INFO-VAX Wed, 03 Dec 2008 Volume 2008 : Issue 645 Contents: Re: %ANALDISK-I-BADHIGHWATER Adding to the Noise WAS Who is left at VMS engineering ? Re: Anyone want a ride on the SS Itanic? Re: Anyone want a ride on the SS Itanic? Re: Banana Republic (was Re: OpenVMS Book Wins award) Re: Good example of C and MACRO It that VAX 11/750 in the background? Re: It that VAX 11/750 in the background? Re: Multicore Is Bad News For Supercomputers Re: Multicore Is Bad News For Supercomputers Re: Multicore Is Bad News For Supercomputers RE: Multicore Is Bad News For Supercomputers Re: Multicore Is Bad News For Supercomputers Re: Open Source Message Oriented Middleware on OpenVMS? Re: OT: How We Found the Missing Memristor Re: RDB plugin/driver for Qt Re: RDB plugin/driver for Qt Re: Synchronize with a remote ntp source Re: Synchronize with a remote ntp source Re: VMS, HP and the recession Re: X-terminals Re: X-terminals Re: X-terminals ---------------------------------------------------------------------- Date: Wed, 3 Dec 2008 01:24:15 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: %ANALDISK-I-BADHIGHWATER Message-ID: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: >In article , moroney@world.std.spaamtrap.com >(Michael Moroney) writes: >> helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: >> >> >Should I be concerned about %ANALDISK-I-BADHIGHWATER? Why does >> >ANA/DISK/REPAIR not repair it? >> >> On PAGEFILE.SYS/SWAPFILE.SYS/SYSDUMP.DMP >Right. These files are normally marked /NOBACKUP, so their contents are not copied during a backup. When a system disk is restored from backup, these files are created, space is allocated and the EOF pointer is set. Since they are never actually written (BACKUP has nothing to write), the highwater mark is left at 0. ANALYZE/DISK considers it a minor error for allocated space to be above the highwater mark. Since these files are only written at a very low level, the highwater mark is not reset. Call it a minor BACKUP bugley, I guess. ------------------------------ Date: Tue, 02 Dec 2008 19:14:46 -0800 From: "Tom Linden" Subject: Adding to the Noise WAS Who is left at VMS engineering ? Message-ID: On Tue, 02 Dec 2008 07:49:51 -0800, Steven M. Schweda wrote: > From: "FredK" > >> [...] For a variety of >> reasons, I have chosen to not spend time in cov. To be honest it wasn't >> even a decision. I gradually spent less and less time checking in on >> the >> newgroup until one day I realized that I hadn't opened it in over a >> month >> and hadn't missed it. The signal to noise ratio was way too high. >> [...] > > Haven't you been paying attention? I've been told repeatedly (for > years, until I (mostly) stopped bothering to complain) that all the > noise was the result of free people exercising their basic rights, and > that that there were no significant adverse consequences to any of the > off-topic rambles. Now you come along and testify to the contrary. > What's a fellow to believe? I usually don't even open what I guess is likely opinions and the like and summarily commit them to the dust bin (although I did here:-) > >> Perhaps things have changed (hope springs eternal)? I'll look in from >> time >> to time over the next few weeks and decide if it is worthwhile to keep >> track >> of things here. > > As in so many things, low expectations would probably be appropriate. > > ------------------------------------------------------------------------ > > Steven M. Schweda sms@antinode-info > 382 South Warwick Street (+1) 651-699-9818 > Saint Paul MN 55105-2547 -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Tue, 02 Dec 2008 21:16:33 -0800 From: Alan Frisbie Subject: Re: Anyone want a ride on the SS Itanic? Message-ID: <49361631.3080009@Flying-Disk.com> Glen Herrmannsfeldt wrote: > Without hard disk, but with a DVD drive and 10GB RAM. > (With disk tray and screws, for easy installation.) I just received mine (the version without the DVD but with the management processor). It had two PCI cards included: 1. Quadrics QM500-A Network adapter, which I am sure is not supported by VMS. I am also sure that it will turn up on ebay before too long. :-) 2. A6829A (LSI Logic 22915-HP) 2-port ultra 160 SCSI It came with two drive sleds (with screws) and one dummy sled. I found additional drive sleds on ebay for $20/each. The seller has them listed at $23 with "Make an Offer", so I offered $20/each for 5 and it was immediately (within a minute) accepted. You might be able to get an even better deal. He has over 40 of them. Search for item 220322487770 HP 9000 3.5-inch SAS/SATA/SCSI Drive Tray/Caddy RP3400 HP 9000 Servers RP3410 RP4410 RP7420 RP8420 Superdome (OK, they are black, but I don't insist on the color matching!) I got a good deal on the HP DVD-ROM drive for it, but would like to get a burner. Does anyone have any suggestions for a suitable drive? Now all I need is a fibre channel card and I'm all set! Alan ------------------------------ Date: Tue, 02 Dec 2008 22:36:24 -0700 From: Glen Herrmannsfeldt Subject: Re: Anyone want a ride on the SS Itanic? Message-ID: Alan Frisbie wrote: (snip) > I just received mine (the version without the DVD but with > the management processor). It had two PCI cards included: > 1. Quadrics QM500-A Network adapter, which I am sure is not > supported by VMS. I am also sure that it will turn up on > ebay before too long. :-) The picture looks like there is built in 10/100 and a separate port for gigabit ethernet. > 2. A6829A (LSI Logic 22915-HP) 2-port ultra 160 SCSI It also shows a built-in SCSI port with a "do not