1 INFO-VAX	Thu, 22 Jun 2006	Volume 2006 : Issue 344       Contents:) Re: Business Week: Tech Has A New Top Dog - Re: console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch 4 how to turn off OPCOM messages on a console terminal8 Re: how to turn off OPCOM messages on a console terminal8 Re: how to turn off OPCOM messages on a console terminal8 Re: how to turn off OPCOM messages on a console terminal8 Re: how to turn off OPCOM messages on a console terminal8 Re: how to turn off OPCOM messages on a console terminal RE: Hung System RWMBX state  Re: Hung System RWMBX state % New OpenVMS Virtualization Whitepaper 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... + Re: Power supply for Alphastation 400 4/233 . Re: Setting up my NIC for Gb Ethernet traffic.. Re: Setting up my NIC for Gb Ethernet traffic.. Re: Setting up my NIC for Gb Ethernet traffic.. Re: Setting up my NIC for Gb Ethernet traffic. surplus DEGPA-SA  F ----------------------------------------------------------------------  % Date: Wed, 21 Jun 2006 19:56:39 -0500 6 From: "David J. Dachtera" <djesys.no@spam.comcast.net>2 Subject: Re: Business Week: Tech Has A New Top Dog- Message-ID: <4499EAC7.48DD5@spam.comcast.net>    bob@instantwhip.com wrote: > 5 > Does this mean they can now advertise OpenVMS more?    I'm sure they CAN.   Question is, will they?   A To quote (more or less) from Tony Robbins, "What people CAN do is = incredible. What people WILL do is frequently disappointing".    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  + Date: Wed, 21 Jun 2006 20:28:37 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: console message involving RUN/STOP switch$ Message-ID: <e7ca5l$k2h$2@online.de>  J > >I have an expansion box with 5 TOSHIBA CD drives in it.  Since the 2100J > >has two tape drives and a CD drive, that is one device too many.  I setC > >the IDs of the CDs in my box such that there is only one SCSI-ID K > >conflict.  When I do this, and put the CD in one of the drives in my box I > >(connected to the SCSI connector near the top of the 2100, not the one J > >on the card in the expansion slot---I don't see anything at the consoleG > >if I connect to it), the machine starts to boot, but stops after the  > >message   > >  > >   jumping to bootstrap code   E The explanation was much simpler: I had the wrong serial port.  I had G been able to access the console via the other serial port, so I thought - that I had the console port, but not really.    G I've seen something similar when the console is set to graphics and one G is working on the serial terminal or vice versa: basic console commands H work, but no more after booting.  If both are connected and switched on,G then it is obvious what is going on, since during boot the stuff starts H showing up on the other screen.  Apparently, a similar arrangement holdsF for the two serial connectors.  On VAX, though, I think that only one F serial connector works as the console, even before one boots into VMS.  G In other words, either one works for the ALPHA console, but only one of E them (labelled "1" and located below the one labelled "2") is the VMS  console.  E It then booted up OK.  However, when connected to the "wrong" serial  F port, I was surprised that even after having time enough to boot, one F couldn't log in here.  (At least some ALPHAs can boot with no console H switched on at all, and one can then log into them from elsewhere after  they have booted.)    + I had set the console to serial, of course.   I Is there any way, when at the console prompt before booting, to find out  E if this is the correct serial port for the "VMS console" (i.e. where  H startup messages are written out etc)?  Or does one just have to know?  H Is there any convention (e.g. the one labelled "1")?  (On VAX, it's the # one with the picture of a printer.)    ------------------------------  # Date: Thu, 22 Jun 2006 02:23:15 GMT   From: John Santos <john@egh.com>6 Subject: Re: console message involving RUN/STOP switch+ Message-ID: <nanmg.26788$YI2.8582@trnddc01>    Alan Adams wrote: * > In message <e7b3eu$22b$1@news.mdx.ac.uk>+ >           david20@alpha2.mdx.ac.uk wrote:  >  > F >>In article <e79q4b$bko$3@online.de>, helbig@astro.multiCLOTHESvax.de4 >>(Phillip Helbig---remove CLOTHES to reply) writes: >>F >>>helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to >>>reply) writes:  >>>  >>> I >>>>I'm testing out some old hardware, an ALPHAserver 2100.  It powers up / >>>>fine, the console sees all the devices etc.  >>>>: >>>>I wanted to boot it off of a CD, but I get the message >>>>O >>>>   dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch " >>>>   failed to open dka600.6.0.1 >>>>O >>>>   dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch " >>>>   failed to open dka600.6.0.1 >>>>C >>>>I tried a 7.3-2 CD and a 7.3-1 CD.  Both are copies, but I have J >>>>successfully booted off of them in other machines in the past.  The CDI >>>>drive in the 2100 is an RRD43.  I have an original 7.1 CD but I can't  >>>>find it at the moment. >>> K >>>I'm pretty sure now that it is due to the fact that I am using copies of J >>>CDs and not original CDs.  (Unfortunately, I can't find my original 7.1K >>>CD.)  