (which does not have% the extenders) can have 24 GB of RAM.   I > That would be kind of interesting to know.  I'm not acquainted with any K > low-end (4-processor and under) PA-RISC products:  if they actually don't  > exist   D Four CPU and fewer PA-RISC products do indeed exist and can be found? on http://www.hp.com/ - look for the rp54XX's and the rp24XX's.   
 rick jones --  . a wide gulf separates "what if" from "if only"F these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...    ------------------------------  $ Date: Fri, 5 Sep 2003 17:41:50 -0400* From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: New Itanium sales figures arrived2 Message-ID: <ps6dnakHwcu_nsSiXTWJkA@metrocast.net>  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message # news:3F583A31.B500FC89@istop.com...  > Bill Todd wrote:J > > HP was selling its McKinley systems for the full year under discussion > > (July, 2002 - June, 2003). > C > But had it made some serious marketing launch for those systems ?    Were you asleep 14 months ago?   ...      It wasn't because theyL > > lacked the cachet (so to speak) of a Microsoft OS:  stable 64-bit Itanic; > > Windows releases have existed for close to 2 years now.  >  > Are you sure about that ?   
 Of course.  J > What version of Windows is available as a full 64 bit implementation for IA64 ?  L Over 2 years ago, people had to use "Windows XP Professional 64-bit Beta 2".G IIRC, a bit under 2 years ago "Windows Advanced Server Limited Edition" K appeared (and subsequently got updated to its '1.2' version).  For close to G a year now, TPC-C submissions have used "Windows Server 2003 Enterprise K Edition" (a.k.a. "Windows .NET Enterprise Server 2003") and "Windows Server I 2003 Datacenter Edition" (which I think may be the equivalent version for  larger processor counts).    - bill   ------------------------------  $ Date: Fri, 5 Sep 2003 17:47:21 -0400* From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: New Itanium sales figures arrived2 Message-ID: <oK6dnaYFcuXtmcSiXTWJhQ@metrocast.net>  3 "Rick Jones" <foo@bar.baz.invalid> wrote in message , news:ba76b.4144$Xd1.3039@news.cpqcorp.net...+ > Bill Todd <billtodd@metrocast.net> wrote: D > > In the period in question, HP's boxes have only one chipset: theE > > zx1.  While that received good initial reviews, they were in part H > > based on erroneous memory latency figures (the original claim of 112G > > ns. was based on a flawed measurement, and the actual number turned D > > out to be 154 ns. - plus another 25 ns. in both cases if you use > > more than 16 GB of memory).  > D > The additional latency is when you use memory extenders (ie in theF > rx5670) and is not _necessarily_ a function of memory quantity.  ForH > example, with the 2GB DIMMs these days, an rx2600 (which does not have' > the extenders) can have 24 GB of RAM.   J Though I don't recall Gordon Haff making that distinction in his excellentF chipset article, it isn't that surprising.  Are the 'memory extenders'I primarily ways to increase fanout to drive additional memory slots, or do # they do something more interesting?    > K > > That would be kind of interesting to know.  I'm not acquainted with any G > > low-end (4-processor and under) PA-RISC products:  if they actually  don't 	 > > exist  > F > Four CPU and fewer PA-RISC products do indeed exist and can be foundA > on http://www.hp.com/ - look for the rp54XX's and the rp24XX's.   D Doh - I should have remembered that:  I've seen them often enough in  scanning the SPECint/fp results.   Thanks,    - bill   ------------------------------  % Date: Fri, 05 Sep 2003 15:08:43 -0700 % From: Greg Cagle <news@gregcagle.com> . Subject: Re: New Itanium sales figures arrived/ Message-ID: <vli2bgp2kqm563@corp.supernews.com>    Bill Todd wrote:  5 > "Rick Jones" <foo@bar.baz.invalid> wrote in message . > news:ba76b.4144$Xd1.3039@news.cpqcorp.net...  F >>Four CPU and fewer PA-RISC products do indeed exist and can be foundA >>on http://www.hp.com/ - look for the rp54XX's and the rp24XX's.  >  > F > Doh - I should have remembered that:  I've seen them often enough in" > scanning the SPECint/fp results.  > Don't forget the workstation variants - the j6xxx machines areB very fast dual processor compute engines in a rackable 2U package,( and are a lot cheaper than the rp24xx's.   --  
