1 INFO-VAX	Mon, 23 Apr 2001	Volume 2001 : Issue 226       Contents:, Re: 1GHz Alpha EV68CB Microprocessor Exposed, Re: 1GHz Alpha EV68CB Microprocessor Exposed, Re: 1GHz Alpha EV68CB Microprocessor Exposed! Re: Compaq (Workstations to NASA)  DFWDAYS a Big Success  FTP failed!  Info-VAX filtering RE: Janitor fixes 90L B Re: looking for VMS-related hardware, particularly VXTs, in Europe) Re: Management Musical Chairs in the BCSG ) Re: Management Musical Chairs in the BCSG / RE: OpenVMS article - please explain last line! / RE: OpenVMS article - please explain last line!  Re: OT: "The Puzzle Palace"  Re: OT: "The Puzzle Palace"  Re: OT: "The Puzzle Palace"  SAS on VMS vs Windoze 2000 Terminal driver in V7.2  Re: Terminal driver in V7.2 , Re: Using 3-phase power converters on VAXen?, Re: Using 3-phase power converters on VAXen? Re: WIS... RIP???  -- NO!  Re: WIS... RIP???  -- NO! - Re: WIS... RIP???  -- NO! - non-standard port   F ----------------------------------------------------------------------  # Date: Sun, 22 Apr 2001 22:48:17 GMT  From: Dirk Munk <munk@home.nl>5 Subject: Re: 1GHz Alpha EV68CB Microprocessor Exposed ' Message-ID: <3AE35FAF.6E1994DE@home.nl>    daytripper wrote:   C > On Fri, 20 Apr 2001 06:00:29 GMT, Dirk Munk <munk@home.nl> wrote:  > >cjt & trefoil wrote:  > > H > >> To crossover threads slightly with another one here on comp.os.vms,G > >> is it going to require a three phase power converter to drive one?  > > Q > >If you mean the ES45 , the answer is no AFAIK. It looks very much like a ES40, P > >and that runs on single phase 230V of here in Europe. Although it can have upO > >to 4 (?) power supplies, and it is possible to connect the power supplies to M > >different phases. So in case one phase goes down, you can keep the machine P > >running on the other phase. Over here a 19" cabinet usualy has 2 power cablesQ > >with 32A CEE form single phase plugs coming from its switchbox, and we connect $ > >these cables to different phases. > ( > fwiw, wrt es4x rack or pedestal units: > N > - up to three "bulk" power supplies; and anything bigger than a quad *needs*O > two supplies, with the third providing n+1 current sharing front-end beasties  > per crate  > M > - depending on memory population & #processors, each (of up to three) cords O > into an ES4x crate can draw enough power that (for most services) it actually 7 > *requires* the use of a unique mains branch per cord.   Q I don't see why. A ES40 power supply is rated at 730W, and with 230V that means a  current of appr. 3.5A.R A normal fuse is 16A, so you could easily run a complete ES40 on one mains branch.       >  > O > - and unless *three* substation sources converge at the crate, only uni's and P > duals could ever hope to obtain 100% mains redundancy (artifact of the triplesJ > and quad's requiring two supplies be *always* functional). With only two@ > station  service, you'd be rolling the dice on every blackout. > I > btw, to the person that asked: any >1g CuEV is a bitch to air-cool in a M > daughterboard (ala es4x) configuration. You really wouldn't want to drop an   > ES45 heat sink on your foot... > 7 > /daytripper ("long" in aluminum and copper shares ;-)  >  > >  > >> Terry C Shannon wrote:  > >> >, > >> > On 20 Apr 2001, Paul Repacholi wrote: > >> >8 > >> > > JF Mezei <jfmezei.spamnot@videotron.ca> writes: > >> > > > >> > > > Dirk Munk wrote:  > >> > >O > >> > > > > We know that 1.6 GHz alpha's already were in existence more than a N > >> > > > > year ago. So producing a reliable 1 GHz system should be possibleO > >> > > > > by know. AFAIK the 1.01 GHz ES45 should be out in a few weeks from  > >> > > > > now.  > >> > >( > >> > > That is good news, if it is so. > >> > >N > >> > Alas, it does not appear to be the way things are gonna be. ES45s do inM > >> > fact exist (have seen same, back in January IIRC) but the original May K > >> > launch date has been pushed out to 4CQ01. Apparently it's supply and N > >> > demand: all the early system orders have been placed by HPTC customers,F > >> > and CPQ can't satisfy the existing orders until late this year. > >> >O > >> > Rather than announce now and deliver in November, CPQ chose to delay the P > >> > announcement. There obviously is more to the matter than just speeding upK > >> > the assembly lines, else CPQ would be doing just that. Especially in 5 > >> > light of what's happening in the Wintel space.  > >> >N > >> > > BTW, anyone know the details of the package for the new 264? It seems= > >> > > to be quite a change from what has been used so far.  > >> > > > >> >P > >> > Well, it weighs more than its predecessor and it's larger. Also difficult: > >> > to de-lid or peel apart as was done for the photos. > >> >O > >> > This is the second version of EV68B I've seen. The one that fell into my F > >> > hands back in October Y2K did not have pins stickin' out of it.   ------------------------------  # Date: Mon, 23 Apr 2001 00:28:33 GMT 0 From: "Yousuf Khan" <ykhan@nospam.home.com.spam>5 Subject: Re: 1GHz Alpha EV68CB Microprocessor Exposed < Message-ID: <RGKE6.85850$61.17687644@news4.rdc1.on.home.com>  : "H.W. Stockman" <stockman3@earthlink.net> wrote in message? news:hgED6.13486$2S6.122077@newsread1.prod.itd.earthlink.net...  > H > "Larry Kilgallen" <Kilgallen@eisner.decus.org.nospam> wrote in message/ > news:5RRkFRazwVih@eisner.encompasserve.org... L > > In article <3ADE7CEE.42FE3B79@home.nl>, Dirk Munk <munk@home.nl> writes:I > > Certain weird segments of the customer base have indicated they would G > > rather have products they buy fully tested.  This is not the market  > > segment of "overclockers". > , > What market segment buys Sun workstations?  