1 INFO-VAX	Fri, 21 Oct 2005	Volume 2005 : Issue 587       Contents:2 Re: "Jumbo Packets" and TCPIP Services for OpenVMS2 Re: "Jumbo Packets" and TCPIP Services for OpenVMS2 Re: "Jumbo Packets" and TCPIP Services for OpenVMS2 Re: "Jumbo Packets" and TCPIP Services for OpenVMS- Re: Can we please have ECP in .PCSI* format ?  Re: LASER symbiont Re: LASER symbiont Re: LASER symbiont" Re: Oracle 9i or 10g under itanium" Re: Oracle 9i or 10g under itanium) Re: OT: Is your HP printer spying on you? ) Re: OT: Is your HP printer spying on you? ) Re: OT: Is your HP printer spying on you?  Re: Porting VMS back to VAX ?  Re: Porting VMS back to VAX ?  Re: Porting VMS back to VAX ? / Re: Sun says HP irrelevant in enterprise space! / Re: Sun says HP irrelevant in enterprise space! = UCX performance on VMS 6.2 - Unexpected rise in CPU usage.... A Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage.... A Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....   F ----------------------------------------------------------------------  # Date: Thu, 20 Oct 2005 17:50:01 GMT / From: "Jeff Goodwin" <jgoodwin@maine.rrr-r.com> ; Subject: Re: "Jumbo Packets" and TCPIP Services for OpenVMS 8 Message-ID: <dNQ5f.73490$Xl2.57888@twister.nyroc.rr.com>  6 "Richard Brodie" <R.Brodie@rl.ac.uk> wrote in message , news:dj8gmn$rdg$1@blackmamba.itd.rl.ac.uk... > 1 > <dave.baxter@bannerhealth.com> wrote in message ? > news:1129825042.115117.303420@g49g2000cwa.googlegroups.com...  > C >> Q1.   Does TCPIP Services for OpenVMS, 5.4 ECO 4 support "Jumbo   >> Packets"???
 > Apparently.  > E >> Q2.   How would I be able to find out if I am already using them??  >  > MC LANCP SHOW DEVICE/PARAM > J >> Q3.   If it does (from Q1), and I am not (from Q2), how do I turn them  >> on? >  > See:G > http://h71000.www7.hp.com/doc/732final/6631/6631pro_006.html#gbe_perf  >   F Q4. If I turn on jumbo frames, is OpenVMS smart enough to negotiate a L smaller frame size when communicating to a peer that does not support jumbo  frames?    -Jeff    ------------------------------    Date: 20 Oct 2005 11:19:19 -0700" From: dave.baxter@bannerhealth.com; Subject: Re: "Jumbo Packets" and TCPIP Services for OpenVMS C Message-ID: <1129832359.740533.314410@g44g2000cwa.googlegroups.com>    Thanks Richard,   C I am now in a position to 'use' the Jumbo Frames, however I have to - cycle TCPIP Services before this will happen. D I really don't want to do this because my experience is that it willF not work correctly (i.e. using TCPIP$CONFIG to shutdown TCPIP ServicesG never seems to shut everything down and then it wont start because some G things are already running.)   I normally wait until I am in a position  to reboot the node.   E however this brought up an further question.    Wen the node was last A booted it was still connected to a 100MB Network segment, and the C 10/100/1000 NIC connected at 100 MB (so said the LED).   When I was B ready to move to GB, I disconnected the Network Cable, changed theA Device speed to 1000/full in LANCP, had Network move me to the GB C segment, and then reconnected the network cable to the NIC.    This G time it connected to the network at 1000/full (so said the LED).    Can D I take this a confirmation that I actually now have a 1GBit pipe, or) does this also require cycling of TCPIP??    Dave.    ------------------------------    Date: 20 Oct 2005 11:19:49 -0700" From: dave.baxter@bannerhealth.com; Subject: Re: "Jumbo Packets" and TCPIP Services for OpenVMS C Message-ID: <1129832389.200522.125920@o13g2000cwo.googlegroups.com>    Thanks Richard,   C I am now in a position to 'use' the Jumbo Frames, however I have to - cycle TCPIP Services before this will happen. D I really don't want to do this because my experience is that it willF not work correctly (i.e. using TCPIP$CONFIG to shutdown TCPIP ServicesG never seems to shut everything down and then it wont start because some G things are already running.)   I normally wait until I am in a position  to reboot the node.   E however this brought up an further question.    Wen the node was last A booted it was still connected to a 100MB Network segment, and the C 10/100/1000 NIC connected at 100 MB (so said the LED).   