 Greg Cagle gregc at gregcagle dot com   ------------------------------  % Date: Fri, 05 Sep 2003 18:02:55 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>. Subject: Re: New Itanium sales figures arrived) Message-ID: <3F5907F7.EBB4A66F@istop.com>    Bill Todd wrote:E > > But had it made some serious marketing launch for those systems ?  >   > Were you asleep 14 months ago?  M I recall much hype from carly & Company about how great IA64 was going to be, J but I don't recall any product specific marketing by HP. The first branded/ products I am aware of is the integrity series.   I I realise that HP did have a few token systems available for purchase (by M developpers mostly) but I don't recall any branding or specific marketing for  those systems.  N > Over 2 years ago, people had to use "Windows XP Professional 64-bit Beta 2".I > IIRC, a bit under 2 years ago "Windows Advanced Server Limited Edition" M > appeared (and subsequently got updated to its '1.2' version).  For close to I > a year now, TPC-C submissions have used "Windows Server 2003 Enterprise M > Edition" (a.k.a. "Windows .NET Enterprise Server 2003") and "Windows Server K > 2003 Datacenter Edition" (which I think may be the equivalent version for  > larger processor counts).     M Are all these versions of windows for IA64 full 64 bits, or are some like VMS > was initially on Alpha; a 32 bit OS running on a 64 bit chip ?   ------------------------------  # Date: Fri, 05 Sep 2003 23:18:11 GMT & From: Rick Jones <foo@bar.baz.invalid>. Subject: Re: New Itanium sales figures arrived0 Message-ID: <TQ86b.4152$th1.29@news.cpqcorp.net>  ) Bill Todd <billtodd@metrocast.net> wrote: B > Though I don't recall Gordon Haff making that distinction in his6 > excellent chipset article, it isn't that surprising.  C Perhaps 2GB DIMM's were not around at the time of his article, only D 1GB's, in which case > 16GB and "use memory extenders" may have been
 synonymous.     A > Are the 'memory extenders' primarily ways to increase fanout to = > drive additional memory slots, or do they do something more  > interesting?  C I do not know for a fact, but have always ass-u-me-d that they were F for increasing the number of DIMM slots and thus total RAM capacity in the system.   
 rick jones --  G oxymoron n, commuter in a gas-guzzling luxury SUV with an American flag F these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...    ------------------------------  # Date: Sat, 06 Sep 2003 00:01:33 GMT 3 From: Robert Klute <robert_klute_removethis@hp.com> . Subject: Re: New Itanium sales figures arrived8 Message-ID: <7j8ilvcq30dp4j98ej47kvk014lt9mdmgl@4ax.com>  B On Fri, 05 Sep 2003 23:18:11 GMT, Rick Jones <foo@bar.baz.invalid> wrote:  * >Bill Todd <billtodd@metrocast.net> wrote:C >> Though I don't recall Gordon Haff making that distinction in his 7 >> excellent chipset article, it isn't that surprising.  > D >Perhaps 2GB DIMM's were not around at the time of his article, onlyE >1GB's, in which case > 16GB and "use memory extenders" may have been  >synonymous.   > B >> Are the 'memory extenders' primarily ways to increase fanout to> >> drive additional memory slots, or do they do something more >> interesting?   F The 'memory extenders' reduce signal load and thus allow more dimms to! be 'hung' off the memory I/O bus.    > D >I do not know for a fact, but have always ass-u-me-d that they wereG >for increasing the number of DIMM slots and thus total RAM capacity in  >the system.   From the HP Web site:    Block diagram of 2-way: < http://www.hp.com/products1/itanium/chipset/2-way_block.html Block diagram of 4-way: < http://www.hp.com/products1/itanium/chipset/4_way_block.html ZX1 FAQ:4 http://www.hp.com/products1/itanium/chipset/FAQ.html