H The Ecache problems with the Ultrasparcs aren't because of overclocking,. it's just underengineering that caused it. :-)  K BTW, that brings up a question in my mind. Who exactly buys Alpha machines, I and what do they use them for? The Tru64 doesn't even show up on the Unix K radar screen compared to Sun, HP and IBM. Is there another Unix marketplace ' out there that I haven't even heard of?        Yousuf Khan    ------------------------------  # Date: Mon, 23 Apr 2001 02:36:02 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net>5 Subject: Re: 1GHz Alpha EV68CB Microprocessor Exposed = Message-ID: <myME6.30372$%_1.8587516@typhoon.ne.mediaone.net>   L "Yousuf Khan" <ykhan@nospam.home.com.spam> wrote in message news:RGKE6.85850C > BTW, that brings up a question in my mind. Who exactly buys Alpha 	 machines, K > and what do they use them for? The Tru64 doesn't even show up on the Unix A > radar screen compared to Sun, HP and IBM. Is there another Unix  marketplace ) > out there that I haven't even heard of?   I I can think of two: the Tru64 installed base (yes, there is one), and the B HPTC marketplace. APAC, the French CEA, the Japanese Atomic EnergyK Commission, ASCI Q, Celera Genomics, the Sangre Centre, NSF/Pittsburgh come  to mind.   ------------------------------  % Date: Sun, 22 Apr 2001 14:17:17 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> * Subject: Re: Compaq (Workstations to NASA)' Message-ID: <3AE32E3D.EA1F76DC@fsi.net>    Wayne Sewell wrote:  > [snip]M > I'm actually kind of surprised.  This is the first indication I've had that G > anyone in compaq even recognized/understood/acknowledged the problem.   F On second read, perhaps I should clarify: the text quoted in the prior1 post was from me to Richard, not Richard's words.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  # Date: Mon, 23 Apr 2001 02:43:06 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net> Subject: DFWDAYS a Big Success= Message-ID: <_EME6.30407$%_1.8591981@typhoon.ne.mediaone.net>   I The third DFWDAYS conference is now history and by all accounts the event F was a big success. Thanks to the folks from the DFWCUG, their partnersH (Tecsys, Sequel Data Systems, Attunity, Digital Networks, Digital India,K etc, etc, etc), and Compaq Computer Corporation. And very special thanks to H Sue Skonetski for marshalling the impressive pool of OpenVMS engineering1 talent who contributed immensely to DFWDAYS 2001.   ' And lest we forget, thanks to HAL 9000!    cheers,   
 Charlie Matco K Currently writing his latest epic: CHARLIE MATCO AND THE DFW TEMPLE OF DOOM    ------------------------------  % Date: Mon, 23 Apr 2001 10:25:51 +0800 5 From: Netsurfer <netsurfer@sentosa.singaporemail.com>  Subject: FTP failed!8 Message-ID: <ti47etcf5edvcl10df0neaj5eo5k3gpjt3@4ax.com>  > It seems to me that the FTP to OpenVMS will fail if I had some6 conditional statements in my SYLOGIN.COM login script.  F Is there a way to allow FTP access to captive users (bound by menus)??          Regards,  	 Netsurfer         ====R For any personal email replies, please remove " sentosa. " from my E-mail address.   ------------------------------  % Date: Sun, 22 Apr 2001 14:50:41 -0400 ' From: "Bill Todd" <billtodd@foo.mv.com>  Subject: Info-VAX filtering ( Message-ID: <9bv94n$2qp$1@pyrite.mv.net>  I I just replied with the message appended below and received the following  automated response:     9 Trend SMEX Content Filter has detected sensitive content.   , Place = Info-VAX@Mvb.Saic.Com; ; ; Bill Todd Sender = Bill Todd3 Subject = Re: Management Musical Chairs in the BCSG 0 Delivery Time = April 22, 2001 (Sunday) 12:41:26 Policy = Dirty Words( Action on this mail = Quarantine message  # Warning message from administrator: 7 Sender, Content filter has detected a sensitive e-mail.     K Now, as I read c.o.v. as a newsgroup, I don't care much what censorship may I be applied by Info-VAX, but some of the rest of you might.  I've placed a H couple of strategic asterisks in what seemed likely the offending places$ below and am now going to try again.   - bill    9 "Paul Repacholi" <prep@prep.synonet.com> wrote in message ' news:87pue4n7h0.fsf@prep.synonet.com...    ...   F > Any one is going to use as low a cost systems as is reasonable. WhatE > is harder to pick is when someone nedds, or has a good call for the D > 'big box' function, but not the grunt. So he is screwed, and oftenH > knows that what he needs is there, just can't use it with out a wad of( > $$. How to p*ss off customers, step 1.  L This is called supply and demand (or, alternatively, value), something we inI the U.S. often consider to be a Good Thing.  It motivates both innovation F and self-improvement by 'have-nots' to become 'haves' so that they canI purchase such goodies, and at the same time is mitigated by the fact that L anything that is *in actuality* over-priced for its value will be the target2 of competitive products at more reasonable prices.   > F > > The real problem at this point may be hardware pricing: as long asH > > the hardware platform costs at least twice as much as a high-qualityE > > Intel (or better, AMD) platform of somewhat similar capabilities, G > > it's not clear that the VMS license pricing is all that much out of C > > line, though reducing it wouldn't hurt (again, assuming there's G > > currently not enough volume in that segment to generate much profit  > > anyway). > C > With hard ware, you have real money at stake. You can't stamp out D > DS10Es for 20c each! H*ll, even if the box was Tooth Faired to theH > door step for nix, there is still a lot of money in stocking, shippingF > and installing.  Plus spares inventory. Real money, mind, money that > must be paid to someone. > H > Licences are essentially free. Some arbitary amount of admin overhead,E > and a bit of paper. So it should be of the order of $25-50 plus and G > extra $1-2 per extra licence in one hit. That is COST, real money. So G > if you give some one a $20,000 licence, you are $50 out of pocket. If G > you give him a $3000 box, that will be, on VMS current margins, $1500  > out of pocket.  