1 INFO-VAX	Mon, 28 Oct 2002	Volume 2002 : Issue 595       Contents:A Re: HP-ETS 2002 Presentations are on-line - Pay-per-view ONLY :-(  Re: Immutable laws of the PC Re: Immutable laws of the PC Installing VMS Jensen   Re: So I went to the HP IT forum  Re: Timezone-change observations  Re: Timezone-change observations  Re: Timezone-change observations  Re: Timezone-change observations2 [FREEWARE] Announcing MAILBOX 0.8 Beta for OpenVMS/ Re: [OpenVMS VAX V7.3] SYSMAN misses /POWER_OFF + [OpenVMS VAX V7.3] SYSMAN misses /POWER_OFF   F ----------------------------------------------------------------------   Date: 27 Oct 2002 21:40:51 GMT& From: peter@abbnm.com (Peter da Silva)J Subject: Re: HP-ETS 2002 Presentations are on-line - Pay-per-view ONLY :-(- Message-ID: <aphmh3$hqv@web.eng.baileynm.com>   H Time for a little consolidation, methinks. Perhaps the remnants of DECUS could merge with Usenix?   --  O I've seen things you people can't imagine. Chimneysweeps on fire over the roofs O of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All L these things will be lost in time, like chalk-paintings in the rain.   `-_-'K Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`    ------------------------------  % Date: Sun, 27 Oct 2002 05:49:40 -0500 * From: "Bill Todd" <billtodd@metrocast.net>% Subject: Re: Immutable laws of the PC 2 Message-ID: <ibGcnb72g6AoWSagXTWcpQ@metrocast.net>  / "xganon" <remailer@xganon.com> wrote in message 3 news:d62fdaca217a98689aec24fe97c6fdf4@xganon.com...    ...    > The sweet spot of 4P to 8P is  > driven by transactions.   G While that's certainly *one* common use, I'd like to see a quantitative K breakdown before accepting the assertion that it's *the* common use of such  systems.  )   How well does the box perform running a B > database.  Several easily locatable studies point out that largeC > on-chip caches coupled with keeping the cache misses down (branch C > prediction coupled with I2 speculative execution) drive the rates 	 > higher.   J The point under discussion (which I snipped for brevity) was not whether aJ large cache helps TPC-C but exactly how cache size needs to scale to allowI performance to scale linearly with CPU clock frequency in the presence of L relatively static memory-access latency.  If those studies that you mention,H for example, indicate that the miss ratio for TPC-C improves at *better*J than the square root of the cache-size increase, then one could expect theJ larger caches in Madison and Montecito to help performance scale more thanK linearly with clock frequency; otherwise, there's no reason to believe that G they'll do any better than keep pace with the baseline set by McKinley.   =   That is why a less powerful 4P Itanium 2 beats all other 4P  > boxes.  G Perhaps.  There's also considerable database set-up art to getting good I TPC-C numbers - plus, of course, the need to be willing to spend money on H the storage.  The fact that HP has an obvious disincentive to try to getF competitive TPC-C performance out of Alpha now that it's committed itsJ future to Itanic makes me somewhat dubious that Itanic really offers aboutK 40% better performance per processor than the current high-end Alpha does - ( and also bodes ill for EV7 TPC-C scores.  L Since Opteron will have a minimum of 1 MB of on-chip cache to work with, andD IIRC 2 MB as an option, it will certainly be far less prone to beingK crippled in this area than the current Alpha is (with only 64 KB of on-chip I data cache plus off-chip cache that's only slightly faster than Opteron's G main memory).  For that matter, McKinley's latency to its large on-chip I cache is 12 ns., while Opteron's open-page latency to main memory is only K about 50 ns. (which is three times as fast as McKinley's main-memory access K latency of 156 ns.).  So I still expect Opteron to blow McKinley out of the 3 water in TPC-C, and quite possibly Madison as well.    > F > When on-chip cache goes to 6MByte (Madison) and 12 MByte (Montecito)7 > the transactions per cpu will be driven quite higher.   E As noted above, the question is whether they'll scale any better than G linearly with the clock frequencies (and I'm charitably leaving out any K possible I/O bottlenecking that may occur, since I don't know how close the 8 zx1 chipset may be to saturation by TPC-C in this area).  F If you'd like to provide some quantitative studies in any of the aboveK areas, I'd be interested in taking a look.  Otherwise, we can just agree to B disagree until actual tests determine who's closer to being right.   - bill   ------------------------------  % Date: Sun, 27 Oct 2002 15:44:39 -0500 - From: JF Mezei <jfmezei.spamnot@videotron.ca> % Subject: Re: Immutable laws of the PC , Message-ID: <3DBC5035.9D5EBE78@videotron.ca>  