 From the FAQ: @ HP zx1 Scalable Memory Expander is an optional component used toE increase memory capacity (up to four times) and increase bandwidth to G the main memory to 12.8 GB/s. Acting like as a memory hub, it decreases B the number of signal loads on the memory bus, thereby allowing theF system to dial up its memory transfer rate. The trade-off of using the@ zx1 memory expander is the cost of the chips and their footprintD together with a modest 25ns of additional memory latency relative to direct attach.     ------------------------------  # Date: Sat, 06 Sep 2003 00:54:03 GMT  From: dada <sman1@stny.rr.com>0 Subject: Re: OpenVMS Pearl - Enterprise printing+ Message-ID: <3F59311D.D03515E9@stny.rr.com>    Dean Woodward wrote:   > Sue Skonetski wrote: > F > > The customer does use DCPS and does consider any information about1 > > their environment as a competitive advantage.  > 5 > VMS is a "competitive advantage". Whodathunkit? ;-)   A Those of us who have developed real, functioning, higly available E applications have known this for years, however try and convince your G boss to convert from buggy/lossy unix or windows and see what happens !    ------------------------------   Date: 5 Sep 2003 11:31:17 -0700 ( From: bob@instantwhip.com (Bob Ceculski)B Subject: Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1= Message-ID: <d7791aa1.0309051031.456a150f@posting.google.com>   q Doc.Cypher <Use-Author-Address-Header@[127.1]> wrote in message news:<20030905074523.8635.qmail@gacracker.org>... : > On 4 Sep 2003, bob@instantwhip.com (Bob Ceculski) wrote:8 > >Nic Clews <sendspamhere@[127.0.0.1]> wrote in message& > >news:<bhg2a4$4aa$1@lore.csc.com>... > >> "Doc.Cypher" wrote: > >> >  G > >> > On Thu, 14 Aug 2003, Nic Clews <sendspamhere@[127.0.0.1]> wrote:  > >> > >Nic Clews wrote: >   J > >> > >The object libraries linked OK on [Alpha] 7.2-2 as well (I did notN > >> > >attempt a build). It provides support for up to 8192 bit key length as5 > >> > >well as an impressive array of compatibility.  > >> >  P > >> > Disastry reported a problem verifying signatures on keys when he built it4 > >> > here, I don't know if that was ever resolved. > >>  I > >> If someone can reproduce the problem or knows more about the issues, - > >> I'll give it a shot with this version...  > > @ > >I would like to compile this under c 6.0 on alpha 7.1-1h2 ... > >will this work? >  > Probably.  > M > Grab a copy from http://vmsbox.cjb.net/pgp/, unzip and look for the command  > file.  > G > It'll give a slew of warnings but you should get a usable executable.  > N > I'll happily host a more friendly build if someone has one they've worked onL > and are happy to share.  It'll go nicely with something I rescued with the; > WayBack machine - http://vmsbox.cjb.net/HowPGPWorks.html.  >  >  > Doc.  B well, it came with a pgp.exe already there dated mar 2002 ... what> version of vms (and machine? alpha/vax?) was this compiled on?  D I went thru the pgpinstal questions and accepted the cc redefinition@ default (cc /vaxc?) and only got 4 warnings on the compile aboutD multiple defined objects in deccshr ... now how do I encrypt a file?# the doc.txt files (2) are empty! :(    ------------------------------   Date: 5 Sep 2003 12:28:15 -0700 ( From: bob@instantwhip.com (Bob Ceculski)B Subject: Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1= Message-ID: <d7791aa1.0309051128.6191146c@posting.google.com>   q Doc.Cypher <Use-Author-Address-Header@[127.1]> wrote in message news:<20030905074523.8635.qmail@gacracker.org>... : > On 4 Sep 2003, bob@instantwhip.com (Bob Ceculski) wrote:8 > >Nic Clews <sendspamhere@[127.0.0.1]> wrote in message& > >news:<bhg2a4$4aa$1@lore.csc.com>... > >> "Doc.Cypher" wrote: > >> >  G > >> > On Thu, 14 Aug 2003, Nic Clews <sendspamhere@[127.0.0.1]> wrote:  > >> > >Nic Clews wrote: >   J > >> > >The object libraries linked OK on [Alpha] 7.2-2 as well (I did notN > >> > >attempt a build). It provides support for up to 8192 bit key length as5 > >> > >well as an impressive array of compatibility.  > >> >  P > >> > Disastry reported a problem verifying signatures on keys when he built it4 > >> > here, I don't know if that was ever resolved. > >>  I > >> If someone can reproduce the problem or knows more about the issues, - > >> I'll give it a shot with this version...  > > @ > >I would like to compile this under c 6.0 on alpha 7.1-1h2 ... > >will this work? >  > Probably.  > M > Grab a copy from http://vmsbox.cjb.net/pgp/, unzip and look for the command  > file.  > G > It'll give a slew of warnings but you should get a usable executable.  > N > I'll happily host a more friendly build if someone has one they've worked onL > and are happy to share.  It'll go nicely with something I rescued with the; > WayBack machine - http://vmsbox.cjb.net/HowPGPWorks.html.  >  >  > Doc.  B neat!  I took a stream_lf rms file, then did after creating my PGP foreign command ...    $ PGP -C file.txt   F then type in my secret phrase and presto!  I then reversed the process< and did a difference on the unencrypted file and they match!  D one question ... when I created a 1024 private key it created a file= called RANDSEED.BIN ... can I rename this to whatever I want?   ? question two ... is the version on your server Doc the same one 7 that HP has out for 7.2-7.3 or is this a different one?    ------------------------------   Date: 5 Sep 2003 12:29:45 -0700 ( From: bob@instantwhip.com (Bob Ceculski)B Subject: Re: pgp 2.6.3ia multi05 finally running on OpenVMS V7.3-1= Message-ID: <d7791aa1.0309051129.4a88695b@posting.google.com>   U Nic Clews <spamhere@[127.0.0.1]> wrote in message news:<bj9n1o$h1n$1@lore.csc.com>...  > Doc.Cypher wrote:  > < > > On 4 Sep 2003, bob@instantwhip.com (Bob Ceculski) wrote: > > 9 > >>Nic Clews <sendspamhere@[127.0.0.1]> wrote in messageS' > >>news:<bhg2a4$4aa$1@lore.csc.com>...  > >> > >>>"Doc.Cypher" wrote: > >>>bF > >>>>On Thu, 14 Aug 2003, Nic Clews <sendspamhere@[127.0.0.1]> wrote: > >>>> > >>>>>Nic Clews wrote:e > >>>  > >>>  > >>>gI > >>>>>The object libraries linked OK on [Alpha] 7.2-2 as well (I did not M > >>>>>attempt a build). It provides support for up to 8192 bit key length as 4 > >>>>>well as an impressive array of compatibility. > >>>>O > >>>>Disastry reported a problem verifying signatures on keys when he built it 3 > >>>>here, I don't know if that was ever resolved.e > >>> I > >>>If someone can reproduce the problem or knows more about the issues, - > >>>I'll give it a shot with this version...e > B > There is a problem with the keys. I've not had time to properly J > investigate it, despite claiming 8192, when trying to use the multi5 to G > work with an existing keyring (and check the keys) or generate a new  6 > key, it complains as you describe. This is at 7.2-2. > @ > Bob, if you have any luck or similar experiences, let me know.  B neat!  I took a stream_lf rms file, then did after creating my PGP foreign command ...-   $ PGP -C file.txt-  F then type in my secret phrase and presto!  I then reversed the process< and did a difference on the unencrypted file and they match!  D one question ... when I created a 1024 private key it created a file= called RANDSEED.BIN ... can I rename this to whatever I want?n  ? question two ... is the version on your server Doc the same onex7 that HP has out for 7.2-7.3 or is this a different one?c   ------------------------------  $ Date: Fri, 5 Sep 2003 20:36:03 +0200) From: "Peter Grevel" <p.grevel@hccnet.nl> & Subject: RDB Interactive SQL questions) Message-ID: <bjakvf$n27$1@news.hccnet.nl>   D I want t present a list from a table which contains a rownumber, see example.7 I thought this is very basic. But I cannot find out how    How do I handle this ?   select FLD1 from TABLE1  where FLD2 < 100;p   FLD1 1  AAt 2  BBF 3  CC  4  XXm    F Another problem I have is how to pass a value from a previous selected: record to the current record, when selecting from a table.< I think it is the some problem I have as as I mention above.    I'm using Oracle RDB for OpenVMS  7 Please post a reply with your comments or send me mail.e   Thanks in advance,   Peter Grevel   ------------------------------   Date: 5 Sep 2003 13:14:46 -0500-+ From: young_r@encompasserve.org (Rob Young)e Subject: Re: RdB on EVA ? 