use" sticker over it. > It came with two drive sleds (with screws) and one dummy sled. > I found additional drive sleds on ebay for $20/each. The > seller has them listed at $23 with "Make an Offer", so I > offered $20/each for 5 and it was immediately (within a minute) > accepted. You might be able to get an even better deal. He > has over 40 of them. Search for item 220322487770 > HP 9000 3.5-inch SAS/SATA/SCSI Drive Tray/Caddy RP3400 > HP 9000 Servers RP3410 RP4410 RP7420 RP8420 Superdome Two should be enough for me for a while. > Now all I need is a fibre channel card and I'm all set! I once was interested in fibrechannel. We had a system with external FC ports, and an FC drive, but the connectors were completely different. Before I got around to figuring out how to connect them, we got onto a different project and didn't need that one. -- glen ------------------------------ Date: Wed, 3 Dec 2008 10:23:00 +0800 From: "Richard Maher" Subject: Re: Banana Republic (was Re: OpenVMS Book Wins award) Message-ID: Hi Mark, > With dynamic content, say using a Python or PHP engine, > Apache does not change. WASD then assumes a per-process model too. Delightful. > "Absolute bollocks" is far too strident. Relative bollocks would be > closer to the mark. :-) > Well, what works for 10 concurrent usages may work for 100 and may not > for 1,000 or 10,000. Or it may. It depends on what needs to be > accomplished. Yeah, I'm advising those who are happy to think along those lines, to use Sockets on the client (As in my JAVA example, or WebSockets or a.n.other) and just use INETd on the server and be done with it. No Middleware required at all! If 1000 threads and 1000 attaches to the database and 1000 Stacks can happily fall under the umbrella of "it depends" then why not bring the 1000 processes in from the rain? Think of the MUTEX savings and consurrency boost. > I have written code that employs POSIX threads. Anyone who has done > anything non-trivial in VMS' AST-driven environment is familiar with > many of the issues. I have never written an Apache module. No. Is the > implication that it cannot or has not be done? No, just that I think there's enough devil-in-the-detail for the issue to require whole chapters in the book of web-implementations and certainly a lot more than a "oh you'd just use that thing" :-) IMO, something also worthy of a chapter is the topic of why ASTs are far superior to the Threading model! (And much easier to implement, even though JAVA does a wonderful job af making threads easy(er)) > The paragraph describing Tier3's pool of flexible processing resources > also *exactly* describes WASD's CGI/CGIplus application environment - I > could not have put it more succinctly myself :-) > > http://wasd.vsm.com.au/ht_root/doc/scripting/scripting_0100.html > > How come? Some or all of these characteristics are more often essential > requirements for accomplishing 'real work'. Partitioning activity is > fundamental to enabling, securing and auditing application environments. > If you don't need or want to pursue it using Web technologies ... Let's put an end to this gratuitous willy-waving and just agree that mine was the far more impressive display :-) If you insist that your chalk does *exactly* what my cheese does then I won't argue with you, but when it comes to the issue of "flexible processing resources" then I do concede that Tier3, WASD CGI/Plus, ACMS, Rdb SQL Services and so on. . . do have similarities. > > No bollocks HTTP, SOAP, XML (unless you really want), Java, Garbage > > Collector, RMI, Threads, WSIT, Axis2, Apache, Tomcat, WASD, OSU, CGI, Perl, > > PHP, Pyhthon! > > > > Just the VMS 3GLs you know and love, Oracle (Rdb or Orrible), and RMS on the > > back-end. (The world's your oyster on the front-end: - HTML, Javascript, > > Java, Flex, Flash, Silverlight, VMS) > > Sounds like having a bob both ways :-) I'm just appealing to that school of thought that no says "The browser *is* the GUI" and even "The browser *is* the computer". It's amazing what can be achieved with these simple and powerful tools! The problem for VMS, as I see it, is that you can pick up a really grunty Intel-based LAMP server for next to nothing and have *all* the latest and greatest versions and security/bug fixes delivered on-time! With VMS everything is 2-to-10 years late (if supported at all) and the equivilant box with cost you ~100x more. What Tier3 lets you do is *integrate* your rich herritage of VMS 3GL code, business-rules, and data, *directly* in with your IIS or LAMP-based web architecture! But without all of the bagage infrastructure described above in the paragraph commencing ">> No bollocks . . .". An out-dated, poorly-performing, unpatched, and bloody-expensive VAMP-server simply can never compete with the ubiquitous LAMP servers or IIS. Just get over it! VMS Management's response is not only will they continue to compete in the xAMP space, but they're also up for a bit of biffo on the desktop browser space :-( It's only license-payer dollars I guess; who carse? Absolute disbelief! > > Sounds optimal :-) > > I wouldn't agree. Perhaps you have misread the question and/or > response. "Yes - the instance is UNavailable." It could be argued this > is suboptimal. Of course this, again, depends. Sorry, an attempt at sarcasm. > > I agree; so that's all two of us then :-( > > Cows outside Oshkosh sometime seem more topical. Been good lately for some reason. > this (and earlier, similarly themed) threads. MONDESI (the > browser-based system monitor) is on the cusp of release. This > discussion spurred me to spend the remaining 80% of the development time > on the outstanding 20% of the work ;-) I look forward to it. (MONDESI - baseball stats? I like it!) Although I still don't know why you just don't use Ajax and a client-based poll-interval? Cheers Richard Maher "Mark Daniel" wrote in message news:0139408d$0$20630$c3e8da3@news.astraweb.com... > Richard Maher wrote: > > Hi Mark, > > Hello again Richard. > > > (Once again, sorry for the delay) > > To quote Weizenbaum's Eliza Doctor, "How long have you been always > apologizing?" :-) > > >> Cross-domain access is one of the holy grails of distributed > >> applications > > > > There is also a strong argument for server-side aggregation, or portal > > functionality. (and good 'ol same-origin policy) > > > >> (at least those that can be mashed together from existing > >> webby technologies) > > > > Granted. > > > >> and are always fraught with security related issues. > > > > To say the least. The JSON script-injection option I find particularly > > scary! (Although I cannot see why, at least for Sockets, some/many people > > still pursue HTTP Access Control at the expense of policy files.) > > All injection vulnerabilities are extraordinarily scary; I know %-{ Of > course with non-interpreted environments you only have > buffer-overflowing, trampolining ones to worry about. SQL interfaces > have presented similar issues. > > >> Of course there probably also is an element of 'HTML people' tending to > >> have only a hammer in their toolbox (no real slight intended). > > > > I think you're right. > > > >> To better convey that this example has some level of sophistication here > >> is a (short-lived) peek at the HMI > >> > >> http://wasd.vsm.com.au/wasd_tmp/mondesi_081116a.png > > > > Looks good! (Although I think I recall some of JFP's Stills looking sexier; > > and with stock standard Ajax?) > > As observed by Mrs Claypool, "I think the Europeans do it better." > > >> > Is this > >>> thread/process serving *all* clients or is there a 1:1 relationship? > >> In all general purpose Web serving there is such a relationship. This > >> is definitely the case in the above application which is written as a > >> CGI script. All VMS Web servers would activate an instance for each > >> client (in fact for a CGI script, all servers period). > > > > So even when you're doing traditional request/response processing, you still > > get one instance per client? It's worse than I thought! But then I also > > thought that "Fast"-CGI (or some such beast) was meant to overcome this > > absolute bollocks? Although, limited in its application to long-polling I > > imagine. > > 'Traditional' request/response HTTP for static content on 'traditional' > Apache (say, a pre-2.0, pre-POSIX-threads build) is a one-to-one, > client-to-server-process model. WASD, using VMS' QIO-AST model, is not. > With dynamic content, say using a Python or PHP engine, Apache does > not change. WASD then assumes a per-process model too. > > FastCGI is an early ('90s) technology from the now assimilated Open > Market e-commerce company, designed to ameliorate the process > instantiation costs (on U*x!) associated with CGI scripting. > > CGIplus is a WASD technology for providing application/engine > persistence. Nothing to do with client concurrency. It's designed to > avoid the particularly expensive VMS process creation coupled with the > often similarly expensive and latent application instantiation (as with, > though not confined to, scripting engines such as Python, Perl, etc.) > > CGIplus is intended for, and is most efficient with, repeated, short > duration requests. Using it for a long-lived request is at best > inefficient, and certainly unnecessary. Otherwise MONDESI would also > have a CGIplus mode for use under WASD. > > "Absolute bollocks" is far too strident. Relative bollocks would be > closer to the mark. Value judgments are dependent on the environment > and objectives. > > > Not to worry, an article has just appeared over at CometDaily that says > > there's nothing wrong with 1000 threads for 1000 users! Doesn't say anything > > about attaching to the database 1000 times, duplicating memory and > > everything else 1000x, paging in/out, but then that's the Comet people for > > ya. > > Well, what works for 10 concurrent usages may work for 100 and may not > for 1,000 or 10,000. Or it may. It depends on what needs to be > accomplished. > > >> Of course in > >> many Web environments there would be nothing preventing the design and > >> implementation of something (like an Apache module) which maintained a > >> single, internal 'application' that serviced multiple, concurrent clients. > > > > OK, something like a single-threaded Apache (or Tomcat?) module that took > > standard Ajax/http requests, kept the connections open, sampled GETRMI data, > > and streamed it back to the client(s)? Perhaps you have one you prepared > > earlier? > > AIUI; Apache 2.n contains a POSIX threads implementation which would > natively support this. Of course many (most?) Apache sites still use > the child-process model which would negate the example advantage. > > I have written code that employs POSIX threads. Anyone who has done > anything non-trivial in VMS' AST-driven environment is familiar with > many of the issues. I have never written an Apache module. No. Is the > implication that it cannot or has not be done? > > > But surely one process or thread is a bottleneck and you'd need an > > application-configurable pool of Execution Server processes/threads to > > allocate the work to, and that pool could grow/shrink (within parameters set > > by the System Manager) to meet workload requirements? Then you might also > > want to know the VMS Username of the client you're performing work for so > > you can perform auditing and security checking? (It's a shame Ian Mugabe > > vetoed Rdb's introduction of SQL> Set Session Authorization Persona > > :ws_integer; But then Rdb doesn't work with threads so you're probably > > stuffed anyway.) > > Of course you do not mean one 'process or thread' *per server*. If it > was it would be a more-than-obvious bottleneck. > > > This is all sounding strangely familiar for some reason. . .oh I've got it - > > "You need Tier3!" As in :- > > The paragraph describing Tier3's pool of flexible processing resources > also *exactly* describes WASD's CGI/CGIplus application environment - I > could not have put it more succinctly myself :-) > > http://wasd.vsm.com.au/ht_root/doc/scripting/scripting_0100.html > > How come? Some or all of these characteristics are more often essential > requirements for accomplishing 'real work'. Partitioning activity is > fundamental to enabling, securing and auditing application environments. > If you don't need or want to pursue it using Web technologies ... > > > http://manson.vistech.net/t3$examples/demo_client_flex.html > > > > Username: TIER3_DEMO > > Password: QUEUE > > > > No bollocks HTTP, SOAP, XML (unless you really want), Java, Garbage > > Collector, RMI, Threads, WSIT, Axis2, Apache, Tomcat, WASD, OSU, CGI, Perl, > > PHP, Pyhthon! > > > > Just the VMS 3GLs you know and love, Oracle (Rdb or Orrible), and RMS on the > > back-end. (The world's your oyster on the front-end: - HTML, Javascript, > > Java, Flex, Flash, Silverlight, VMS) > > Sounds like having a bob both ways :-) > > >>> Is the thread/process unavailable for servicing > >>> other requests while it's streaming its long-poll (or words to that > > effect > >>> :-) > >> Yes. > > > > Sounds optimal :-) > > I wouldn't agree. Perhaps you have misread the question and/or > response. "Yes - the instance is UNavailable." It could be argued this > is suboptimal. Of course this, again, depends. > > >> An interesting . > > > > I agree; so that's all two of us then :-( > > Cows outside Oshkosh sometime seem more topical. > > Of course the more-than-casual c.o.v. audience is severely constrained > these days. > > >> and opportune thread, > > > > How so? > > Purely personal. My experimenting with some of these > Ajaxian/Cometish/HTML5/Web2 (choose your poison) technologies during > this (and earlier, similarly themed) threads. MONDESI (the > browser-based system monitor) is on the cusp of release. This > discussion spurred me to spend the remaining 80% of the development time > on the outstanding 20% of the work ;-) > > > Anyway, for anyone else out there who may be reading, let me reiterate an > > alternative architecture for asynchronous client event notification; it's > > called "UDP"! (Plus or minus Broadcasting and Multicasting functionality > > depending on the network intra/internet etc) A single client socket can > > receive messages from any number of server processes who in turn could be > > sending message events to any number of clients. Use this in combo with a > > middleware backbone based on a reliable transport such as TCP/IP and all of > > your application architecture needs will have been met! > > > > See below for a PUSH technology example (in case you missed a much earlier > > post to COV) > > > > Cheers Richard Maher > 8< snip 8< some fine code and 8< snip 8< excessive quotation 8< snip 8< ------------------------------ Date: Wed, 03 Dec 2008 01:01:32 +0100 From: Johnny Billquist Subject: Re: Good example of C and MACRO Message-ID: Bob Koehler skrev: > In article <6pdujdF7jt18U1@mid.individual.net>, billg999@cs.uofs.edu (Bill Gunshannon) writes: >> Not possible? I really expected someone to recognize RAD50 (aka RADIX-50) >> when I posted this. :-) > > But you said ASCII. The RAD50 character set is a subset of the ASCII > character set, which is why it works. Hey! RAD50 is just as much a subset of EBCDIC. Try to prove me wrong. :-) Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt@softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol ------------------------------ Date: Tue, 2 Dec 2008 16:17:13 -0800 (PST) From: paul_hansford@o2.co.uk Subject: It that VAX 11/750 in the background? Message-ID: See the you tube link from The UK (and ?) Renault Megane ad. http://uk.youtube.com/watch?v=rTthvxpzNDM 'I would never get a computer' 11-13 seconds into the ad. P ------------------------------ Date: Tue, 02 Dec 2008 19:27:15 -0600 From: Michael Austin Subject: Re: It that VAX 11/750 in the background? Message-ID: paul_hansford@o2.co.uk wrote: > See the you tube link from The UK (and ?) Renault Megane ad. > http://uk.youtube.com/watch?v=rTthvxpzNDM > 'I would never get a computer' 11-13 seconds into the ad. > P In the center it looks like a MVII or 11/73 or 11/83 with the RA disk on top and the one to the left does look like the 11/750... ------------------------------ Date: Tue, 2 Dec 2008 11:38:27 -0800 (PST) From: Christopher Subject: Re: Multicore Is Bad News For Supercomputers Message-ID: On Dec 2, 10:55=A0am, DaveG wrote: > On Dec 1, 7:20=A0pm, "Main, Kerry" wrote: > > > > > > -----Original Message----- > > > From: Neil Rieck [mailto:n.ri...@sympatico.ca] > > > Sent: December 1, 2008 3:04 PM > > > To: Info-...