When I was B ready to move to GB, I disconnected the Network Cable, changed theA Device speed to 1000/full in LANCP, had Network move me to the GB C segment, and then reconnected the network cable to the NIC.    This G time it connected to the network at 1000/full (so said the LED).    Can D I take this a confirmation that I actually now have a 1GBit pipe, or) does this also require cycling of TCPIP??    Dave.    ------------------------------  # Date: Thu, 20 Oct 2005 19:57:16 GMT % From: Rick Jones <rick.jones2@hp.com> ; Subject: Re: "Jumbo Packets" and TCPIP Services for OpenVMS 4 Message-ID: <wES5f.15037$WK5.12044@news.cpqcorp.net>  . Jeff Goodwin <jgoodwin@maine.rrr-r.com> wrote:E > Q4. If I turn on jumbo frames, is OpenVMS smart enough to negotiate A > a smaller frame size when communicating to a peer that does not  > support jumbo frames?   D A comment not specific to VMS, but about Jumbo Frames and TCP/UDP in general....   C When the two TCP's exchange MSS options, if the other system is not C using Jumbo Frames, the "typical" MSS will be send and that is what E the TCP on the Jumbo Frames-enabled system will use.  That is, if the C other side doesn't also support Jumbo Frames, they will not be used  and there will be no benefit.   ? If you are using UDP, there is no exchange and so both ends and 8 everything in the middle _MUST_ understand Jumbo Frames.  D There is another feature of some GbE NICs (no idea if VMS TCP stacksA support it or not) called "large send" or "TSO" (TCP Segmentation C Offload).  This can be thought of as "poor man's" Jumbo Frames - in F this case, TCP is told it can have a virtual MSS up to 64KB or so, andC it is the NIC that further segments the TCP segment into a suitable D size.  The frames on the network then look just like they would withE the standard MTU, and neither the intermediate devices nor the remote D end system have to have support.  This drops CPU util on the sender," but does nothing for the reciever.  , Generally, it is not used for UDP, only TCP.  
 rick jones --  G oxymoron n, Hummer H2 with California Save Our Coasts and Oceans plates F these opinions are mine, all mine; HP might not want them anyway... :)D feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...   ------------------------------  % Date: Fri, 21 Oct 2005 04:15:25 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)6 Subject: Re: Can we please have ECP in .PCSI* format ?; Message-ID: <43584f3d.524144494f47414741@radiogaga.harz.de>   7 Peter 'EPLAN' LANGSTOEGER <peter@langstoeger.at> wrote: E > As I understand, one needs TDC on OpenVMS Alpha V8.2 for collecting L > performance data and ECP (V5.5A and up) for displaying them (graphically). > Right?   AFAIK, yes.   M > What I currently don't understand is the difference between TDC and TDC_RT. C > TDC_RT is installed with VMS V8.2, but what for? Collecting data? L > Why is then this other TDC product? Displaying data? I, thought, ECP is...@ > (On V7.3-2, there is no TDC_RT, only TDC, which collects data) > K > Or is TDC the TDC_RT plus the SDK? Do I need the SDK for collecting data? K > Probably not. So, TDC_RT is sufficient. But TDC gets updates. Will TDC_RT K > get updates, too? Probably not. So, is the use of TDC better than the use K > of TDC_RT? Likely. So, it is better to install TDC. And deinstall TDC_RT?    See Technical Journal #5:   F For logistical reasons, two distribution variants of the TDC software G exist: TDC_RT for the software installed with OpenVMS, and TDC for the  H software downloaded from the web site (URL at the end of this article). I TDC_RT contains the software required to run TDC on a specific platform,  B while TDC contains software to run TDC on any supported platform. E Assuming a common TDC build identification (for example, Version 2.1- I 70), there is no functional difference between the software installed by  D TDC_RT and the software that is installed on the same system by TDC.  > The downloadable TDC kit will be revised as needed to deliver E maintenance updates and functionality enhancements, while the TDC_RT  H variant will be updated only with new OpenVMS releases. Installation of B an updated TDC variant might or might not remove files previously H installed by TDC_RT, depending on your system configuration. Since both @ variants utilize the same startup file (TDC$STARTUP.COM), which E initializes the system environment to use the most recent TDC images  F available on the system, an older TDC_RT installation and a newer TDC I installation can safely co-exist. Do not remove TDC_RT from your system,  2 even if you have installed an updated TDC variant.   