 xganon wrote: B > database.  Several easily locatable studies point out that largeC > on-chip caches coupled with keeping the cache misses down (branch C > prediction coupled with I2 speculative execution) drive the rates F > higher.  That is why a less powerful 4P Itanium 2 beats all other 4P > boxes.  N Does IA64 have branch prediction ? Doesn't that sort of go against the concept; of EPIC where the compiler does all the optimisation work ?    ------------------------------  % Date: Sun, 27 Oct 2002 18:42:58 +0200 $ From: "Dr. Dweeb" <Dweeb@nospam.com> Subject: Installing VMS Jensen0 Message-ID: <VIUu9.106$eB4.2093@news.get2net.dk>  B I have a Jensen which has had NT running on it for some years.  ItL originally also had VMS installed.  Anyhow, I would like to reinstall VMS on it now.   J I ran the configuration utility, changed the boot to VMS etc. ROM 2.2 SRMM 370 : SHOW DEVICE showed the RRD42 (?) as ROMDISK RX CD-ROM 4.5d  J However, an attenpt to boot standalone backup from the CD results in error 84  J This does not help to much I know, because that is the error when there is no CD in the drive.    Any obvious clues ?    Dweeb.   ------------------------------  % Date: Sun, 27 Oct 2002 15:37:15 -0500 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ) Subject: Re: So I went to the HP IT forum , Message-ID: <3DBC4E7A.B36915BB@videotron.ca>  ) Few commentas about educational licenses.   N From what I have read here, it seems that there is a (bnew ?) person in charge; of educational licenses who is eager to do something right.   H While there is absolutely no reason not to believe that there is so trueJ "keenness" in that person to improve things, I have vague recollections ofC similar things in the past. They get in, promise to listen and make J improvements, but the only changes that the deliver aren't really enough.   C Obviously HP will have certain needs with regards to edu programme. K (marketing, legal, financial). If HP were to make those needs very clear to N instuitutions, and institiutions make their wishes very clear, perhaps both of/ them could come up with an acceptable solution.   F For instance, institutions complain about the amount of work to have aM gazillion individual licences instead of one big infinite licence. Perhaps HP K needs to have some count of how many students use VMS and the only way they E could think of is to provide individual licenses. But by talking with K institutions perhaps some more efficient method could be derived. (eg: each N semester, institution sends a report on all the SYSUAF.DAT it has on campus to; HP which can then have an idea of how many students use it.   J Also, if HP donates a wildfire to an university, would the university thenN generate service/maintenance revenus ? (are those also deeply discounted or doC universities have no problem with regular maintenance contracts ?).    ------------------------------  % Date: Sun, 27 Oct 2002 15:54:27 -0500 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ) Subject: Re: Timezone-change observations , Message-ID: <3DBC5281.9FFD33BC@videotron.ca>   Peter LANGSTOEGER wrote:E > It's the new SYS$LOCALTIME and the AUTO_DLIGHT_SAV which is new and N > unfortunately different on VAX and Alpha. On VAX there is no AUTO_DLIGHT_SAV   And guess what: 	 MC SYSMAN  SYSMAN> SET ENV/CLUSTER ! SYSMAN> DO SET TIME="-0 01:00:00"  SYSMAN> EXIT  M Is a lot quickler than fiddling around , looking into files and searching the S system diskc to find out where rules are stored and why it doesn't quite work well.   @ During the time I had DECnet-5, it did make the time zone changeJ automatically, but you know what ? I much prefer the quietness of the realJ decnet all year long even if it means that twice a year, I have to issue aK command to change the time. (heck, one can submit the command to execute at 3 the right time, while one is asleep or at a party).   7 And out of curiosity, how does one go about setting the G SYS$TIMEZONE_DIFFERENTIAL logical ? (on VAX)  It seems to have been set K properly (-5 hours) at one point, but I have no idea when/how this was set, V and even less how to properly set it to the right one when summer/winter times change.   ------------------------------    Date: 27 Oct 2002 19:34:49 -0800( From: bob@instantwhip.com (Bob Ceculski)) Subject: Re: Timezone-change observations = Message-ID: <d7791aa1.0210271934.10a3b4fc@posting.google.com>   k peter@langstoeger.at (Peter LANGSTOEGER) wrote in message news:<ErMu9.176047$N_6.2542849@news.chello.at>... D > Because summertime ended today, I thought I share my observations. > ; > 1) Local Time (SHOW TIME) was correct on VAX and Alpha !! * > 	I'm running VMS V7.3 and TCPware V5.6-2I > 	but no longer DECdts (which made perfect timezone changes for years !)  >   B why are you doing it the hard way?  all you have to do is set up 2* tcpware_configure.com files, for me it was  7 tcpware_configure.com_edt and tcpware_configure.com_est   @ a dcl command procedure can be written to run once in the spring< and once in the fall on the appropriate date and time at 2am; and swap in the correct tcpware configure com file and do a  simple    