In terms of data, my copies are exact copies.  The problem is that D >>>the original CDs have small pits for the bits and the copies haveH >>>different reflectivity.  Old drives can't read copies.  Some somewhatG >>>newer drives can read them but they won't work as a boot drive.  New I >>>drives don't have any problems at all.  (I have installed and upgraded & >>>VMS many times from copies of CDs.) >>> J >>>I have an expansion box with 5 TOSHIBA CD drives in it.  Since the 2100J >>>has two tape drives and a CD drive, that is one device too many.  I setC >>>the IDs of the CDs in my box such that there is only one SCSI-ID K >>>conflict.  When I do this, and put the CD in one of the drives in my box I >>>(connected to the SCSI connector near the top of the 2100, not the one J >>>on the card in the expansion slot---I don't see anything at the consoleG >>>if I connect to it), the machine starts to boot, but stops after the 
 >>>message >>>  >>>  jumping to bootstrap code >>> J >>>Like I said, there are two devices with the same SCSI ID on the bus andB >>>also the cable is quite long.  Either or both of these could beJ >>>contributing to the problem (perhaps along with the fact that the driveK >>>is perhaps not completely able to read the copy CD, though I'm sure that 5 >>>it can handle it if it is not being used to boot).  >>> I >>>Unfortunately, trying a shorter cable would mean getting the heavy box F >>>down off of a high shelf and removing one of the devices on the busG >>>would mean opening up the expansion box or the 2100 itself.  I don't  >>>have time to test this now. >>>  >>M >>I'd suspect that is what is causing the RUN/STOP problem with the CD Drive. O >>I had this happen on my PWS 600au home system just recently but this was with P >>real VMS OS CDs not copies. Hence I thought the CD drive had just given up the* >>ghost and wasn't able to read the media.H >>I purchased a replacement drive from Island however at the same time II >>purchased a new dual port SCSI card since I was using both internal and H >>external drives on the inbuilt scsi card (which according to something
 >>I'd read >  > I > That riungs a distant bell. I think the controller is terminated. That  I > means it must be at the end of the cable, and if you use both internal  G > and external cables, it is in the middle. You then end up with three  E > terminators, and poor signal quality. The result can be apparently  = > random, and things work if both cables are kept very short.  >   B Similar experience with an 8mm tape drive.  Daisy-chained it to anG external shelf with a bunch of RZ29's.  Forgot to remove the terminator C in the shelf, so the bus was triple-terminated.  It worked fine for ? about a year until someone who will remain nameless (not me!!!) B attempted to jam a tape into the drive upside-down.  DEC/Compaq/HPD (forget who at the time) replaced the 8mm drive, but the replacementG was very flaky.  We went through about 4 new drives before I discovered E the triple-termination (I suspected something was loose.)  Pulled out G the extra terminator and it's worked fine ever since.  (Actually, we've F been through another couple of drives, but that was mechanical.  8mm's+ are cheap, and you get what you pay for :-)     H > I could be wrong, but the distant memory is of some similar effect on " > a 3100 Vax - about 10 years ago. >  >   & This was on an Alphaserver 1200, FWIW.    N >>you shouldn't be able to do - though it had been working fine (including theG >>CD drive until recently). Also I wanted to add in a few more drives).  >>H >>When I installed the new scsi card so that the internal drives were on
 >>one scsiK >>connection and the external on another then the original CD drive started  >>working again. >> >> >>David Webb >>Security team leader >>CCSS >>Middlesex University >> >> >> >> >> >> >> >>H >>>Any serious hardware errors would have showed up at the console or onH >>>the alphanumeric display on the machine itself.  The fact that it gotB >>>this far indicates to me that there is no real problem with it. >>>  >  >  >      --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  + Date: Wed, 21 Jun 2006 17:59:32 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)= Subject: how to turn off OPCOM messages on a console terminal $ Message-ID: <e7c1e4$11j$1@online.de>  I Although it is not always strictly necessary, at least in my experience,  I I think it is recommended to leave the console switched on.  I do this.   D Sometimes, the stuff printed there (even when one is not logged in)  might be interesting.   B I'm now experiencing a break-in attempt via FTP (several tries per@ second with the same username).  I've restarted FTP, which in myB experience breaks off the attack (and set the remote address to beC rejected, complain to the owner of said address etc).  However, the G screen is several minutes behind.  Thus, I would like to switch off the A OPCOM messages.  I could do this if I could log in to the console G terminal, but I can't do that.  I can, of course, log in to the machine  over the network.   G I just discovered that SET TERM/NOBROAD does the trick, but a) only if  I one adds the /PERMANENT qualifier.  