3 Message-ID: <F7+4HzpfIJJ5@eisner.encompasserve.org>K  S In article <bjah6e$i7p$1@news2.tilbu1.nb.home.nl>, Dirk Munk <munk@home.nl> writes:c > jlsue wrote:F >> On Wed, 03 Sep 2003 22:40:01 +0200, Dirk Munk <munk@home.nl> wrote: >> o >> >S >>>In general it is recommended to put files with mainly random access on one disk TG >>>group, and mainly sequentially accessed files on another disk group.7 >>>1/ >>>Within the disk groups you define your luns.e >>   >> ,K >> I do not believe that this is true (at least, not anymore, not sure when  >> you learned this).A > S > HPQ storage engineers gave me this advice several times over the past years, the cR > last time during a hands-on session at HP world 2003 in Atlanta a few weeks ago. > J > A good example would be to put Oracle redo files on a 'sequential' disk 9 > group, and the database files on a 'random' disk group.+ >   B 	This advice doesn't make sense.  A redo log is a write only file.C 	The EVA would catch the writes and flush to disk in between reads.<   >  >> 6I >> There is a balance between flexibility, manageability, and performancepL >> here.  I really doubt you'd be able to measure much significant gain fromJ >> the effort to split these into two groups.  If you've got 100 spindles,M >> splitting into 50 for sequential and 50 for random access will not provide < >> much benefit from just leaving all 100 in the same group. > O > I can imagine that the automatic tuning algorithm of the EVA can distinguish $P > between sequential and random access, thus optimizing disk groups accordingly. >   E 	With tons of random IO , you want as many spindles as possible.  The D 	question would be would the sequential IO interfere "somehow."  I'mE 	not sure how and short of a good technical explanation wouldn't know A 	how.  In the case of redo logs, surely the EVA could destage thenA 	sequential writes without interfering with servicing random read = 	traffic.  I'm sure Bill has some ideas as this is his forte.    				Rob    ------------------------------  % Date: Fri, 05 Sep 2003 19:00:14 -0400 + From: John Johnstone <jj_usenet2@yahoo.com>y3 Subject: Slow keyboard autorepeat rate for new DS10.) Message-ID: <3F58DD3E.50DCDAB8@yahoo.com>d  B I've just setup a new AlphaServer DS10 and it appears that the new? console firmware V6.5-15 has a REAL slow autorepeat rate on theyC keyboard (LK462-A2 A01).  Hmmm, we finally get a graphics card (ATI D Radeon 7500) that's fairly fast and somebody slows down the keyboard? autorepeat rate!  I guess that it's the new firmware that's the @ culprit since I didn't see this with the previous version V6.4. A There's no mention of this in the V6.5 firmware release notes.  Ib# could roll back to V6.4 to be sure.d  E Has anyone else seen this?  Is there any way to change the autorepeatoE rate back up to what it was before?  There were some flaky autorepeatoF rate problems with some Personal Workstations (the one that I'm typing@ this on actually), but this is the first time I've seen anything unusual with a DS10.   ------------------------------  % Date: Fri, 05 Sep 2003 14:21:02 -0500- From: paul@wren.cc.kux.edu! Subject: VMS on the trading floorL8 Message-ID: <afohlvca6qaqhsqac996d38jebr0j3vn48@4ax.com>  ? Look for the reference to onExchange Inc.'s Extensible ClearingD System.R   http://tinyurl.com/mdspa   ------------------------------  % Date: Fri, 05 Sep 2003 14:26:34 -0400@* From: JF Mezei <jfmezei.spamnot@istop.com>: Subject: Re: Will OpenVMS I64 run on a Dell PowerEdge 3250) Message-ID: <3F58D54F.3AB6C32C@istop.com>>   Bill Gunshannon wrote:J > DELL.  Certified VMS systems will cost more.  I didn't say that, various8 > people from HP said right here in the last few weeks.   M Is this a case of a markup for systems destined for VMS, or just a case of HP'K only certifying the more expensive systems for VMS (continuing the trend too= prevent VMS from being scalable from desktop to data centre).s  H For low end systems, if there were a will, there would be a way. When HPL produces a IA64 thing that is sold with Linux on it, does HP spend megabucksL to certify Linux on that box resulting in a higher price for the box ? SinceN HP can't recover any money on Linux licenses, does HP then charge more for theB actual system, even if the customer won't be running Linux on it ?  