@Mvb.Saic.Com > > > Subject: OT: Multicore Is Bad News For Supercomputers > > > > This article should not surprise anyone in this NG, but it is nice to > > > know others are working on a fix. > > > > Multicore Is Bad News For Supercomputers > > >http://spectrum.ieee.org/nov08/6912 > > > > Neil Rieck > > > Kitchener/Waterloo/Cambridge, > > > Ontario, Canada. > > >http://www3.sympatico.ca/n.rieck/ > > > Its not only multi-cores that is the issue, but rather new buses like I= ntel's > > new QuickPath (formerly called CSI). > > > The new X86/Itanium bus architecture is NUMA based and that is a new pa= radigm > > that for very high performance requires accessing local memory much mor= e than > > remote memory. Hence, the OS and App's need to be aware and be able to = maximize > > perf with this architecture. > > > Reference:http://www.intel.com/pressroom/archive/reference/whitepaper_Q= uickPath... > > > Btw, anyone on this newsgroup know of an OS that already has a great de= al of > > experience With NUMA architectures? > > > :-) > > > Regards > > > Kerry Main > > Senior Consultant > > HP Services Canada > > Voice: 613-254-8911 > > Fax: 613-591-4477 > > kerryDOTmainAThpDOTcom > > (remove the DOT's and AT) > > > OpenVMS - the secure, multi-site OS that just works.- Hide quoted text = - > > > - Show quoted text - > > Most here know the answer to your question. =A0Do the "right" people @ > HP know? =A0That's the real =A0question. > > Dave... Does anyone at HP know? There is surprisingly little customer demand for openvms. For example, SAS doesn't support openvms at all. ------------------------------ Date: Tue, 2 Dec 2008 11:58:01 -0800 (PST) From: DaveG Subject: Re: Multicore Is Bad News For Supercomputers Message-ID: <262e0ac4-b4e2-468c-adac-9abf705b92db@j35g2000yqh.googlegroups.com> On Dec 2, 1:38=A0pm, Christopher wrote: > On Dec 2, 10:55=A0am, DaveG wrote: > > > > > > > On Dec 1, 7:20=A0pm, "Main, Kerry" wrote: > > > > > -----Original Message----- > > > > From: Neil Rieck [mailto:n.ri...@sympatico.ca] > > > > Sent: December 1, 2008 3:04 PM > > > > To: Info-...@Mvb.Saic.Com > > > > Subject: OT: Multicore Is Bad News For Supercomputers > > > > > This article should not surprise anyone in this NG, but it is nice = to > > > > know others are working on a fix. > > > > > Multicore Is Bad News For Supercomputers > > > >http://spectrum.ieee.org/nov08/6912 > > > > > Neil Rieck > > > > Kitchener/Waterloo/Cambridge, > > > > Ontario, Canada. > > > >http://www3.sympatico.ca/n.rieck/ > > > > Its not only multi-cores that is the issue, but rather new buses like= Intel's > > > new QuickPath (formerly called CSI). > > > > The new X86/Itanium bus architecture is NUMA based and that is a new = paradigm > > > that for very high performance requires accessing local memory much m= ore than > > > remote memory. Hence, the OS and App's need to be aware and be able t= o maximize > > > perf with this architecture. > > > > Reference:http://www.intel.com/pressroom/archive/reference/whitepaper= _QuickPath... > > > > Btw, anyone on this newsgroup know of an OS that already has a great = deal of > > > experience With NUMA architectures? > > > > :-) > > > > Regards > > > > Kerry Main > > > Senior Consultant > > > HP Services Canada > > > Voice: 613-254-8911 > > > Fax: 613-591-4477 > > > kerryDOTmainAThpDOTcom > > > (remove the DOT's and AT) > > > > OpenVMS - the secure, multi-site OS that just works.- Hide quoted tex= t - > > > > - Show quoted text - > > > Most here know the answer to your question. =A0Do the "right" people @ > > HP know? =A0That's the real =A0question. > > > Dave... > > Does anyone at HP know? =A0There is surprisingly little customer demand > for openvms. =A0For example, SAS doesn't support openvms at all.- Hide qu= oted text - > > - Show quoted text - My point exactly. ------------------------------ Date: Tue, 2 Dec 2008 12:03:34 -0800 (PST) From: DaveG Subject: Re: Multicore Is Bad News For Supercomputers Message-ID: On Dec 2, 1:38=A0pm, Christopher wrote: > On Dec 2, 10:55=A0am, DaveG wrote: > > > > > > > On Dec 1, 7:20=A0pm, "Main, Kerry" wrote: > > > > > -----Original Message----- > > > > From: Neil Rieck [mailto:n.ri...@sympatico.ca] > > > > Sent: December 1, 2008 3:04 PM > > > > To: Info-...@Mvb.Saic.Com > > > > Subject: OT: Multicore Is Bad News For Supercomputers > > > > > This article should not surprise anyone in this NG, but it is nice = to > > > > know others are working on a fix. > > > > > Multicore Is Bad News For Supercomputers > > > >http://spectrum.ieee.org/nov08/6912 > > > > > Neil Rieck > > > > Kitchener/Waterloo/Cambridge, > > > > Ontario, Canada. > > > >http://www3.sympatico.ca/n.rieck/ > > > > Its not only multi-cores that is the issue, but rather new buses like= Intel's > > > new QuickPath (formerly called CSI). > > > > The new X86/Itanium bus architecture is NUMA based and that is a new = paradigm > > > that for very high performance requires accessing local memory much m= ore than > > > remote memory. Hence, the OS and App's need to be aware and be able t= o maximize > > > perf with this architecture. > > > > Reference:http://www.intel.com/pressroom/archive/reference/whitepaper= _QuickPath... > > > > Btw, anyone on this newsgroup know of an OS that already has a great = deal of > > > experience With NUMA architectures? > > > > :-) > > > > Regards > > > > Kerry Main > > > Senior Consultant > > > HP Services Canada > > > Voice: 613-254-8911 > > > Fax: 613-591-4477 > > > kerryDOTmainAThpDOTcom > > > (remove the DOT's and AT) > > > > OpenVMS - the secure, multi-site OS that just works.