Also, this shuts out ALL broadcasts;  5 is there any way to shut out just the OPCOM messages?   ( Interestingly, set term/broad opa0: says  4    %SET-W-NOPERM, permanent qualifier not specified   & while set term/nobroad opa0: does not.   ------------------------------  # Date: Wed, 21 Jun 2006 19:36:02 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>A Subject: Re: how to turn off OPCOM messages on a console terminal 1 Message-ID: <Cchmg.2145$4c6.789@news.cpqcorp.net>   / Phillip Helbig---remove CLOTHES to reply wrote: K > Although it is not always strictly necessary, at least in my experience,  K > I think it is recommended to leave the console switched on.  I do this.   F > Sometimes, the stuff printed there (even when one is not logged in)  > might be interesting.   G    It can be, and there are a couple of messages that are only sent to  G the console.  (Though to be cautious, these same critical messages can  I sometimes be lost within a blizzard of OPCOM messages, so many folks can  4 and do turn the default console OPCOM displays off.)  D > I'm now experiencing a break-in attempt via FTP (several tries per" > second with the same username).   3    I tend to use a firewall for this sort of thing.   G    One of the disadvantages of having OpenVMS right at the edge of the  D abyss is the denial of service attack or its cousin the distraction D (covering) attack, where the blizzard of logging can either trigger I problems (breakin evasion or disk capacity limits, for instance), or can  ; seek to cover an actual attack within a blizzard of audits.   I > I just discovered that SET TERM/NOBROAD does the trick, but a) only if  K > one adds the /PERMANENT qualifier.  Also, this shuts out ALL broadcasts;  7 > is there any way to shut out just the OPCOM messages?   H    Um, there's this document that I maintain -- you might have heard of F it -- that has all sorts of interesting information on OpenVMS within ? it.  Answers to various questions that seem to arise with some  G frequency.  There might well be something for this question, too, as I  I have a vague recollection of it.  Now if I could just remember what that  H pesky document is called or what the URL might be, we'd be all set.  :-)  ;    But seriously, I think the following FAQ text will help:   D >           8.4  Using REPLY/LOG from DCL? Disabling Console OPCOMs? > J >                    Your terminal must be enabled as an operator terminalH >                    before the REPLY/LOG command can be used, but a DCLH >                    procedure (batch command file, system startup, etc)G >                    does not have an associated terminal. To make this I >                    work, use the following sequence to enable the OPA0: I >                    console as the operator terminal, then the REPLY/LOG . >                    command will be accepted: > 5 >                    $ DEFINE/USER SYS$COMMAND _OPA0:   >                    $ REPLY/LOG5 >                    $ DEFINE/USER SYS$COMMAND _OPA0: # >                    $ REPLY/ENABLE  > I >                    To disable the system console terminal (OPA0:) as an B >                    operator terminal, use the following command: > 5 >                    $ DEFINE/USER SYS$COMMAND _OPA0: $ >                    $ REPLY/DISABLE > J >                    Also see SYLOGICALS.COM (and SYLOGICALS.TEMPLATE) forG >                    information on configuring the behaviour of OPCOM, F >                    including the (default) use of the system consoleF >                    (OPA0:) as an operator terminial and the specificK >                    contents and behaviour of the system operator log file " >                    OPERATOR.LOG.   ------------------------------  + Date: Wed, 21 Jun 2006 20:23:12 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)A Subject: Re: how to turn off OPCOM messages on a console terminal $ Message-ID: <e7c9rg$k2h$1@online.de>  > In article <Cchmg.2145$4c6.789@news.cpqcorp.net>, Hoff Hoffman" <hoff-remove-this@hp.com> writes:   M > > Although it is not always strictly necessary, at least in my experience,  M > > I think it is recommended to leave the console switched on.  I do this.   H > > Sometimes, the stuff printed there (even when one is not logged in)  > > might be interesting.  > I >    It can be, and there are a couple of messages that are only sent to  I > the console.  (Though to be cautious, these same critical messages can  K > sometimes be lost within a blizzard of OPCOM messages, so many folks can  6 > and do turn the default console OPCOM displays off.)  % Via the logical name OPC$OPA0_ENABLE?   F > > I'm now experiencing a break-in attempt via FTP (several tries per$ > > second with the same username).  > 5 >    I tend to use a firewall for this sort of thing.  > I >    One of the disadvantages of having OpenVMS right at the edge of the  F > abyss is the denial of service attack or its cousin the distraction F > (covering) attack, where the blizzard of logging can either trigger K > problems (breakin evasion or disk capacity limits, for instance), or can  = > seek to cover an actual attack within a blizzard of audits.   I In general, I want to be accessible via FTP.  As I said, once an address  E does this, it gets blocked forever.  (I suspect this is some sort of  I virus; can any hacker be so stupid as not to notice that he is making no  E progress?  Nothing else fishy seems to be going on at the same time.)   K > > I just discovered that SET TERM/NOBROAD does the trick, but a) only if  M > > one adds the /PERMANENT qualifier.  Also, this shuts out ALL broadcasts;  9 > > is there any way to shut out just the OPCOM messages?  > J >    Um, there's this document that I maintain -- you might have heard of H > it -- that has all sorts of interesting information on OpenVMS within A > it.  Answers to various questions that seem to arise with some  I > frequency.  There might well be something for this question, too, as I  K > have a vague recollection of it.  Now if I could just remember what that  J > pesky document is called or what the URL might be, we'd be all set.  :-) > = >    But seriously, I think the following FAQ text will help:  > F > >           8.4  Using REPLY/LOG from DCL? Disabling Console OPCOMs? > > L > >                    Your terminal must be enabled as an operator terminalJ > >                    before the REPLY/LOG command can be used, but a DCLJ > >                    procedure (batch command file, system startup, etc)I > >                    does not have an associated terminal. To make this K > >                    work, use the following sequence to enable the OPA0: K > >                    console as the operator terminal, then the REPLY/LOG 0 > >                    command will be accepted: > > 7 > >                    $ DEFINE/USER SYS$COMMAND _OPA0: " > >                    $ REPLY/LOG7 > >                    $ DEFINE/USER SYS$COMMAND _OPA0: % > >                    $ REPLY/ENABLE   B Right, but I want to keep the logging, just turn off the messages.  K > >                    To disable the system console terminal (OPA0:) as an D > >                    operator terminal, use the following command: > > 7 > >                    $ DEFINE/USER SYS$COMMAND _OPA0: & > >                    $ REPLY/DISABLE  # Yes, that's what I was looking for!    ------------------------------  % Date: Wed, 21 Jun 2006 16:33:12 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> A Subject: Re: how to turn off OPCOM messages on a console terminal , Message-ID: <4499AD01.6DD95A77@teksavvy.com>   Hoff Hoffman wrote: F > > I'm now experiencing a break-in attempt via FTP (several tries per# > > second with the same username).  > 5 >    I tend to use a firewall for this sort of thing.   F The fireall won't prevent some hacker from initiating a series of whatH could have been valied FTP connections but whcih turn out to be a a bonaH fide password cracking attempt.  FTP does trigger the breaking detectionD and evasion, so it is not a threath to VMS security, but those OPCOMH messages are a real pain. (and as you say, can hide a true attack hidden( amongst the billions of opcom messages).  A By the time you get to your router and add that IP to the list of D blocked IPs, there are still a few hundred million attempts done andG OPCOM is sending each to a 9600 baud VT terminal with smooth scrolling, E so you're still hearing bells for a long time after the attempts stop : while OPCOM tries push that data out to the slow terminal.      7 > >                    $ DEFINE/USER SYS$COMMAND _OPA0: & > >                    $ REPLY/DISABLE    D In a scenario where the audit server (or whoever) has generated 1000F opcom messages which would take say 10 minutes to push out to a VT200,G and you issue the REPLY/DISABLE command after 1 minute, would the OPCOM G messages that were already made (but not yet transmitted to that serial H port) be zapped, or would the command take effect for new OPCOM messages> that would no longer be queued for delivery to that terminal ?    B Also I noticed that if you hold-screen a terminal for long enough,G broadcasts to it eventually vaporize and when to unfreeze the terminal, F you won't see them.  Is this a terminal driver feature ? Would it also? apply to OPCOM messages ? And more importantly, is that timeout , documented and adjustable via some logical ?   ------------------------------  # Date: Wed, 21 Jun 2006 20:49:48 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>A Subject: Re: how to turn off OPCOM messages on a console terminal 2 Message-ID: <Mhimg.2146$je6.1922@news.cpqcorp.net>  / Phillip Helbig---remove CLOTHES to reply wrote: @ > In article <Cchmg.2145$4c6.789@news.cpqcorp.net>, Hoff Hoffman$ > <hoff-remove-this@hp.com> writes:  >> ...so many folks can 8 >> and do turn the default console OPCOM displays off... > ' > Via the logical name OPC$OPA0_ENABLE?   D    Correct.  Or via the (older) REPLY command mechanism.  Both work.  I    The OPCOM mechanisms discussed in SYLOGICALS provide more flexibility  A than does the "bigger hammer" that is the REPLY command, as (for  C instance) you could choose to display specific subset of the OPCOM  = messages displayed at the console using SYLOGICALS mechanism.    ------------------------------  % Date: Wed, 21 Jun 2006 17:03:42 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> A Subject: Re: how to turn off OPCOM messages on a console terminal , Message-ID: <4499B426.42F9F045@teksavvy.com>   Hoff Hoffman wrote: J >    The OPCOM mechanisms discussed in SYLOGICALS provide more flexibilityB > than does the "bigger hammer" that is the REPLY command, as (forD > instance) you could choose to display specific subset of the OPCOM? > messages displayed at the console using SYLOGICALS mechanism.     G Instead of pointing you to the FAQ , I'll point you to another document  that is often mentioned:   	$HELP  :-)     H HELP REPLY shows that you can disable specific classes of opcom messages) just like the logical schemes in SYLOGIN.    aka REPLY/DISABLE=SECURITY   ------------------------------    Date: 21 Jun 2006 14:47:51 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) $ Subject: RE: Hung System RWMBX state3 Message-ID: <E5+Rz7g5+o$+@eisner.encompasserve.org>   m In article <Pine.LNX.4.61.0606211040480.25385@localhost.localdomain>, Rob Brown <mylastname@gmcl.com> writes: . > On Wed, 21 Jun 2006, Farrell, Michael wrote: > B >> When we tried to log in, it logged us in, in the sense that it G >> clearly accepted your userid and password, gave the welcome message  ! >> but never got to a DCL prompt.  > H > I think that is also consistent with PAGEFILE full.  When it happened H > to me (more than once, I am a slow learner), I did eventually get the ' > DCL prompt, but it took many minutes.   C    It could simply be a performance problem due to that COM process 
    he saw.  E    When I knew I was going to have to investigate issues like this on E    my VAXen, I set the login priority of SYSTEM to 31, then turned it "    down in sys$manager:login.com .  E    So normally when I logged in I went right to a reasonable priority G    before anyone else got hurt.  But when the system was "hung" I'd log !    in /nocommand and look around.   A    If the pagefile is full and you have a console, you will see a D    message saying so.  RWMBX is not indicative of pagefile full, theF    crash dump should show lots of processes in RWMPW or something like    that.   ------------------------------  % Date: Wed, 21 Jun 2006 16:47:12 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Hung System RWMBX state, Message-ID: <4499B048.32F3CF42@teksavvy.com>   Suggestion:    SDA> SHOW MAILBOX/STATUS=RWMBX  C This would scan through all processes in RWMBX status, and find out E which mailbox devices are causing the RWMBX status, and for each such E mailbox, would list all processes that have channels assigned to that G mailbox and whether their status is RMBBX because of an attempted write  to that mailbox.  D Obviously, if the server process doesn't have a read QIO issued, theG system won't be able to tell you which server process isn't reading the E messages from the mailbox, but you'd still see that server process as 2 having a channel assigned to that problem mailbox.  U (or perhaps this could be done outside of $ANA/SYSTEM with $ANA/MAILBOX/STATUS=RWMBX)    ------------------------------  % Date: Wed, 21 Jun 2006 20:53:34 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com> . Subject: New OpenVMS Virtualization WhitepaperT Message-ID: <FA60F2C4B72A584DBFC6091F6A2B86840161352F@tayexc19.americas.cpqcorp.net>   All -=20  > The following may be of interest to this list: (includes a few% references to VMS V8.3 and Montecito)   F http://h71000.www7.hp.com/openvms/whitepapers/index.html - see OpenVMS Virtualization WP. Also at: 8 http://h71028.www7.hp.com/ERC/downloads/4AA0-5801ENW.pdfD This white paper outlines how HP OpenVMS is embracing virtualizationF technologies contributing to the HP Adaptive Enterprise initiative, so  that customers can begin to planE integration of their current and future OpenVMS environments into the 9 Adaptive Enterprise and Virtual Server Environment (VSE).    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------    Date: 21 Jun 2006 14:42:32 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) > Subject: Re: OT: IA64's speed beaten by orders of magnitude...3 Message-ID: <ACmHpKLXpo+A@eisner.encompasserve.org>   [ In article <epcmg.2133$u_5.1545@news.cpqcorp.net>, John Reagan <john.reagan@hp.com> writes:  > I > I noticed that they didn't say what kind of chip.  I doubt it was some  ; > Power CPU, right?  What was it?  Some silly JK flip-flop?  > D > Making a "simple" chip run at 350Ghz isn't like making a Power or  > Itanium run at 350Ghz.  A    Back when small VAXen were getting up to 20 VUP we saw a video ?    clip at a War Stories session which revealed that DEC has an 0    experimental chip doing over one billion IPS.  G    A couple years later DEC shipped Alphas.  I think the beta shipments C    were over 750 MIPS and when I bought the smallest, slowest Alpha     DEC made it ran over 1 GIPS.   E    Tomorrow or Friday I'll see a youngster I know who designs PPC for F    IBM and see what I can pump him for without getting him in trouble.   ------------------------------  % Date: Wed, 21 Jun 2006 16:40:58 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude..., Message-ID: <4499AED3.C2217653@teksavvy.com>   Bob Koehler wrote:G >    Tomorrow or Friday I'll see a youngster I know who designs PPC for H >    IBM and see what I can pump him for without getting him in trouble.  