N It seems to me that VMS engineers should be able to certify a single "vanilla"G configuration and then "support" any system that abides by those rules.E  K The term "support" is used loosely. HP could have two types of VMS support:5K one for certified mission critical systems, and one for "industry standard"4 systems with vanilla configs.=    K > asked here back at the time the port to IA64 was beginning if it could be-> > expected that VMS would run on any industry standard system.  L If IA64 systems will be so specific that each one must truly be individuallyJ certified, then the architercture cannot be called "industry standard" and! Carly must stop usuing that term.1   ------------------------------   Date: 5 Sep 2003 17:15:51 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)n: Subject: Re: Will OpenVMS I64 run on a Dell PowerEdge 32503 Message-ID: <TnKygxbrHc$E@eisner.encompasserve.org>S  h In article <bjageu$f6o5m$1@ID-135708.news.uni-berlin.de>, bill@gw5.cs.uofs.edu (Bill Gunshannon) writes:  K > Chips don't run OSes.  Systems do.  And, it has been very recently stated 7 > that VMS will only be supported on certified systems.h  F Perhaps it is a recent statement to you, but to me it is tautological.E The meaning of certification is that VMS support is available for it.tH (Note there is at least one Unix-only Alphaserver which is not certified for use with VMS.)   > Those do not include DELL.  C I don't think anyone, including Mark Gorham, can say that for sure. A VMS Development would likely certify a system from another vendoro< if that vendor was willing to pay the costs of that process.  C > Certified VMS systems will cost more.  I didn't say that, various-7 > people from HP said right here in the last few weeks.   - If nothing else, certifying them costs money.c   > The question wasK > asked here back at the time the port to IA64 was beginning if it could begJ > expected that VMS would run on any industry standard system.  The answerM > was "engineering is doing nothing to make VMS rely on special HP hardware."hL > I think most people took this to mean that VMS would be capable of runningK > on any "industry standard" (whatever that means) Itanium system you couldt > buy.  E Which of course is entirely different from being _supported_ on thosel uncertified boxes.   ------------------------------  # Date: Fri, 05 Sep 2003 20:18:16 GMTe0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com>- Subject: Re: [Q] Learning to use Queues in C. > Message-ID: <cc66b.84697$bo1.59500@news-server.bigpond.net.au>  F Did you try printing the sizeof(...) or the base address of each item?  8 It seems your use of pragma is not doing what you hoped.   Matt.c   -- p= -------------------------------------------------------------e OpenVMS TCP/IP Engineering Enterprise Computing Group Hewlett-Packard Companyh Gold Coast, AUSTRALIAi= -------------------------------------------------------------t    G "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de> wrote in messagee2 news:bj9fka$g2veq$1@ID-56200.news.uni-berlin.de... > Matt Muggeridge wrote:I > > Your queueitem contains the record structure which is not quadword in I > > size. Since you have this in an array, only the first element will besE > > quad aligned.  The remaining array items will appear at long-wordh > > offsets. >iC > But I still get %SYSTEM-F-ROPRAND if I surround the queueitem andeC > queueheader definitions with "#pragma nomemberalignment quadword"t4 > (so that each array member shold be quad aligned). >y > cu, 
 >   Martin > -- sH >   OpenVMS:                | Martin Vorlaender  |  VMS & WNT programmer5 >    The operating system   | work: mv@pdv-systeme.decH >    God runs the           |   http://www.pdv-systeme.de/users/martinv/< >    earth simulation on.   | home: martin@radiogaga.harz.de >a   ------------------------------   End of INFO-VAX 2003.493 ************************