I But if you sell, say, 10,000 licenses at cost ($50), but would still have K sold, say, 2000 of them at $20,000 apiece, you're out (really out:  this is ( money you would have had) $39.5 million.  I That's what value is all about, and it's only tenuously related to actual E manufacturing/handling cost (though it does have something to do with K overall cost, including development, because that's what allows competition I to spring up if you price things too much out of proportion).  And that's K why the platform cost *is* relevant, because as long as the license cost is F not too unreasonable as a percentage of the total system cost it won't" affect sales volume all that much.   > C > (Note that I'm just handwaving the media kits, OK Given a descent 4 > pressing run, the cost also falls into the noise.) > G > So how do you spice things up at the moment? Cut to the bone where it H > will not COST you, even if that throws away potential revenue.  If you8 > do it well, that money WILL come back in a few months,  I That's easy for you to say:  it's not your business on the line if you're  wrong.  F If Compaq in fact *is* generating significant revenue with its currentI entry-level systems and pricing, then the portion of that revenue derived F from license prices is at risk.  And whether it would make up anywhereG nearly that amount in increased hardware revenue (actually, profit from K those increased sales, and perhaps from later service) is a function of the J detailed shape of the demand curve, which it's extremely unlikely you haveH any real insight into at that level of detail (even very thorough market+ analysis can get it wrong, and often does).   H Conversely, if no significant revenue is being generated in this segmentF today, then there's no significant risk to dropping license prices and+ seeing what happens.  Which is what I said.    - bill   ------------------------------  % Date: Mon, 23 Apr 2001 12:54:35 +0800 + From: "Davies, Huw" <Huw.Davies@compaq.com>  Subject: RE: Janitor fixes 90LO Message-ID: <5AD7782D2FB2D11183CE0000F8023ACD04146058@meoexc1.dhcp.meo.dec.com>   " True story - at a former employer!  J About 4 hours into an upgrade I was getting bored so start practicing dropK kicking a small paper ball. Connected a little too well and hit the Big Red B Button with force. Wait for the silence to descend but no. FurtherC investigation leads to discovering that the Big Red Button has been  disconnected....  
 Huw Davies0 COMPAQ SERVICES - Platform Technologies Practice+ 564 St Kilda Rd., Melbourne, Australia 3004    E-mail: Huw.Davies@compaq.com    * Personal Opinion Only *    ------------------------------  % Date: Sun, 22 Apr 2001 22:10:36 -0400 , From: "islandco.com" <dbturner@islandco.com>K Subject: Re: looking for VMS-related hardware, particularly VXTs, in Europe / Message-ID: <te73j61isgtv8f@news.supernews.com>    Philip  1 Have you ever paid for anything ?!?!?!?!!?  (:0)>    David    -- We sell Alpha's ! % Want to buy an Alpha or Alpha Parts ?  Go to http://www.islandco.com & Hardware for Alpha VMS, Tru64 & Linux.   Island Computers US Corporation  2700 Gregory Street  Savannah GA 31404  Tel: 912 447 6622  Fax:912 201 0096 sales@islandco.com4 "Phillip Helbig" <helbig@man.ac.uk> wrote in message' news:9bsecd$171k$1@godfrey.mcc.ac.uk...  | Hi!  | C | As always, I am looking for free or cheap hardware to run VMS on.  | C | I have just realised that a VXT or two would make a nice addition  | to my collection.  | 
 | See also | > |    http://www.astro.multivax.de:8000/helbig/VMS/hardware.txt | J | Before you throw anything away, give me a chance to give it a good home! |  | 	 | Thanks,  | 	 | Phillip  |    ------------------------------    Date: 23 Apr 2001 01:56:59 +0800, From: Paul Repacholi <prep@prep.synonet.com>2 Subject: Re: Management Musical Chairs in the BCSG- Message-ID: <87pue4n7h0.fsf@prep.synonet.com>   ) "Bill Todd" <billtodd@foo.mv.com> writes:   C > "Doc.Cypher" <Use-Author-Address-Header@[127.1]> wrote in message 2 > news:20010422094305.15034.qmail@nym.alias.net... >  > ...   F > > I certainly think that a cut in licensing costs would give sales aB > > real shot in the arm. However, I doubt if the current economicB > > climate would make it seem like a wise move to the top people.  E > That really ought to depend on whether they're currently making anyo? > non-negligible profit on the lowest-end systems: if not, theneC > there's nothing significant to lose, as long as they don't take aB! > noticeable loss on each system.e  = > People have occasionally conjectured that having too low aneE > entry-level price could cause cannibalization of higher-end systems C > by clusters of low-end systems, but that could be limited by fiatVC > (enforced in the code) if necessary - and they've already startedsC > offering clustering sub-set functionality at lower prices (thoughR) > how useful that is may be in question).   D Any one is going to use as low a cost systems as is reasonable. WhatC is harder to pick is when someone nedds, or has a good call for the B 'big box' function, but not the grunt. So he is screwed, and oftenF knows that what he needs is there, just can't use it with out a wad of& $$. How to piss off customers, step 1.  D > The real problem at this point may be hardware pricing: as long asF > the hardware platform costs at least twice as much as a high-qualityC > Intel (or better, AMD) platform of somewhat similar capabilities,oE > it's not clear that the VMS license pricing is all that much out ofrA > line, though reducing it wouldn't hurt (again, assuming there's E > currently not enough volume in that segment to generate much profit?