H You might wish to review that wording.... It could easily be interpretedG into a very different way which would get YOU into a lot of trouble :-)  :-) (pumping a youngster)        Back to topic: ##H The experiment is part of a project to explore the ultimate speed limitsE of silicon-germanium (SiGe) chips. SiGe chips are similar to standard E silicon chips, but they also contain germanium for better performance  and lower power consumption.   ##  F Since this mentions a "chip",  and later on discusses wafers, it wouldG appear to be more than a simple transistor. But there isn't any mention " of the chip doing any useful work.    F BTW, when you get to such low temperatures, wouldn't crystals oscilateC at slower speed ? (and at 0K, wouldn't crystals stop oscilating ?).    ------------------------------    Date: 21 Jun 2006 13:43:00 -0700- From: "Doug Phillips" <dphill46@netscape.net> > Subject: Re: OT: IA64's speed beaten by orders of magnitude...C Message-ID: <1150922580.775019.324560@g10g2000cwb.googlegroups.com>    Bob Koehler wrote:] > In article <epcmg.2133$u_5.1545@news.cpqcorp.net>, John Reagan <john.reagan@hp.com> writes:  > > J > > I noticed that they didn't say what kind of chip.  I doubt it was some= > > Power CPU, right?  What was it?  Some silly JK flip-flop?  > > E > > Making a "simple" chip run at 350Ghz isn't like making a Power or  > > Itanium run at 350Ghz. > C >    Back when small VAXen were getting up to 20 VUP we saw a video A >    clip at a War Stories session which revealed that DEC has an 2 >    experimental chip doing over one billion IPS. > I >    A couple years later DEC shipped Alphas.  I think the beta shipments E >    were over 750 MIPS and when I bought the smallest, slowest Alpha ! >    DEC made it ran over 1 GIPS.  > G >    Tomorrow or Friday I'll see a youngster I know who designs PPC for H >    IBM and see what I can pump him for without getting him in trouble.    B Seems the Georgia Tech article has given away most of the secrets;( (Also see link posted by Richard Brodie)  C < http://gtresearchnews.gatech.edu/newsrelease/half-terahertz.htm >    ##F "The silicon-germanium heterojunction bipolar transistors built by theB IBM-Georgia Tech team operated at frequencies above 500 GHz at 4.5D Kelvins (451 degrees below zero Fahrenheit) - a temperature attained? using liquid helium cooling. At room temperature, these devices E operated at approximately 350 GHz. Performance measurements were made = using a specialized high-frequency test system in the Georgia  Electronic Design Center.   G The devices used in the research are from a prototype fourth-generation D SiGe technology fabricated at IBM on a 200-millimeter wafer using anD older un-optimized mask set. Simulations suggest that the technologyA could ultimately support much higher (near-Terahertz) operational 1 frequencies at room temperature, Cressler said. "  ##   ------------------------------  % Date: Wed, 21 Jun 2006 17:00:55 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude..., Message-ID: <4499B37F.A21E9C03@teksavvy.com>   Doug Phillips wrote:D > Seems the Georgia Tech article has given away most of the secrets;* > (Also see link posted by Richard Brodie) > E > < http://gtresearchnews.gatech.edu/newsrelease/half-terahertz.htm >    Thanks for the pointer.   E I can just see a new burgeonning industry: liquid helium delivered to C homes so that the teenagers can boost the speed of their PC to play  faster games :-)  F What I find most significant is that they are able to run that chip atE 350ghz at room temperature.  What is not said however is whether that E chip had samples from all the building blocks of a CPU (OR, XOR, AND, $ NAND  (are there others I forgot ?))    H Of you can only implement "AND" at that speed, then it isn't very useful for building CPUs.   ------------------------------  % Date: Wed, 21 Jun 2006 14:28:02 -0700 # From: "Tom Linden" <tom@kednos.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude...) Message-ID: <op.tbilw0emzgicya@hyrrokkin>   / On Wed, 21 Jun 2006 14:00:55 -0700, JF Mezei  =   % <jfmezei.spamnot@teksavvy.com> wrote:    > Doug Phillips wrote:E >> Seems the Georgia Tech article has given away most of the secrets; + >> (Also see link posted by Richard Brodie)  >>F >> < http://gtresearchnews.gatech.edu/newsrelease/half-terahertz.htm > >  > Thanks for the pointer.  > G > I can just see a new burgeonning industry: liquid helium delivered to E > homes so that the teenagers can boost the speed of their PC to play  > faster games :-) > I > What I find most significant is that they are able to run that chip at=   G > 350ghz at room temperature.  What is not said however is whether that G > chip had samples from all the building blocks of a CPU (OR, XOR, AND, & > NAND  (are there others I forgot ?))  I Article didn't say other than it was built on a 200 mm wafer, which migh=  t  indicate more complex circuity   >  > I > Of you can only implement "AND" at that speed, then it isn't very usef=  ul > for building CPUs.   ------------------------------    Date: 21 Jun 2006 14:48:11 -0700- From: "Doug Phillips" <dphill46@netscape.net> > Subject: Re: OT: IA64's speed beaten by orders of magnitude...A Message-ID: <1150926491.541124.41340@r2g2000cwb.googlegroups.com>    JF Mezei wrote:  > Doug Phillips wrote:F > > Seems the Georgia Tech article has given away most of the secrets;, > > (Also see link posted by Richard Brodie) > > G > > < http://gtresearchnews.gatech.edu/newsrelease/half-terahertz.htm >  >  > Thanks for the pointer.  > G > I can just see a new burgeonning industry: liquid helium delivered to E > homes so that the teenagers can boost the speed of their PC to play  > faster games :-) > H > What I find most significant is that they are able to run that chip atG > 350ghz at room temperature.  What is not said however is whether that G > chip had samples from all the building blocks of a CPU (OR, XOR, AND, & > NAND  (are there others I forgot ?)) >   J > Of you can only implement "AND" at that speed, then it isn't very useful > for building CPUs.    = Silicon-germanium is fast, but expensive. And, just because a F transistor *can* switch at 350Ghz doesn't mean that any CPU built fromC SiGe will run at that speed. Cell phones are a big SiGe user now, I  think.  G The logic gate switching time would be relative to (among other things) G the number of components in series, but each like transistor would have G the same speed. What I get from the article is that they were measuring  transistor switching speed.   > Maybe one of the chip guys will jump in and explain it better.   ------------------------------  % Date: Wed, 21 Jun 2006 17:48:59 -0400 2 From: Bob Willard <BobwBSGS@TrashThis.comcast.net>> Subject: Re: OT: IA64's speed beaten by orders of magnitude...: Message-ID: <MPydnQdD_vjcIgTZnZ2dnUVZ_sudnZ2d@comcast.com>   Tom Linden wrote: 0 > On Wed, 21 Jun 2006 14:00:55 -0700, JF Mezei  ' > <jfmezei.spamnot@teksavvy.com> wrote:  >  >> Doug Phillips wrote:  >>F >>> Seems the Georgia Tech article has given away most of the secrets;, >>> (Also see link posted by Richard Brodie) >>> G >>> < http://gtresearchnews.gatech.edu/newsrelease/half-terahertz.htm >  >> >> >> Thanks for the pointer. >>H >> I can just see a new burgeonning industry: liquid helium delivered toF >> homes so that the teenagers can boost the speed of their PC to play >> faster games :-)  >>I >> What I find most significant is that they are able to run that chip at H >> 350ghz at room temperature.  What is not said however is whether thatH >> chip had samples from all the building blocks of a CPU (OR, XOR, AND,' >> NAND  (are there others I forgot ?))  >  > K > Article didn't say other than it was built on a 200 mm wafer, which might   > indicate more complex circuity >  >> >>K >> Of you can only implement "AND" at that speed, then it isn't very useful  >> for building CPUs.  >  >   N The Georgia Tech PR article referenced above said the device was a transistor.K And, since GT & IBM were going for bragging rights, I suspect the frequency M claimed is the alpha cutoff frequency of the transistor -- rather higher than ' one could reliably clock complex logic.   # But still, very impressive results.  --   Cheers, Bob    ------------------------------    Date: 21 Jun 2006 18:43:21 -0700$ From: "AEF" <spamsink2001@yahoo.com>> Subject: Re: OT: IA64's speed beaten by orders of magnitude...B Message-ID: <1150940601.886310.59440@g10g2000cwb.googlegroups.com>   JF Mezei wrote:  [...]  >  > H > BTW, when you get to such low temperatures, wouldn't crystals oscilateE > at slower speed ? (and at 0K, wouldn't crystals stop oscilating ?).   E Well, cooling down to liquid-helium temperatures probably changes the B physical properties of the chip/transistor/whatever A LOT. Perhaps< there is even some type of phase transition somewhat akin to water-->ice.  C At 0K, which you can't achieve anyway, you'd still have the quantum C mechanical "zero-point" energy. Remember the uncertainty principle: B delta-x * delta-p .ge. ~ h-bar. (p is momentum which is mass timesG velocity) And while h-bar (Plank's constant divided by 2*pi) is really, D really, really small, it is not zero. IOW, if you stop the vibrationB completely you can't satisfy the uncertainty principle because youB would know exactly both the position and the momentum: delta-x andD delta-p would both be zero, violating the law. Knowing exactly whereD something is would mean you have no idea what its momentum is, whichC means the next time you look it will most likely be somewhere else. C (And if you did find a particle, how would you know it was the same G one!) Similarly, if you knew the momentum exactly, you'd not be able to G tell where it is. (These are idealized extreme cases given just to make C the point, which is the following:  Quite generally, the better you E know the poistion, the worse you can determine the momentum, and vice  versa.)   D This is why the energy of the ground state (state of lowest possibleF energy) of a harmonic oscillator is 1/2 * h-bar * frequency, not zero.  ( Ah..., the wonders of quantum mechanics.   ------------------------------  % Date: Wed, 21 Jun 2006 23:22:54 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude..., Message-ID: <449A0CEE.56B6501B@teksavvy.com>  