cu,    Martin --  B    Emacs would be a great   | Martin Vorlaender  |  OpenVMS rules!5    operating system,        | work: mv@pdv-systeme.de H    if only it came with     |   http://www.pdv-systeme.de/users/martinv/<    a decent editor...       | home: martin@radiogaga.harz.de   ------------------------------    Date: 20 Oct 2005 18:11:23 -0700/ From: "GeorgeC" <gconstantinides@myrealbox.com>  Subject: Re: LASER symbiont C Message-ID: <1129857083.267010.175650@o13g2000cwo.googlegroups.com>   * Thank you all for your considered replies.& It has given me some food for thought.  ;  1. It appears that the "LASER" symbiont is not that common <     any more in the VMS community and has not been worked on     for some time.  8  2. Some time ago I set up TELNETSYM symbiont queues for>     those printers and connected them directly to the network,=     eliminating the print servers, but when I found that some =     applications were not working, I quickly went back to the :     LAT (and print server) setup. Your suggestion that the9     TELNETSYM should be able to handle forms makes sense. ?     Maybe I had not specified the /LIBRARY qualifier correctly. <     What I am not sure about is whether the form definitions>     that are installed in the system are specific to the LASER?     symbiont  (/PROCESSOR=LASER). Some more investigating here. =     Some legacy applications do a VMS $COPY to the LAT device >     which will not work with a TCPIP queue, but that should be     easy to fix.  :  3. We do not do a lot of sophisticated printing from this@     box to warrant installing DCPS. I want to avoid the overheadB     of installing, configuring and maintaining yet another layered?     product, but if someone tells me it is dead easy to install .     and configure I will keep it as an option.  ! Thanks again for your suggestions    GeorgeC    GeorgeC wrote:	 > Hi all,  > / > I am running OpenVMS V7.2-2 on an Alpha 1200.  > ? > I have a few LAT print queues using the LASER print symbiont. A > These printers are connected to the network via Emulex (NETQUE) B > print servers.  The printers are fitted with ethernet interfacesA > so they are capable of being connected to the network directly.  > F > I want to connect the printers directly on the network and eliminateB > the print servers. I can do that by converting the LAT queues toH > TCPIP queues and use the default TCPIP print symbiont TCPIP$TELNETSYM.4 > In fact, I do this already with some of my queues. > However there is a snag... > I > User written applications rely on queue form definitions (LANDSCAPE,..) 4 > that do not work with the default TELNET symbiont. > G > What are my options?  At this stage I want to avoid using DCPS, which B > I now believe is free and included in the standard distribution. > G > Is there a more recent version of LASER (or similar) which works with  > TELNET queues? >  > Thanks > G.   ------------------------------  % Date: Thu, 20 Oct 2005 22:44:26 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>  Subject: Re: LASER symbiont , Message-ID: <435855FE.5B91B928@teksavvy.com>   GeorgeC wrote:< >  3. We do not do a lot of sophisticated printing from thisB >     box to warrant installing DCPS. I want to avoid the overheadD >     of installing, configuring and maintaining yet another layeredA >     product, but if someone tells me it is dead easy to install 0 >     and configure I will keep it as an option.  H The local squirrels say that it is dead easy to install and configure itE :-) (But I had to bribe them with some peanuts for them to say it ;-)     K http://h71000.www7.hp.com/doc/732final/documentation/pdf/dcps023_mguide.pdf   N Look at chapter 2. It provides a good overview of supported network protocols.  D If you have a supported printer and network connection, then DCPS isH really easy. The startup procedure contains documented examples for eachG type of network connection, so you just need to essentially fill in the ! blanks and uncomment those lines.   F If your printers are not postscript, then DCPS doesn't fit your needs.8 As of version 2.3, DCPS doesn't support telnet protocol.  