 > anyway).  A With hard ware, you have real money at stake. You can't stamp outeB DS10Es for 20c each! Hell, even if the box was Tooth Faired to theF door step for nix, there is still a lot of money in stocking, shippingD and installing.  Plus spares inventory. Real money, mind, money that must be paid to someone.  F Licences are essentially free. Some arbitary amount of admin overhead,C and a bit of paper. So it should be of the order of $25-50 plus andgE extra $1-2 per extra licence in one hit. That is COST, real money. SorE if you give some one a $20,000 licence, you are $50 out of pocket. IfmE you give him a $3000 box, that will be, on VMS current margins, $1500S out of pocket.  A (Note that I'm just handwaving the media kits, OK Given a descente2 pressing run, the cost also falls into the noise.)  E So how do you spice things up at the moment? Cut to the bone where it F will not COST you, even if that throws away potential revenue.  If you? do it well, that money WILL come back in a few months, plus yousC perhaps could claim an increase in 'System' sales, so you are on of D the few not slidding into the mud. Being the only vendor with growth2 in a recession is priceless for punter confidence.  E Also, note well that I'm refering to licence costs above, not supporteC costs. It may well be that you need to add a support fee to the daygD one cost. Of course, if you want 24/7, 1hr max fix, that will cost a8 'little' more :) And that is where the money really is!!   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.>@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.   ------------------------------  % Date: Sun, 22 Apr 2001 14:38:37 -0400a' From: "Bill Todd" <billtodd@foo.mv.com>c2 Subject: Re: Management Musical Chairs in the BCSG( Message-ID: <9bv8e3$2eo$1@pyrite.mv.net>  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in message ' news:87pue4n7h0.fsf@prep.synonet.com...    ...>  F > Any one is going to use as low a cost systems as is reasonable. WhatE > is harder to pick is when someone nedds, or has a good call for the D > 'big box' function, but not the grunt. So he is screwed, and oftenH > knows that what he needs is there, just can't use it with out a wad of( > $$. How to piss off customers, step 1.  L This is called supply and demand (or, alternatively, value), something we inI the U.S. often consider to be a Good Thing.  It motivates both innovation F and self-improvement by 'have-nots' to become 'haves' so that they canI purchase such goodies, and at the same time is mitigated by the fact that L anything that is *in actuality* over-priced for its value will be the target2 of competitive products at more reasonable prices.   > F > > The real problem at this point may be hardware pricing: as long asH > > the hardware platform costs at least twice as much as a high-qualityE > > Intel (or better, AMD) platform of somewhat similar capabilities,oG > > it's not clear that the VMS license pricing is all that much out of C > > line, though reducing it wouldn't hurt (again, assuming there'shG > > currently not enough volume in that segment to generate much profitk > > anyway). > C > With hard ware, you have real money at stake. You can't stamp out D > DS10Es for 20c each! Hell, even if the box was Tooth Faired to theH > door step for nix, there is still a lot of money in stocking, shippingF > and installing.  Plus spares inventory. Real money, mind, money that > must be paid to someone. >6H > Licences are essentially free. Some arbitary amount of admin overhead,E > and a bit of paper. So it should be of the order of $25-50 plus andiG > extra $1-2 per extra licence in one hit. That is COST, real money. So>G > if you give some one a $20,000 licence, you are $50 out of pocket. IftG > you give him a $3000 box, that will be, on VMS current margins, $1500u > out of pocket.  I But if you sell, say, 10,000 licenses at cost ($50), but would still have K sold, say, 2000 of them at $20,000 apiece, you're out (really out:  this isa( money you would have had) $39.5 million.  I That's what value is all about, and it's only tenuously related to actualeE manufacturing/handling cost (though it does have something to do withdK overall cost, including development, because that's what allows competitionsI to spring up if you price things too much out of proportion).  And that's K why the platform cost *is* relevant, because as long as the license cost is1F not too unreasonable as a percentage of the total system cost it won't" affect sales volume all that much.   >oC > (Note that I'm just handwaving the media kits, OK Given a descentn4 > pressing run, the cost also falls into the noise.) >iG > So how do you spice things up at the moment? Cut to the bone where itWH > will not COST you, even if that throws away potential revenue.  If you8 > do it well, that money WILL come back in a few months,  I That's easy for you to say:  it's not your business on the line if you'reh wrong.  F If Compaq in fact *is* generating significant revenue with its currentI entry-level systems and pricing, then the portion of that revenue derivedeF from license prices is at risk.  And whether it would make up anywhereG nearly that amount in increased hardware revenue (actually, profit fromeK those increased sales, and perhaps from later service) is a function of thefJ detailed shape of the demand curve, which it's extremely unlikely you haveH any real insight into at that level of detail (even very thorough market+ analysis can get it wrong, and often does).'  