 AEF wrote:E > At 0K, which you can't achieve anyway, you'd still have the quantum E > mechanical "zero-point" energy. Remember the uncertainty principle:   E The uncertanty principle did not take into account the possibility to H look at particles from subspace with the help of a dilithium matrix that. is coupled to a matter/anti-matter engine. :-)  G Besides, if you dip your clock and cpu into 0K water (with antifreeze, E of course) and it stops working, then you know for sure the clock has  stopped its oscilations :-)    ------------------------------  + Date: Tue, 20 Jun 2006 19:23:02 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 4 Subject: Re: Power supply for Alphastation 400 4/233( Message-ID: <e79hum$1pg$1@pcls4.std.com>  " Z <SpamDumpst3r@yahoo.com> writes:  L >Is the power supply in an Alphastation 400 4/233 a vanilla AT power supply?  G You have to replace the motherboard plugs, but an ATX power supply will E fit and work, I've done it.  The wire color codes are different, too.    ------------------------------    Date: 21 Jun 2006 14:29:23 -0700" From: dave.baxter@bannerhealth.com7 Subject: Re: Setting up my NIC for Gb Ethernet traffic. B Message-ID: <1150925363.102276.46360@y41g2000cwy.googlegroups.com>   Colin,F        I use two GB NICs (Fiber) for Cluster Interconnect SCS traffic.F   Are they going to be bothered by setting the LAN_FLAGS parameter (to? 64 I guess), (since it applies to ALL GB Ethernet Adapters) ???    Dave   Colin Butcher wrote:L > I suspect that you need to enable Jumbo frames before starting DECnet (and2 > maybe before starting SCS if this is a cluster). > J > The SYSGEN parameter LAN_FLAGS (bit 6) should enable Jumbo frames. LANCPL > DEFINE might be a good thing too. Unfortunately my test systems are out of/ > service at the moment, so I can't check this.  >  > -- >  > Hope this helps, Colin. + > colin DOT butcher AT xdelta DOT co DOT uk G > It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------    Date: 21 Jun 2006 14:29:28 -0700" From: dave.baxter@bannerhealth.com7 Subject: Re: Setting up my NIC for Gb Ethernet traffic. C Message-ID: <1150925368.506908.173680@c74g2000cwc.googlegroups.com>    Colin,F        I use two GB NICs (Fiber) for Cluster Interconnect SCS traffic.F   Are they going to be bothered by setting the LAN_FLAGS parameter (to? 64 I guess), (since it applies to ALL GB Ethernet Adapters) ???    Dave   Colin Butcher wrote:L > I suspect that you need to enable Jumbo frames before starting DECnet (and2 > maybe before starting SCS if this is a cluster). > J > The SYSGEN parameter LAN_FLAGS (bit 6) should enable Jumbo frames. LANCPL > DEFINE might be a good thing too. Unfortunately my test systems are out of/ > service at the moment, so I can't check this.  >  > -- >  > Hope this helps, Colin. + > colin DOT butcher AT xdelta DOT co DOT uk G > It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------  % Date: Wed, 21 Jun 2006 14:38:36 -0700 # From: "Tom Linden" <tom@kednos.com> 7 Subject: Re: Setting up my NIC for Gb Ethernet traffic. ) Message-ID: <op.tbimemgszgicya@hyrrokkin>   I On Wed, 21 Jun 2006 14:29:23 -0700, <dave.baxter@bannerhealth.com> wrote=  :    > Colin,I >        I use two GB NICs (Fiber) for Cluster Interconnect SCS traffic.=   I >   Are they going to be bothered by setting the LAN_FLAGS parameter (to=   A > 64 I guess), (since it applies to ALL GB Ethernet Adapters) ???  >  > Dave >  > Colin Butcher wrote:I >> I suspect that you need to enable Jumbo frames before starting DECnet=    =    >> (and 3 >> maybe before starting SCS if this is a cluster).  >>I >> The SYSGEN parameter LAN_FLAGS (bit 6) should enable Jumbo frames. LA=  NCP I >> DEFINE might be a good thing too. Unfortunately my test systems are o=  ut  =    >> of 0 >> service at the moment, so I can't check this. >> >> --  >> >> Hope this helps, Colin., >> colin DOT butcher AT xdelta DOT co DOT ukI >> It's not mine, but I like this definition: Legacy =3D stuff that work=  s. >   I BTW, I have 3 DEGPA-SA cards which I don't need, if anyone is interested=  , 1 before I put them on ebay, send me a mail offline  Tom    ------------------------------  # Date: Wed, 21 Jun 2006 22:58:06 GMT % From: Rick Jones <rick.jones2@hp.com> 7 Subject: Re: Setting up my NIC for Gb Ethernet traffic. 2 Message-ID: <2akmg.2150$b86.1406@news.cpqcorp.net>  @ Colin Butcher <colin_DOT.butcher_AT@xdelta_dot.co_dot.uk> wrote:E > Shouldn't be an issue. AFAIK then if the board and switch can agree A > on Jumbo frames then they will do so, else they'll fall back to   > normal ethernet sized frames.   F Is there really a negotiation for JumboFrame?  My recollection is thatC there is not.  Are you perhaps thinking of TCP MSS exchange between > _hosts_ where when both ends are JumboFrame enabled, they willD exchange the larger MSS, but if one is not the connection falls-back to the "standard" MSS?  
 rick jones --  B No need to believe in either side, or any side. There is no cause.E There's only yourself. The belief is in your own precision.  - Jobert 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: Wed, 21 Jun 2006 15:05:04 -0700 # From: "Tom Linden" <tom@kednos.com>  Subject: surplus DEGPA-SA ) Message-ID: <op.tbinmqjtzgicya@hyrrokkin>   < Have three if anyone is interested before I put them on ebay Tom    ------------------------------   End of INFO-VAX 2006.344 ************************