G If you can get DCPS into your environment,  there are advantages. There H are many features that are enabled with DCPS which you may not need now.F But if some of your users ask "can you do this on your system", you'll be able to answer "YES".  F So DCPS allows you to increase the capabilities of your VMS system andG may help provent shifting work over to other platforms such as Windows.   C The installation itself is very simple. There are a couple of files L installed in the system and no dramatic heart transplant type of stuff done.   ------------------------------  # Date: Fri, 21 Oct 2005 03:19:41 GMT   From: John Santos <john@egh.com> Subject: Re: LASER symbiont , Message-ID: <h7Z5f.13888$2Y2.12720@trnddc05>   GeorgeC wrote:, > Thank you all for your considered replies.( > It has given me some food for thought.   <snip>  ? >     Some legacy applications do a VMS $COPY to the LAT device @ >     which will not work with a TCPIP queue, but that should be >     easy to fix.  < You can leave the setup of the LTA devices exactly the same, but then  )    $ set device/spooled=TCPIP_queue LTAx:   ? which will redirect them to the TELNETSYM (or whatever) queues.   = It doesn't matter if the physical LAT server doesn't exist or 0 goes away, since VMS never actually talks to it.  ? Especially if the the LTAx: devices are already spooled to your = current queues, this is transparent to the apps.  (If the app ; actually looks at device characteristics, e.g. to determine < formating or line width or something, it might have problems> because sys$getdvi will report on the spooling device (usually? a disk) rather than the spooled device.  However, it would have : the same problem with LTA's spooled to LAT queues, so this probably won't be an issue.)     --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------    Date: 20 Oct 2005 10:37:02 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) + Subject: Re: Oracle 9i or 10g under itanium , Message-ID: <imgV4gJCGz+n@malvm9.mala.bc.ca>  C In article <1129824347.026269.194300@f14g2000cwb.googlegroups.com>, +   "itanium" <ulloa.edgar@gmail.com> writes:  > G > Some one knows if oracle 9i or 10g is available to work with ovms 8.2  > or 8.2-1 under itanium..?  >   L   According to Oracle's website 10gR2 (both Enterprise and Standard edition)< is planned to be available for OVMS 8.2-1 Itanium in Q1CY06.  @   Don't see any mention of 9i so I'd suspect it won't be ported.  M  (disclaimer: That's just what I read, if you really want to know ask Oracle)    ------------------------------    Date: 20 Oct 2005 12:54:17 -0700; From: "johnhreinhardt@yahoo.com" <johnhreinhardt@yahoo.com> + Subject: Re: Oracle 9i or 10g under itanium C Message-ID: <1129838057.868040.139420@f14g2000cwb.googlegroups.com>    Malcolm Dunnett wrote:E > In article <1129824347.026269.194300@f14g2000cwb.googlegroups.com>, - >   "itanium" <ulloa.edgar@gmail.com> writes:  > > I > > Some one knows if oracle 9i or 10g is available to work with ovms 8.2  > > or 8.2-1 under itanium..?  > >  > N >   According to Oracle's website 10gR2 (both Enterprise and Standard edition)> > is planned to be available for OVMS 8.2-1 Itanium in Q1CY06. > B >   Don't see any mention of 9i so I'd suspect it won't be ported. > O >  (disclaimer: That's just what I read, if you really want to know ask Oracle)   C That's their plan, though they were about 4 months late (and 1 year B behind the UNIX and Windows versions) on the 10gR1 release for VMSG Alpha so it will be interesting to see if the 10gR2 release is anywhere ; near on time.  And there will be no VMS IA64 version of 9i.      John H. Reinhardt    ------------------------------  % Date: Thu, 20 Oct 2005 19:53:11 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: OT: Is your HP printer spying on you?0 Message-ID: <11lgb2bsejcc0cc@corp.supernews.com>   Z wrote: > Dave Froble wrote: > H >> I'm thinking that this may have been a reasonable request, and not a E >> requirement.  If 'requirement' is the correct word, then the next   >> revolution is way past due. >  >   > A revolution?! Overreact much?  G So, you're Ok with some government agency telling mfgs to secretly put  K things into their products, which you will then buy?  Where should I start?   ) 1) I'm paying for something I don't want. 1 2) The product advertising/specs are not correct. 