H Conversely, if no significant revenue is being generated in this segmentF today, then there's no significant risk to dropping license prices and+ seeing what happens.  Which is what I said.t   - bill   ------------------------------  % Date: Sun, 22 Apr 2001 14:12:53 -0500o+ From: "Main, Kerry" <Kerry.Main@compaq.com> 8 Subject: RE: OpenVMS article - please explain last line!R Message-ID: <DC4745D1A85CA04180C83CDC706A9D180D9551@cthexc02.americas.cpqcorp.net>   Andrew,l  * Sorry Andrew, but it is you who is wrong.   K Read what I stated ie. "they [posted stream results] do not reflect currentt: Alpha capabilities, so it is false to claim that they do."  I These stream results posted are EV67 667 Mhz systems dated from June 2000rI and do NOT reflect current EV68 833Mhz ES40's Alpha capabilities. I don't3L know how much better the numbers for current servers would be, but one couldF reasonably expect the numbers to improve (new and faster cpu, enhanced compiler changes etc). G   How much? Don't know.,  G Anyway, I'm not sure why the stream numbers for current Alpha's are not I posted, but under pure speculation heading - it might be because the 1GhznI Alpha stuff is relatively close ..I have no idea what the real reason is.o  H >>> You will also note from the result that the scalability isn't linear+ indicating that it is the memory subsystem/n  backplane which is the issue.<<<  J With no supporting data, I love the way you can make such expert analysis. :-),  K >>> As is making claims for existing systems which are not supported by thec% data you youselves have published.<<<   G As I stated before - these posted stream numbers do not reflect current E Alpha systems and should not be used in any type of "here is how a P4N1 compares to the "latest" Alpha .." type analysis.c  L One could however rightfully state - "Comparing apples to apples (keeping inF mind the P4 is currently only available as a single cpu) heres how theI latest P4 stream numbers compare to Sun systems and the older EV67 667Mhz  Alpha systems:"0  H Machine ID                      ncpus    COPY    SCALE      ADD    TRIAD  H Generic_Pentium4-1400              1   1437.2   1431.6   1587.7   1575.4H Compaq_AlphaServer_ES40-6/667      1   1331.2   1249.3   1275.4   1338.5H Compaq_AlphaServer_GS320          32  19592.0  19200.1  21022.0  20425.6I Sun_UE_10000_400                   1    364.2    215.4    287.4    296.1 MI Sun_Blade_1750                     1    809.2    814.8    931.4    890.8 6I Sun_UE_10000_400                  64  12141.6   7970.4   8599.0   8600.4    L [Apologies if I did not find the fastest Sun stream numbers - please correct me if I am wrong]   F Note - while the P4 has done well to increase its single cpu benchmarkB performance ratings, one also needs to keep the following in mind:= http://www.inqst.com/articles/p4bandwidth/p4bandwidthmain.htme   Regards,  
 Kerry Main Senior Consultantg Compaq Canada Inc. Professional Servicesp Voice: 613-592-4660e Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----7 From: andrew harrison [mailto:andrew.nospam@uk.sun.com]s Sent: April 20, 2001 12:01 PM  To: Info-VAX@Mvb.Saic.Como8 Subject: Re: OpenVMS article - please explain last line!     "Main, Kerry" wrote: > G > >> My numbers were not false; they were the actual stream results.<<<  > E > Regardless of where you got them, they do not reflect current Alphae5 > capabilities, so it is false to claim that they do.i >    Sorry Kerry you are wrong.  7 They do represent the ES40's current capabilites, your o5 numbers are peak bandwidth which are about as usefull,5 as a chocolate teapot and which are numbers that onlyn8 get published by marketing staff who should know better.  6 His numbers are based on the STREAMS results published by Compaq for the ES40.    They are at   9 http://www.cs.virginia.edu/stream/standard/Bandwidth.htmla   The results for the ES40 are  H Machine ID                      ncpus    COPY    SCALE      ADD    TRIADI Compaq_AlphaServer_ES40-6/667      4   2472.8   2363.3   2314.8   2369.1 aI Compaq_AlphaServer_ES40-6/667      2   1778.8   1712.2   1798.8   1840.1 sH Compaq_AlphaServer_ES40-6/667      1   1331.2   1249.3   1275.4   1338.5  7 You will notice that despite your claim of 5+ GB/s thate4 the ES40 tops out at half your claimed performance.   7 You will also note from the result that the scalability 8 isn't linear indicating that it is the memory subsystem/ backplane which is the issue.g    C > >>> Just as I'm sure people understand the difference between the 	 bandwidth 4 > of a system, and the bandwidth of a processor. <<< > J > Can you help me understand how having a fast CPU waiting for memory dataL > somehow will impress a user who is only interested in overall performance?' > Do they care where the bottleneck is?0 >   = Quite and as your own STREAMS results show the Alpha CPU's in 6 the ES40 are fast CPU's having to wait for the memory 8 subsystem, if they wern't then you would have got better scalability then you did.   I > >>> It is correct that _current_ P4 systems (which also do 3.2 GB/s) do  have? > a lower peak bandwidth than the highest end Alpha systems.<<<o > I > Again, the current low end Alpha DS20 systems use a 5.2Gb/sec cross baroF > switch technology. No one thinks of the DS20 as a "highest end Alpha