 $ @restart   isn't this simple?   ------------------------------  % Date: Sun, 27 Oct 2002 22:16:10 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> ) Subject: Re: Timezone-change observations 9 Message-ID: <ZZ1v9.2682$Nf2.216152@news20.bellglobal.com>   : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3DBC5281.9FFD33BC@videotron.ca... > Peter LANGSTOEGER wrote:G > > It's the new SYS$LOCALTIME and the AUTO_DLIGHT_SAV which is new and @ > > unfortunately different on VAX and Alpha. On VAX there is no AUTO_DLIGHT_SAV  >  > And guess what:  > MC SYSMAN  > SYSMAN> SET ENV/CLUSTER # > SYSMAN> DO SET TIME="-0 01:00:00"  > SYSMAN> EXIT > K > Is a lot quickler than fiddling around , looking into files and searching  the J > system diskc to find out where rules are stored and why it doesn't quite
 work well. > B > During the time I had DECnet-5, it did make the time zone changeL > automatically, but you know what ? I much prefer the quietness of the realL > decnet all year long even if it means that twice a year, I have to issue aJ > command to change the time. (heck, one can submit the command to execute at5 > the right time, while one is asleep or at a party).  > 9 > And out of curiosity, how does one go about setting the I > SYS$TIMEZONE_DIFFERENTIAL logical ? (on VAX)  It seems to have been set H > properly (-5 hours) at one point, but I have no idea when/how this was set,J > and even less how to properly set it to the right one when summer/winter
 times change.   H JF. Changing the time manually (or via batch) misses the point. Some dayG soon, all computers on the planet will set their internal clocks to UTC K (this includes file stamps too). Users will be presented with an individual K local view of that internal time and it is this UTC-to-Local view that will A be adjusted twice a year, not the system clock. (depending on the H application, changing the system clock can be a really bad idea). In theK mean time, getting these rules tables fixed brings OpenVMS systems into the  modern computer era.  J This year, I was happy to retire my 15 year old home-brewed DCL script andL allow a combination of AUTO_DLIGHT_SAV and NTP take care of the time change.E As far as I'm aware, the AUTO_DLIGHT_SAV feature in OpenVMS-7.3 still B changes the system clock but I've been led to believe that OpenVMSL Engineering is working toward a total UTC solution (this is probably the one1 thing that UNIX did correctly right off the bat).   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------  # Date: Mon, 28 Oct 2002 06:09:30 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)) Subject: Re: Timezone-change observations 5 Message-ID: <uw4v9.187963$N_6.2763110@news.chello.at>   h In article <d7791aa1.0210271934.10a3b4fc@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes:l >peter@langstoeger.at (Peter LANGSTOEGER) wrote in message news:<ErMu9.176047$N_6.2542849@news.chello.at>...E >> Because summertime ended today, I thought I share my observations.  >>  < >> 1) Local Time (SHOW TIME) was correct on VAX and Alpha !!+ >> 	I'm running VMS V7.3 and TCPware V5.6-2 J >> 	but no longer DECdts (which made perfect timezone changes for years !) > C >why are you doing it the hard way?  all you have to do is set up 2 + >tcpware_configure.com files, for me it was  > 8 >tcpware_configure.com_edt and tcpware_configure.com_est > A >a dcl command procedure can be written to run once in the spring = >and once in the fall on the appropriate date and time at 2am < >and swap in the correct tcpware configure com file and do a >simple  >  >$ @restart  >  >isn't this simple?    Yup, but you missed the point.E The point was to get VMS being in the line of many Opsys which do the B DST switches automatically. And progress have been made obviously.  G But, as I wrote, something is still not ok. The VMS time switch on VAX, J the DST rules on Alpha and the neccessity to change TCPware startup files.   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------    Date: 27 Oct 2002 19:40:17 -08003 From: mailbox@mailbox-vms.org (MAILBOX for OpenVMS) ; Subject: [FREEWARE] Announcing MAILBOX 0.8 Beta for OpenVMS = Message-ID: <581c42ed.0210271940.608b3f36@posting.google.com>   ? I am delighted to announce the availability of the Beta version + of a major upcoming new release of MAILBOX:  MAILBOX 0.8 for OpenVMS.  ? MAILBOX 0.7 has been downloaded by several hundred users. While @ I have extremely little time to work on MAILBOX, it is rewardingD and amazing to receive mails from users living all around the world.  