 And so on.  G Hey, if the government wants to use some of the ink, let them help pay   for the ink cartridges.   C Overreact?  No, I don't think so.  The people are suppost to be in  F charge here.  When that is no longer so, then it's once again time to  make it so.   F Understand, I have no problem with the concept.  If the package has a G notice that all printed material will have hidden coding, for whatever  G purposes, no problem.  When it's done without my knowledge, then there  
 is a problem.   C How about the RIAA?  Do you want them to be allowed to invade your  H computers?  Same principal.  I don't condone stealing, but I sure don't H condone some of the (in my view) 'illegal' methods used to prevent such.  A It's like Microsoft and XP (and beyond).  I don't like having to  F register with them to have the OS continue to operate.  But they tell A you up front, and why.  Then I have the option to comply, or not  E purchase.  It's when it's hidden, and you're not given options, that   it's wrong.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Thu, 20 Oct 2005 21:24:30 -0700  From: Z <Z@no.spam> 2 Subject: Re: OT: Is your HP printer spying on you?* Message-ID: <14_5f.9518$RG4.1084@fe05.lga>   JF Mezei wrote: I >>For one thing, it makes identifying the source easy and that could help  >>secure a criminal conviction.   I > If you are using a $10,000 printer to forge a million dollar fraud, and F > you know the printer leaves a watermark, then you simply destroy theJ > printer after you've printed your documents. They may be able to get the  H And if you're going to use a truck to try to blow up the WTC, you don't F rent it ... and you certainly don't go back to the rental company for  your deposit, right?  - Yet that's exactly what Mohammad Salameh did.   G I'm sure that back when carmakers started stamping car and truck parts  H with hidden VINs, some people were up in arms and talking about how the E government was spying on them and how criminals would just steal the  ( vehicles to make the hidden VINs futile.  A Yet those hidden VINs nailed both Salameh and the worst domestic    terrorist ever, Timothy McVeigh.  C You think too highly of the average criminal and average terrorist.     E > So in the end, criminals will learn not to store any information on I > their computers but rather hide it on the internet somwhere without any E > information in their computer on where it is hidden. So a search of F > their home/computer would not reveal any information on the location) > where the information is really stored.   G Yes, right about the time they start learning not to use rental trucks.     A > So, in the end, arresting those people without any evidence and F > detaining them for 90 days or whatever will not add to security. TheD > only way to catch them will be to get a warrant to intercept theirI > communications and find out where they store their information and then ) > arrest them once you have the evidence.   G WTF are you rambling about? This has nothing to do with the discussion   at hand.   ------------------------------  % Date: Thu, 20 Oct 2005 21:28:54 -0700  From: Z <Z@no.spam> 2 Subject: Re: OT: Is your HP printer spying on you?* Message-ID: <98_5f.9790$RG4.7380@fe05.lga>   Dave Froble wrote:I >>> I'm thinking that this may have been a reasonable request, and not a  F >>> requirement.  If 'requirement' is the correct word, then the next  >>> revolution is way past due.   ! >> A revolution?! Overreact much?   I > So, you're Ok with some government agency telling mfgs to secretly put  M > things into their products, which you will then buy?  Where should I start?   ' What is this, False Dichotomy Thursday?     E > Overreact?  No, I don't think so.  The people are suppost to be in    > You don't think it's an overreaction to start talking about a A revolution?! Are you 4-king nuts?! A REVOLUTION because of this?!    That's a HUGE overreaction!    ------------------------------  % Date: Thu, 20 Oct 2005 21:33:20 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> & Subject: Re: Porting VMS back to VAX ?: Message-ID: <ezX5f.21735$GH1.516878@news20.bellglobal.com>  5 "Dave Froble" <davef@tsoft-inc.com> wrote in message  * news:11lbgnb68fg1879@corp.supernews.com... > Neil Rieck wrote:  [...snip...] > G > Enough people, in positions that would allow them to speak with some  F > intimate knowledge, have made statements that EPIC was not going to L > challenge Alpha/Power, all else being equal.  Curly made sure that things I > wouldn't be equal for Alpha, but look at how Power continues to vastly   > outperform the itanic. > D > I remember reading a post from Paul in Australia that even the HP L > developers told their masters that they should change course.  Don't know B > where he got his info, so I cannot comment on it's authenticity. > I > So what?  If we can continue to get itanics, at reasonable prices, and  K > continue to run VMS, that's all that counts to me.  Alpha is truly dead,  H > and x86 is just one of JF's dreams, at this time.  My only concern is ) > being able to continue getting systems.  >  > --  6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > 170 Grimplin Road  > Vanderbilt, PA  15486   J I hear what you are saying and agree with most of it. I guess the point I G was trying to make was that there was a time when Alpha was new and it  H wasn't (yet) a good idea to jump from VAX. The price of Alpha came down G while the performance shot ahead of VAX which made most people jump to  . Alpha. I feel like I'm in the same boat today.  J On a related note, I contacted HP last week to get a quote on a brand new K Alpha Server DS25. The salesman did try to sell me an Itanium (we were not  J interested), but the quote for the DS25 came in yesterday and it was much  lower than I expected.  J Like you, I am more interested in OpenVMS than the underlying hardware. I L don't even care if I would be OpenVMS running on extended IA32 as long as I % could get RAM with parity protection.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Thu, 20 Oct 2005 22:04:58 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> & Subject: Re: Porting VMS back to VAX ?: Message-ID: <U0Y5f.21754$GH1.520170@news20.bellglobal.com>  ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  & news:43559F06.9C381681@teksavvy.com... > Neil Rieck wrote: K >> I was never convinced that RISC was better than CISC until I experienced J >> OpenVMS on Alpha. (CISC was much more important when RAM was expensive  >> but >> those days are long gone).  >  > 0 > What is the difference between CISC and RISC ? >  [...snip...]  1 I'm hoping that this is a rhetorical question :-)   7 If it is not, then here is my overly simplistic answer:   L When memory was expensive and slow, computer engineers wanted to read a few H instruction words from memory then have the CPU go off and do something M complicated (the VAX polynomial instruction comes to mind but they had other  H cool instructions like inserting entries into a linked list etc.). CISC G worked by translating the instruction just read from RAM into a longer  H sequence of microcode instructions internal to the CPU. This is usually A visible by looking at the clock driving the CPU's internal logic.   E One problem with CISC was not always being able to service important  M interrupts (like page fault) unless certain instructions were restartable so  L they did implement some on VAX. Designers started to think that, instead of I restarting instructions, it just might make more sense to load a form of  H microcode into RAM (now you were restarting a smaller instruction). The I original RISC design rule stipulated that each CPU instruction needed to  K execute in one clock cycle.This seemed OK since 90% of the time most CPU's  I where only doing load/store operations it's just that you needed more of  E these clock cycles to do the same job done by fewer cycles on a CISC  F machine. But hey, memory was getting faster and cheap, and many newer I hardware applications demanded faster load/store (RISC) CPUs. (If memory  A serves, RISC was originally designed with packet routing in mind)   F Simplifying CPU design allowed for the development of some other cool M features like out-of-order execution, speculative execution, etc. A Motorola  K engineer told be that RISC didn't mean "Reduced Instruction Set Computing"  3 it meant "Relegate Important Stuff to the Compiler"   " Those are the basic points anyway.  