 > system". >   6 Again this is just a marketing BS number which is not ! supported by your STREAMS result.A    K > >>> However, the ServerWorks Grand Champion architecture will do 6.4 GB/soJ > with up to 4 P4 processors and has PCI-X. It blows the DS20 and ES40 out of > the water to be sure. :-)<<c > L > Sure, whenever its available, we will be glad to compare apples to apples.C > Right now, comparing future designs to current HW is a bit silly.o >   7 As is making claims for existing systems which are not i3 supported by the data you youselves have published.e   regardsy Andrew Harrison  Enterprise IT Architect    ------------------------------   Date: 22 Apr 2001 19:38:05 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)8 Subject: RE: OpenVMS article - please explain last line!, Message-ID: <9bvbut$ci6@gap.cco.caltech.edu>   In article <DC4745D1A85CA04180C83CDC706A9D180D9551@cthexc02.americas.cpqcorp.net>, "Main, Kerry" <Kerry.Main@compaq.com> writes: >aJ >These stream results posted are EV67 667 Mhz systems dated from June 2000J >and do NOT reflect current EV68 833Mhz ES40's Alpha capabilities. I don'tM >know how much better the numbers for current servers would be, but one couldfG >reasonably expect the numbers to improve (new and faster cpu, enhancedd >compiler changes etc).  >i >How much? Don't know.  L Did the memory system change going from the 667->833?  I don't believe that I part was touched, so there should be no improvement at all, assuming thatlD the compiler is and was generating code that's maximally efficient.   D The STREAM benchmark measures system memory bandwidth (transfers areJ intentionally so large that nothing will fit in cache) and if that is rateG limiting you can quadruple the processor speed and see no change in the  numbers at all.    Regards,   David Mathog mathog@caltech.edu? Manager, sequence analysis facility, biology division, Caltech iJ **************************************************************************J *                                RIP VMS                                 *J **************************************************************************   ------------------------------  # Date: Sun, 22 Apr 2001 20:26:09 GMT.2 From: "Richard B. Gilbert" <Dragon@CompuServe.Com>$ Subject: Re: OT: "The Puzzle Palace". Message-ID: <3AE33EB2.E46DF773@CompuServe.Com>  J     No, no, no, Arne.  Title 18 United States Code Sections 793,  794, andM 798 provide for a penalty of ten years, $10,000 or both.  (Unless, of course, I these sections have been revised since I was last required to read them.)lG These cases are seldom tried; a trial would have to be public and wouldlH probably reveal as much, or more, classified information as the original; violation.  I think they just quietly  have you killed! :-)9   Arne Vajhj wrote:   > Scott Vieth wrote:H > > After a few folks in c.o.v. raved about the book by James Bamford, I > > orderedm > > a copy from Amazon.  > >aK > > It's not quite what I expected.  I was looking more for the techie sidej > > ofE > > what the NSA can do as opposed to the biographical stories of the  > > variousl' > > folks who held the DIRNSA position.: >08 > I think too detailed and too uptodate info about NSA'sA > technical side will reward the author with 50 years in prison !  >  > :-)i >  > Arne   ------------------------------  # Date: Sun, 22 Apr 2001 20:45:13 GMTa2 From: "Richard B. Gilbert" <Dragon@CompuServe.Com>$ Subject: Re: OT: "The Puzzle Palace". Message-ID: <3AE34329.97E186B2@CompuServe.Com>  J These details are not shared for pretty much the same reason that detailedE information on the construction of nuclear weapons is not shared; theiO information could and would be used against us!  I suspsect that the revelationoO of NSA's cryptological techniques might be more dangerous to our interests than 6 the revelation of nuclear weapons technology would be.  N And, assuming that all the cryptologic information was freely  available, whatH would you do with it?  Unless you are trying to secure your military andL diplomatic communications or to read the communications of your enemies, theO techniques are not particularly useful.  Nor are the cryptanalytical techniquesaO anything that the average citizen could employ even if he had them; it requiresyN a great deal of knowledge, a great deal of computing capacity, a great deal ofN time, and a great deal of persistance to break a good encryption system.  BillL Gates has the resources; you and I don't and are never likely to have.  I am$ happy to have the stuff kept secret!   Christopher Smith wrote:   > > -----Original Message-----F > > From: young_r@encompasserve.org [mailto:young_r@encompasserve.org] > I >         Hmmmm.  Much of the technical abilities of the NSA are shroudedeH >         in mystery.  Things of national security and whatnot are stillF >         very crucial to trying to get a leg up on others set against >         our country. >aK > Well said, but somewhat incorrect, IMO.  If it were really "our country,"eN > we'd know what was going on with it, don't you think?  It's "their country,"# > and they're not willing to share.h >t