- Hide quoted tex= t - > > > > - Show quoted text - > > > Most here know the answer to your question. =A0Do the "right" people @ > > HP know? =A0That's the real =A0question. > > > Dave... > > Does anyone at HP know? =A0There is surprisingly little customer demand > for openvms. =A0For example, SAS doesn't support openvms at all.- Hide qu= oted text - > > - Show quoted text - On the SAS subject.... A while back they said no more VMS support then came this earlier this year: +++++++++++ SAS delivers SAS 9.2 on HP OpenVMS Integrity Servers - hear what's new! You're invited to attend a joint HP / SAS Webinar to learn about this powerful combination of hardware and software. SAS, as the leader in business analytics software and services, is the first company to call when you need to solve complex business problems, achieve key objectives and more effectively manage your information assets. As you may now know, SAS recently delivered the SAS 9.2 (Foundation) products on OpenVMS for HP Integrity Servers. Many of our customers have been eagerly awaiting this release and are very excited - hear why. Join us and learn about HP's roadmap for OpenVMS and the exciting new features already available in the latest release and plans for future releases. OpenVMS on HP Integrity Servers provides an exciting combination of flexibility, lower TCO (total cost of ownership), and help IT Managers do more with less. ------------------------------ Date: Tue, 2 Dec 2008 20:37:56 +0000 From: "Main, Kerry" Subject: RE: Multicore Is Bad News For Supercomputers Message-ID: <9D02E14BC0A2AE43A5D16A4CD8EC5A593EDB45E92D@GVW1158EXB.americas.hpqcorp.net> > -----Original Message----- > From: DaveG [mailto:david.gudewicz@abbott.com] > Sent: December 2, 2008 3:04 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Multicore Is Bad News For Supercomputers > > On Dec 2, 1:38 pm, Christopher wrote: > > On Dec 2, 10:55 am, DaveG wrote: > > > > > > > > > > > > > On Dec 1, 7:20 pm, "Main, Kerry" wrote: > > > > > > > -----Original Message----- > > > > > From: Neil Rieck [mailto:n.ri...@sympatico.ca] > > > > > Sent: December 1, 2008 3:04 PM > > > > > To: Info-...@Mvb.Saic.Com > > > > > Subject: OT: Multicore Is Bad News For Supercomputers > > > > > > > This article should not surprise anyone in this NG, but it is > nice to > > > > > know others are working on a fix. > > > > > > > Multicore Is Bad News For Supercomputers > > > > >http://spectrum.ieee.org/nov08/6912 > > > > > > > Neil Rieck > > > > > Kitchener/Waterloo/Cambridge, > > > > > Ontario, Canada. > > > > >http://www3.sympatico.ca/n.rieck/ > > > > > > Its not only multi-cores that is the issue, but rather new buses > like Intel's > > > > new QuickPath (formerly called CSI). > > > > > > The new X86/Itanium bus architecture is NUMA based and that is a > new paradigm > > > > that for very high performance requires accessing local memory > much more than > > > > remote memory. Hence, the OS and App's need to be aware and be > able to maximize > > > > perf with this architecture. > > > > > > > Reference:http://www.intel.com/pressroom/archive/reference/whitepaper_Q > uickPath... > > > > > > Btw, anyone on this newsgroup know of an OS that already has a > great deal of > > > > experience With NUMA architectures? > > > > > > :-) > > > > > > Regards > > > > > > Kerry Main > > > > Senior Consultant > > > > HP Services Canada > > > > Voice: 613-254-8911 > > > > Fax: 613-591-4477 > > > > kerryDOTmainAThpDOTcom > > > > (remove the DOT's and AT) > > > > > > OpenVMS - the secure, multi-site OS that just works.- Hide quoted > text - > > > > > > - Show quoted text - > > > > > Most here know the answer to your question. Do the "right" people > @ > > > HP know? That's the real question. > > > > > Dave... > > > > Does anyone at HP know? There is surprisingly little customer demand > > for openvms. For example, SAS doesn't support openvms at all.- Hide > quoted text - > > > > - Show quoted text - > > On the SAS subject.... A while back they said no more VMS support > then came this earlier this year: > > +++++++++++ > SAS delivers SAS 9.2 on HP OpenVMS Integrity Servers - hear what's > new! > [snip...] An additional link of interest: "What's New in SAS 9.2 OpenVMS Companion" http://tinyurl.com/5agogx Or real link (beware wrap) http://support.sas.com/documentation/cdl/en/hostvms/61616/HTML/default/host= vmswhatsnew902.htm Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-254-8911 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Tue, 02 Dec 2008 15:23:29 -0700 From: Keith Parris Subject: Re: Multicore Is Bad News For Supercomputers Message-ID: Main, Kerry wrote: > Its not only multi-cores that is the issue, but rather new buses like Intel's > new QuickPath (formerly called CSI). > > The new X86/Itanium bus architecture is NUMA based and that is a new paradigm > that for very high performance requires accessing local memory much more than > remote memory. Hence, the OS and App's need to be aware and be able to maximize > perf with this architecture. It's NUMA, but not NUMA like you remember from the Wildire (GS-320, GS-160, GS-80) series, which had such severe performance problems because of the 3X difference in latency between local (QBB) and remote (across-QBB) memory accesses. Intel's QuickPath is NUMA of the EV7 flavor, which had remote access times only 1.something times as slow as local memory access, and where the slowest (farthest) memory access was still faster than the fastest memory access in a QBB-based design. ------------------------------ Date: Tue, 02 Dec 2008 19:05:04 -0800 From: "Tom Linden" Subject: Re: Open Source Message Oriented Middleware on OpenVMS? Message-ID: On Tue, 02 Dec 2008 05:32:17 -0800, Jan-Erik Söderholm wrote: > MessageQ was written in PL/I, and I know that BEA looked at rewritinbg > it > in C to port it to Itanium. I don't if that succeeded or not. > According to the BEA support site, there is a BMQ 5.0 kit > for OVMS Itanium... > I wonder if anyone is using it. I think there was not a lot of effort put into the rewrite. -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Tue, 02 Dec 2008 19:17:52 -0800 From: "Tom Linden" Subject: Re: OT: How We Found the Missing Memristor Message-ID: On Tue, 02 Dec 2008 10:05:53 -0800, John Smith (not the one @ HP) wrote: > > "Neil Rieck" wrote in message > news:a2896748-cdbf-4325-8850-f433b8dc7413@41g2000yqf.googlegroups.com... > On Dec 1, 11:57 pm, JF Mezei wrote: >> > Neil Rieck wrote: >> >> For all you hardware hackers: How We Found the Missing Memristor >> >> >>http://spectrum.ieee.org/dec08/7024 >> >> isn't that the type of research that won't happen again after Hurd >> scaled HP Labs down to focus on less esoteric stuff ? > > I'm not sure what the relationship is between HP corporate and Hp > labs, but the discovery of the invention could do two really cool > things: I will find out, one of my neighbors works in the labs > > 1) bring a ton of money to HP in the form of royalties (which could > lead to other future discoveries) > 2) revolutionize the electronics industry in a way not seen since the > invention of the transistor > > > Exactly. Like the revolutionary CYMK color ink droplet - which > automagically > adjusts the intensity of each component upon reaching the page based on > the > needs of the photo. Just like a cuttlefish. > > -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Tue, 02 Dec 2008 15:36:26 -0600 From: Michael Austin Subject: Re: RDB plugin/driver for Qt Message-ID: yyyc186 wrote: > Hello, > > I'm doing some work with Qt and have seen various messages about > people porting Qt to OpenVMS. What I haven't seen is any messages > about an RDB pluging/driver for Qt that would let an application > running on a 64-bit Ubuntu desktop access an RDB database being served > up by an OpenVMS cluster. > > Anyone know if such a beast exists? > > Thanks, > Roland Hmm... I don't really know what Qt is, but how about a jdbc connection? Unfortunately I have not used Rdb in 2+ years, but if memory serves correctly, they have done a lot of work with SQL SERVICES to be able to handle jdbc and if not, you should be able to use OCI (Oracle connection) by configuring the services correctly. ------------------------------ Date: Tue, 02 Dec 2008 20:24:31 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: RDB plugin/driver for Qt Message-ID: <4935dfc7$0$90266$14726298@news.sunsite.dk> Michael Austin wrote: > yyyc186 wrote: >> I'm doing some work with Qt and have seen various messages about >> people porting Qt to OpenVMS. What I haven't seen is any messages >> about an RDB pluging/driver for Qt that would let an application >> running on a 64-bit Ubuntu desktop access an RDB database being served >> up by an OpenVMS cluster. > > Hmm... I don't really know what Qt is, Possible: http://en.wikipedia.org/wiki/Qt_(toolkit) Arne ------------------------------ Date: Tue, 2 Dec 2008 17:12:59 -0800 (PST) From: FrankS Subject: Re: Synchronize with a remote ntp source Message-ID: <79a8669b-6e78-42ec-bdbb-5ceebf524799@33g2000yqm.googlegroups.com> On Dec 2, 11:13=A0am, "Richard B. Gilbert" wrote: > I'm not all sure that it CAN be done. =A0V6.2 is not an issue but the > version of TCP/IP Services (a/k/a UCX or the Ultrix Connection) could be > critical! > I don't think you're paying attention. Of course it can be done, just depends on which version of TCP/IP he's using on his OpenVMS system. Working from memory, it's something like: $ @sys$manager:tcpip$define_commands $ ntpdate -q time-a.nist.gov The -q will just query the remote NTP service (time-a.nist.gov) and display the difference between system time and the remote server's time. Remove the -q and the system time will get updated. It's very important that the time zone be set correctly. ------------------------------ Date: Tue, 02 Dec 2008 20:45:53 -0500 From: "Richard B. Gilbert" Subject: Re: Synchronize with a remote ntp source Message-ID: FrankS wrote: > On Dec 2, 11:13 am, "Richard B. Gilbert" > wrote: >> I'm not all sure that it CAN be done. V6.2 is not an issue but the >> version of TCP/IP Services (a/k/a UCX or the Ultrix Connection) could be >> critical! >> > > I don't think you're paying attention. Of course it can be done, just > depends on which version of TCP/IP he's using on his OpenVMS system. > Didn't I just say that?!?! AIRC, he didn't say which version of UCX he was running. > Working from memory, it's something like: > > $ @sys$manager:tcpip$define_commands > $ ntpdate -q time-a.nist.gov > > The -q will just query the remote NTP service (time-a.nist.gov) and > display the difference between system time and the remote server's > time. Remove the -q and the system time will get updated. > > It's very important that the time zone be set correctly. I've done that myself but I'm running VMS v7.2-1 and TCP/IP Services V5.1. It's very back level at this point but it works for me. My memory grows DIMM but AIRC NTP was not always there and did not always work. I don't recall if NTP was present and working in, say, UCX V3.3 or even V4.0. AIRC from UCX 2.7 through V3.3 ECO 14, we were damned lucky if ANYTHING worked! DEC ported the raw Berkeley code several YEARS after TGV, Wollongong and other major players had ported