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Thu, 20 Oct 2005 23:26:41 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> & Subject: Re: Porting VMS back to VAX ?, Message-ID: <43585FE4.623CF304@teksavvy.com>   Neil Rieck wrote: G > Simplifying CPU design allowed for the development of some other cool N > features like out-of-order execution, speculative execution, etc. A MotorolaL > engineer told be that RISC didn't mean "Reduced Instruction Set Computing"5 > it meant "Relegate Important Stuff to the Compiler"     E You described some complexity in CISC that made RISC better. However, D were those unique to VAX or were they truly widespread in other CISCC architectures ? Seems most other CISC were *much* simpler than VAX.     B When you look at the fact that Intel implemented many stolen AlphaD features into its pentium 3 (and now does it freely), it seems to meE that many RISC tricks are also applicable to some CISC architectures.   C And since the 8086 has matured into a very respectable chip that is E right up there in terms of performance, isn't that a proof that it is , possible to make CISC compete against RISC ?  E Was it just VAX that was just so complex that it couldn't get all the  performance tricks ?   ------------------------------  % Date: Thu, 20 Oct 2005 19:56:52 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 8 Subject: Re: Sun says HP irrelevant in enterprise space!0 Message-ID: <11lgb9740pcq447@corp.supernews.com>   Bill Todd wrote: > Dave Froble wrote: >  >> Alan Greig wrote: >> >>> Dave Froble wrote: >>>  >>>> bob@instantwhip.com wrote:  >>>>/ >>>>> http://www.theinquirer.net/?article=27069  >>>  >>>  >>>  >>> I >>>> "Singer says Sun hears a consistent message from HP customers: "They G >>>> have no idea where HP is going." Even Intel, he says, has begun to K >>>> eschew the Itanium in favour of Montecito but when you talk to HP they " >>>> have one direction - Itanium" >>>  >>>  >>>  >>> H >>> It makes sense if you assume someone mixed up Montecito and Paxville >>> (the dual core Xeon).  >> >> >>J >> Hell of an assumption Alan.  Maybe for some, but this is suppost to be K >> someone intimate with computers.  Sounds like a real loose cannon to me.  >>E >> E-mailed the author of the piece.  Told him his work was a little  D >> shoddy, if all he could do is regurgutate Sun propaganda without 3 >> calling the guy on some of the obvious mistakes.  >  > G > Why do you assume that the mistake was made by Singer and not by the  J > article's author?  Hales does not regularly cover processors and is not I > always the sharpest tool in The Inquirer's shed:  if he were, he would  G > presumably have commented upon the inconsistency himself rather than  M > merely passed it on (let alone in a paraphrase rather than a direct quote).  >  > - bill  I Good point, and in my e-mail to him, I was a bit more detailed and asked   that question.  B For all I know, he wasn't actually able to ask any such questions.   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Thu, 20 Oct 2005 21:35:56 -0400 ( From: Bill Todd <billtodd@metrocast.net>8 Subject: Re: Sun says HP irrelevant in enterprise space!= Message-ID: <_7ednaBrdbCd28XeRVn-tg@metrocastcablevision.com>    Dave Froble wrote: > Bill Todd wrote: >  >> Dave Froble wrote:  >> >>> Alan Greig wrote:  >>>  >>>> Dave Froble wrote:  >>>>  >>>>> bob@instantwhip.com wrote: >>>>> 0 >>>>>> http://www.theinquirer.net/?article=27069 >>>> >>>> >>>> >>>> >>>>J >>>>> "Singer says Sun hears a consistent message from HP customers: "TheyH >>>>> have no idea where HP is going." Even Intel, he says, has begun toL >>>>> eschew the Itanium in favour of Montecito but when you talk to HP they# >>>>> have one direction - Itanium"  >>>> >>>> >>>> >>>> >>>>I >>>> It makes sense if you assume someone mixed up Montecito and Paxville  >>>> (the dual core Xeon). >>>  >>>  >>>  >>> H >>> Hell of an assumption Alan.  Maybe for some, but this is suppost to I >>> be someone intimate with computers.  Sounds like a real loose cannon  