 > Regards, >h > Chrisc > # > Christopher Smith, Perl Developerr > Amdocs - Champaign, IL >v > /usr/bin/perl -e 'A > print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");a > 'e   ------------------------------  # Date: Mon, 23 Apr 2001 01:44:16 GMTb4 From: "Terry C. Shannon" <terryshannon@mediaone.net>$ Subject: Re: OT: "The Puzzle Palace"= Message-ID: <QNLE6.30356$%_1.8554026@typhoon.ne.mediaone.net>d  = "Richard B. Gilbert" <Dragon@CompuServe.Com> wrote in messagey( news:3AE34329.97E186B2@CompuServe.Com...L > These details are not shared for pretty much the same reason that detailedG > information on the construction of nuclear weapons is not shared; theyF > information could and would be used against us!  I suspsect that the
 revelationL > of NSA's cryptological techniques might be more dangerous to our interests than8 > the revelation of nuclear weapons technology would be. >sK > And, assuming that all the cryptologic information was freely  available,u whatJ > would you do with it?  Unless you are trying to secure your military andJ > diplomatic communications or to read the communications of your enemies, them) > techniques are not particularly useful.c  G As one who has been there and done that (was in the USASA from 70-72) Iu) completely concur with your observations.r   ------------------------------  % Date: Sun, 22 Apr 2001 17:20:15 -0500f/ From: "Stuart, Ed" <Ed.Stuart@austinenergy.com>.# Subject: SAS on VMS vs Windoze 2000aR Message-ID: <92EFB80E551BD511B39500D0B7B0CDCCF3B63A@ohms.electric.ci.austin.tx.us>  I I've been asked to spec a SAS server for our load forecasting group.  Our K production database servers run Oracle on OpenVMS, but the load forecasting0F team is very pro-Windoze.  I'm looking for pointers to help support myJ argument.  Also, from what I can tell the interface to SAS on OpenVMS is X< based so I would have to include an x-term for the clients.   $ Ed Stuart                           ( Manager, Systems and Desktop Services	  + Information Technology & Telecommunications  City of Austin, Austin Energy  Ed.Stuart@austinenergy.com  + "Glittering prizes and endless compromises ./ shatter the illusion of integrity" - Neil Peartg  B *Please apply a generous amount of all the usual disclaimers here*   ------------------------------  % Date: Sun, 22 Apr 2001 21:43:57 +0200c+ From: Leo Klein <leoklein@netvision.net.il>d  Subject: Terminal driver in V7.20 Message-ID: <3AE3347A.60F40DC0@netvision.net.il>  F Having just upgraded to Alpha VMS V7.2 (from V7.1), I've come across a( problem I'd be grateful for comments on.  G I'm on a PC running a terminal emulator (Ericom's Powerterm, though I'moB not sure it matters) connected to VMS. I paste the contents of theF Windows 98 clipboard into the VMS window - in fact, to an application,F though the problem can also be seen at the DCL prompt. Under VMS V7.1,D any quantity of data could be pasted. However, under V7.2, I get theG errors RMS-F-RER (file read error) and -SYSTEM-W-DATAOVERUN after about  160 characters.a  F This isn't a TCPIP V5.1 Telnet problem, because I also see it on a LATD connection. Apparently something has changed in the terminal driver,9 which now probably requires increased quotas - but which?    ------------------------------  % Date: Sun, 22 Apr 2001 14:10:20 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>r$ Subject: Re: Terminal driver in V7.2' Message-ID: <3AE32C9C.46B8A3ED@fsi.net>a   Leo Klein wrote: > H > Having just upgraded to Alpha VMS V7.2 (from V7.1), I've come across a* > problem I'd be grateful for comments on. > I > I'm on a PC running a terminal emulator (Ericom's Powerterm, though I'myD > not sure it matters) connected to VMS. I paste the contents of theH > Windows 98 clipboard into the VMS window - in fact, to an application,H > though the problem can also be seen at the DCL prompt. Under VMS V7.1,F > any quantity of data could be pasted. However, under V7.2, I get theI > errors RMS-F-RER (file read error) and -SYSTEM-W-DATAOVERUN after about0 > 160 characters.e > H > This isn't a TCPIP V5.1 Telnet problem, because I also see it on a LATF > connection. Apparently something has changed in the terminal driver,; > which now probably requires increased quotas - but which?e  # Neither. Do this in your LOGIN.COM:   & $ IF F$MODE .EQS. "INTERACTIVE" THEN - $ SET TERMINAL/HOSTSYNCa  C That "guarantees" that the terminal driver will send XOFF when youroH type-ahead buffer is almost full and XON when it has been cleared down a bit.  1 Note, however, the quotes in the foregoing... :-)f  H Alternatively, you could increase the system parameters TTY_TYPAHDSZ and/ TTY_ALTYPAHD, but HOSTSYNC maybe more reliable.e   -- e David J. Dachteraa dba DJE Systemsw http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.,   ------------------------------  % Date: Sun, 22 Apr 2001 22:09:47 +0100r; From: Malcolm MacArthur <malcolmm@rustic-place.demon.co.uk>e5 Subject: Re: Using 3-phase power converters on VAXen?,8 Message-ID: <3AE3489B.3CAF2291@rustic-place.demon.co.uk>   Paul Sture wrote:d [...]r > K > The UK does have a ground to earth. Sometimes just a rod into the ground,uJ > but I've seen it tied to water pipes, and gas pipes as well. Whether theI > latter is just a good ground or a way of ensuring that gas and electrice) > charge don't mix I honestly don't know.   F Sometimes, ground can be tied to neutral instead. Which leads me to an amusingoF story about a former flatmate. He had just moved in and plugged in his	 TV - onlyg# to find out it didn't work at all. a  H After much headscratching, he asked me, and after further headscratching I saitG we should open up the plug and try replacing the fuse, as it might have: blown orH broken in transit. So I duly opened up the plug to find that the neutral wire: was connected to earth! Rewired the plug and TV worked OK.  C It could only have worked if his house had ground tied to neutral -  correct? TheG flat we stayed in was ca. 100 years old (when the telephone cable brokeu and wasoG replaced, the engineer said it was over 60 years old!) - back then they  used (as they do now again) earth rods.  C Oh, and the amusing bit: he was studying electrical engineering :-)6   -Malcolm   >  > ___a > Paul Sture