it. It was not an easy port to do and the early days of UCX were a rough ride indeed! ------------------------------ Date: Tue, 2 Dec 2008 11:27:05 -0800 (PST) From: steel_and_alum_engr Subject: Re: VMS, HP and the recession Message-ID: <184aa11b-ce4e-4fb8-8b6b-7ae446a7fefd@j35g2000yqh.googlegroups.com> On Dec 2, 10:33=A0am, JF Mezei wrote: > steel_and_alum_engr wrote: > > One benefit that we've enjoyed with virtual machines -- it's very easy > > to create one for testing purposes. =A0 Creating a VM is certainly > > easier than locating a spare piece of hardware, loading the OS, > > downloading patches & updates, loading applications, etc. > > But you can't test new harware cards, network interfaces etc without > taking down the whole server which means taking down all instances that > run on it. > > When you have a separare test machine, you can test hardware to your > hearts content without affecting production. Testing hardware is a whole 'nuther proposition. As usual, there are always exceptions. VMware isn't going to fit every situation, and it isn't going to provide broad support for everyone's specific hardware needs. If your situation involves a "special" board or interface, you might not be able to virtualize. We have a number of systems that fall into that category. For example, we have systems that use PCI reflective memory boards. Others have special fiber optic interface boards for high-speed data collection. These systems have not been (and probably never will be) virtualized. Other more garden-variety servers are more easily moved into the virtual world. As for networking, you can do considerable tweaking on individual VMs without any need of taking the host server down. In our experience, VMware's ESX seems very solid. (Probably not VMS solid, but solid.) But, your mileage may vary. I don't work for VMware and can only speak from my own experience -- which, so far, is quite positive. ------------------------------ Date: 2 Dec 2008 20:59:26 GMT From: billg999@cs.uofs.edu (Bill Gunshannon) Subject: Re: X-terminals Message-ID: <6pllteF8nde3U1@mid.individual.net> In article , tadamsmar writes: > On Dec 2, 8:58 am, Robin Schipper <"thespriteman wrote: >> tadamsmar schreef:> I need to buy 3 X-terminals for a DS-10 running VMS 7.3-2. >> >> > Any suggestions? >> >> > Thanks! >> >> hi, >> >> where are u from? >> >> we used to have some X-terminals, perhaps i can rescue a few from the scrap >> >> With Regards, >> >> Robin > I am in Chapel Hill, NC. > But I would like to look at some new ones, too. This upgrade is to > some extent cosmetic, but the ones you have might be good enough. > Thanks! Any reason not to use cheap PC's with Linux based X-terminal software installed? I have used thin-client software that supported both RDP and X11 quite nicely. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves billg999@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 2 Dec 2008 15:04:33 -0600 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: X-terminals Message-ID: tadamsmar schreef: > I need to buy 3 X-terminals for a DS-10 running VMS 7.3-2. > > Any suggestions? > > Thanks! We'll soon recommend that the government excess about half a dozen X terminals. If they follow that recommendation they will eventually auction them off, most likely in lots with a bunch of stuff you don't want. You'ld have to be local, and probably a US citizen. Most folks would just get a PC and load an X server on it. Some folks would get a Mac. ------------------------------ Date: 2 Dec 2008 21:06:10 GMT From: billg999@cs.uofs.edu (Bill Gunshannon) Subject: Re: X-terminals Message-ID: <6plma2F8nde3U2@mid.individual.net> In article , m.kraemer@gsi.de (Michael Kraemer) writes: > In article , > tadamsmar writes: >> On Dec 2, 8:58=A0am, Robin Schipper <"thespriteman wrote: >> > tadamsmar schreef:> I need to buy 3 X-terminals for a DS-10 running VMS 7= >> .3-2. >> > >> > > Any suggestions? >> > >> > > Thanks! >> > >> >> I am in Chapel Hill, NC. > > I remotely remember I attended a "Radiation Damage to DNA" conference > loong ago, in 1999 I think. Accommodated in Carolina Inn. > >> But I would like to look at some new ones, too. This upgrade is to >> some extent cosmetic, but the ones you have might be good enough. > > I think there are no "new" classical X-terminals any more, > unfortunately. Tektronix' business was eaten by NCD, who themselves > went belly up in 2000-something. Considering one of their last products, > the NC900, they deserved it. Ummm.... What makes you think they went belly-up? Not only are they still around in the thin-client world but they still list an X-terminal in their stable of products. > However, the classical products such as Explora or the XP series > are still usable, IMHO. Much less hassle than a Linux PC or those > so called thin-clients. Say what? I am looking at changing our labs to thn clients from ordinary PC's. I can netboot the PC's from a Windows or Unix box with trivial setup and then they just run. I can also put the image on the hard disk, on a CDROM or even on a USB stick. No hassle at all. > I have a couple of the classical hardware, but it is not worth it to > ship them across the pond. So you better watch out for some used > boxes in your area. I still have some HP and DEC X-terminals that I play with but their capabilities fall far short of what you can get from a thin client based system. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves billg999@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ End of INFO-VAX 2008.645 ************************