B MAILBOX 0.8 Beta brings many exciting improvements to MAILBOX 0.7,- and responds to a number of popular requests.   ? * Huge parts of the code have been completely rewritten, making -   MAILBOX more stable and easier to maintain.     > * A new, full-featured widget library for console applications?   has been designed especially for MAILBOX. This widget library A   brings a new streamlined look to MAILBOX as well as a number of     enhancements to the interface.  A     - It is now possible to select several objects within a list. ?       Sort, file, and delete your messages in the bliof an eye.   B     - A new file selection dialog lets you browse easily your file/       systems, and apply filters to file names.   ?     - Navigate throughout MAILBOX using the TAB and arrow keys; C       use radio buttons, checkbox and list boxes. MAILBOX has never        been so easy to use.  B This Text Widget Library will be released in the months to come as@ OpenSource software for the OpenVMS platform. Future versions of1 MAILBOX will leverage the updates to the library.     @ * The Options screen unveals a bunch of options that were hidden"   in previous versions of MAILBOX.  "        - Open any .MAI file easily           - Choose your text editor  <        - Choose SMTP or VMSMail as your mail transport agent  @        - Choose whether you want read messages to go to the MAIL          folder automatically.          - And more.    A * A number of new features are available. For example, the Search >   feature lets you search messages by sender, recipient, date,   and subject.    ; * MAILBOX Single-User Edition enables regular users lacking E   administrative privileges to install MAILBOX on their own accounts. 0   This is in response to a very popular request.    D * MAILBOX now has its own Web site. Go to http://www.mailbox-vms.orgA   to learn about new releases, browse the documentation or report    bugs.     = Please give a try to MAILBOX 0.8 Beta for OpenVMS, and take a B minute to fill our registration form. MAILBOX is free software andH we are looking forward to learning a little about you through this form.  E You will find answers to most common questions regarding this release ) in our FAQ at http://www.mailbox-vms.org.   A We welcome your comments and encourage you to report bugs. Please D remember this is a Beta version and refer to our web site for a list of known bugs.  G We hope you enjoy MAILBOX 0.8 for OpenVMS and we are looking forward to  hearing from you!      Patrick    MAILBOX for OpenVMS  mailbox@mailbox-vms.org  http://www.mailbox-vms.org   ------------------------------    Date: 27 Oct 2002 14:06:07 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) 8 Subject: Re: [OpenVMS VAX V7.3] SYSMAN misses /POWER_OFF3 Message-ID: <+gsTN84PBfPh@eisner.encompasserve.org>   f In article <ehYu9.184104$N_6.2713545@news.chello.at>, peter@langstoeger.at (Peter LANGSTOEGER) writes:   > SYSMAN>SHUT NOD/POWER_OFF Q > %CLI-W-IVQUAL, unrecognized qualifier - check validity, spelling, and placement  >  \POWER_OFF\ >  > N > What do you think, is this a coding error or is this a documentation error ? > M > Is /POWER_OFF a Alpha_Only feature (No, I don't mean that a VAX will follow K > a /POWER_OFF in OPCCRASH.EXE, I mean a VAX must be able to request such a K > behaviour from an Alpha which should be able to satisfy it of course) and F > someone forgot to mention it in HELP or is /POWER_OFF simply not yet$ > implemented in SYSMAN.EXE on VAX ?  G I would think it is a design error; that the programmer had not thought  of the scenario you describe.    ------------------------------  # Date: Sun, 27 Oct 2002 20:47:06 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)4 Subject: [OpenVMS VAX V7.3] SYSMAN misses /POWER_OFF5 Message-ID: <ehYu9.184104$N_6.2713545@news.chello.at>   4 On my OpenVMS VAX V7.3 system I have in SYSMAN help:   SYSMAN>HELP SHUT NODE /POWER   SHUTDOWN     NODE       /POWER_OFF  B        Specifies that the system is to power off after shutdown is        complete.    / however SYS$SYSTEM:SYSMAN.EXE thinks otherwise:    SYSMAN>SHUT NOD/POWER_OFF O %CLI-W-IVQUAL, unrecognized qualifier - check validity, spelling, and placement   \POWER_OFF\    L What do you think, is this a coding error or is this a documentation error ?  K Is /POWER_OFF a Alpha_Only feature (No, I don't mean that a VAX will follow I a /POWER_OFF in OPCCRASH.EXE, I mean a VAX must be able to request such a I behaviour from an Alpha which should be able to satisfy it of course) and D someone forgot to mention it in HELP or is /POWER_OFF simply not yet" implemented in SYSMAN.EXE on VAX ?   TIA    -Peter   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------   End of INFO-VAX 2002.595 ************************