 > Switzerlandm   ------------------------------  # Date: Mon, 23 Apr 2001 04:58:25 GMTn From: Dirk Munk <munk@home.nl>5 Subject: Re: Using 3-phase power converters on VAXen?.' Message-ID: <3AE3B671.10710C94@home.nl>'   Malcolm MacArthur wrote:   > Paul Sture wrote:' > [...]  > >tM > > The UK does have a ground to earth. Sometimes just a rod into the ground,tL > > but I've seen it tied to water pipes, and gas pipes as well. Whether theK > > latter is just a good ground or a way of ensuring that gas and electrict+ > > charge don't mix I honestly don't know.h > H > Sometimes, ground can be tied to neutral instead. Which leads me to an	 > amusingaH > story about a former flatmate. He had just moved in and plugged in his > TV - onlyb$ > to find out it didn't work at all. >.J > After much headscratching, he asked me, and after further headscratching > I saihI > we should open up the plug and try replacing the fuse, as it might haves
 > blown orJ > broken in transit. So I duly opened up the plug to find that the neutral > wire< > was connected to earth! Rewired the plug and TV worked OK. >lE > It could only have worked if his house had ground tied to neutral -u
 > correct?  K No, it means his house had a bad ground connection. As pointed out earlier,cN neutral is connected to ground, but quite often only at the source = the powerK companies big transformer. So it should be possible to run equipment with a3O connection between hot and ground. Maybe you know those small screwdrivers with2N a small neon light in them. You can use them to check if a wire is 'hot'. TheyH (should) only light up when you touch the hot wire, never when you touchK neutral, because there is suppose to be no (big) voltage difference between  neutral and ground.        > The I > flat we stayed in was ca. 100 years old (when the telephone cable brokew	 > and wasoI > replaced, the engineer said it was over 60 years old!) - back then theyr
 > used (as  > they do now again) earth rods. >nE > Oh, and the amusing bit: he was studying electrical engineering :-)t > 
 > -Malcolm >e > >a > > ___r > > Paul Sture > > Switzerland    ------------------------------  + Date: Sun, 22 Apr 2001 14:06:16 -0600 (MDT)M" From: John Nebel <nebel@csdco.com>" Subject: Re: WIS... RIP???  -- NO!G Message-ID: <Pine.OSF.4.21.0104221403300.17662-100000@athena.csdco.com>o   See:  " http://www.support.compaq.com/wis/    - On Sun, 22 Apr 2001, David J. Dachtera wrote:r  ( > "Brian Schenkenberger, VAXman-" wrote:
 > > [snip]. > > >relay1.support.compaq.com [192.208.35.18] > > 0 > > ... but thanks for the IP as that does work. > > D > http://192.208.35.18/ not responding as of 11:16 (a.m.) on Sunday, > 22-Apr-2001. >  > -- r > David J. Dachterae > dba DJE Systemsr > http://www.djesys.com/ > < > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/e > H > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.c >    ------------------------------  % Date: Sun, 22 Apr 2001 20:24:21 -050021 From: "David J. Dachtera" <djesys.nospam@fsi.net>v" Subject: Re: WIS... RIP???  -- NO!' Message-ID: <3AE38445.56B70999@fsi.net>r   John Nebel wrote:1 >  > See: > $ > http://www.support.compaq.com/wis/  F ...but still redirects you to a non-standard TCP port. Corporate usersE will likely find it blocked by their company's firewall, unless theirbA workstations are so unfortunate as to be not behind the firewall.a   --   David J. Dachterag dba DJE Systemsa http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/<  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.p   ------------------------------    Date: 22 Apr 2001 22:41:52 -05007 From: hamilton@encompasserve.org (Bradford J. Hamilton) 6 Subject: Re: WIS... RIP???  -- NO! - non-standard port3 Message-ID: <eyZjidUdOspC@eisner.encompasserve.org>e  	 Hi David,e  N Actually, my first contact with this site was from behind a corporate firewall* (we were tired of dialing out to DSNLink).  V Can you educate me a little regarding the use of a non-standard TCP/IP port?  Is thereJ any risk to using such a port, either to my workstation, or to my company?   Thanks,M Brad  \ >In article <3AE38445.56B70999@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes: > John Nebel wrote:a >> s >> See:t >> s% >> http://www.support.compaq.com/wis/  > H > ...but still redirects you to a non-standard TCP port. Corporate usersG > will likely find it blocked by their company's firewall, unless theirHC > workstations are so unfortunate as to be not behind the firewall.i >  > -- i > David J. Dachterat > dba DJE Systemse > http://www.djesys.com/ > < > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/3 > H > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.n   ------------------------------   End of INFO-VAX 2001.226 ************************