 >>> to me. >>> F >>> E-mailed the author of the piece.  Told him his work was a little E >>> shoddy, if all he could do is regurgutate Sun propaganda without  4 >>> calling the guy on some of the obvious mistakes. >> >> >>H >> Why do you assume that the mistake was made by Singer and not by the G >> article's author?  Hales does not regularly cover processors and is  H >> not always the sharpest tool in The Inquirer's shed:  if he were, he I >> would presumably have commented upon the inconsistency himself rather  E >> than merely passed it on (let alone in a paraphrase rather than a   >> direct quote).  >>	 >> - bill  >  > K > Good point, and in my e-mail to him, I was a bit more detailed and asked   > that question. > D > For all I know, he wasn't actually able to ask any such questions.  I My point was that he would have commented upon it *in the article*.  The  G lack of comment there suggests that he was clueless about the matter -  E hence could easily have screwed it up himself by misstating what had  D been said to him (e.g., simply substituted Montecito for Xeon while B typing it in:  The Inquirer doesn't seem to do much in the way of  proofreading, you know).   - bill   ------------------------------    Date: 20 Oct 2005 11:23:15 -0700( From: "Lawrie" <stroker_ace@hotmail.com>F Subject: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....C Message-ID: <1129832595.676578.305030@z14g2000cwz.googlegroups.com>    Hi,   D I have recently upgraded a single threaded legacy application on VMS 6.2.  D The application originally used two serial ports (using QIOs)to sendG data. I modified it to use multiplexed TCP/IP sockets implemented using  the DEC C socket.h library.   B If I compare the CPU usage of the serial and TCP/IP versions of myF application I notice that the TCP/IO version uses nearly twice as much CPU at the same data rates.   E Does anyone know if this is an overhead of using UCX and/or the DEC C B socket library or is the phenomenom is more likely to be caused by3 inefficient coding in my particular implementation?     Many thanks for any help/advice.   Lawrie.    ------------------------------    Date: 20 Oct 2005 13:30:53 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) J Subject: Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....3 Message-ID: <g8VUI2ygUxYI@eisner.encompasserve.org>   n In article <1129832595.676578.305030@z14g2000cwz.googlegroups.com>, "Lawrie" <stroker_ace@hotmail.com> writes:  F > I have recently upgraded a single threaded legacy application on VMS > 6.2. > F > The application originally used two serial ports (using QIOs)to sendI > data. I modified it to use multiplexed TCP/IP sockets implemented using  > the DEC C socket.h library.  > D > If I compare the CPU usage of the serial and TCP/IP versions of myH > application I notice that the TCP/IO version uses nearly twice as much > CPU at the same data rates.  > G > Does anyone know if this is an overhead of using UCX and/or the DEC C D > socket library or is the phenomenom is more likely to be caused by5 > inefficient coding in my particular implementation?   B To get closest to the same performance, your rewritten application should also use QIOs.   D But make sure you are not calling QIO separately for each character.   ------------------------------  % Date: Thu, 20 Oct 2005 20:15:31 -0400 ' From: Dave Froble <davef@tsoft-inc.com> J Subject: Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....0 Message-ID: <11lgcc7l804ioe1@corp.supernews.com>  
 Lawrie wrote:  > Hi,  > F > I have recently upgraded a single threaded legacy application on VMS > 6.2. > F > The application originally used two serial ports (using QIOs)to sendI > data. I modified it to use multiplexed TCP/IP sockets implemented using  > the DEC C socket.h library.  > D > If I compare the CPU usage of the serial and TCP/IP versions of myH > application I notice that the TCP/IO version uses nearly twice as much > CPU at the same data rates.  > G > Does anyone know if this is an overhead of using UCX and/or the DEC C D > socket library or is the phenomenom is more likely to be caused by5 > inefficient coding in my particular implementation?  > " > Many thanks for any help/advice. > 	 > Lawrie.  >   E TCP/IP is not a particularly 'slim' protocol.  For example, LAT uses  C less resources.  Your old communications, direct QIOs, is probably    largly data and little overhead.  K If the overhead isn't a problem, then use it since it provides flexibility.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------   End of INFO-VAX 2005.587 ************************