1 INFO-VAX	Sun, 24 Jul 2005	Volume 2005 : Issue 410       Contents:N Re: Building SIMH V3.4 on OpenVMS Alpha V7.3-2 / CC V6.5 - errors - any ideas? Re: Exabyte 220 robot blues.3 Re: perfomance question about bool variables in cxx 3 Re: perfomance question about bool variables in cxx 3 Re: perfomance question about bool variables in cxx ) Re: Platform Support vs. Business Support ) Re: Platform Support vs. Business Support ) Re: Platform Support vs. Business Support   F ----------------------------------------------------------------------    Date: 24 Jul 2005 02:30:44 -0700' From: "Kevin Murrell" <kevin@ps8.co.uk> W Subject: Re: Building SIMH V3.4 on OpenVMS Alpha V7.3-2 / CC V6.5 - errors - any ideas? C Message-ID: <1122197444.245412.253620@g43g2000cwa.googlegroups.com>    Hi  F Quick change in snprintf.c in [pcap-vci] to uncomment the {hash}define? HAVE_SNPRINTF will allow overcome the multiply defined symbols.   E Regarding the list of 4 undefined symbols, there is a line missing in 1 descip.mms - the appropriate section should read:    VAX_DIR = SYS$DISK:[.VAX] # VAX_LIB = $(LIB_DIR)VAX-$(ARCH).OLB 7 VAX_SOURCE = $(VAX_DIR)VAX_CPU1.C,$(VAX_DIR)VAX_CPU.C,\ 5              $(VAX_DIR)VAX_FPA.C,$(VAX_DIR)VAX_IO.C,\   @ $(VAX_DIR)VAX_CIS.C,$(VAX_DIR)VAX_OCTA.C,$(VAX_DIR)VAX_CMODE.C,\9              $(VAX_DIR)VAX_MMU.C,$(VAX_DIR)VAX_STDDEV.C,\ 9              $(VAX_DIR)VAX_SYS.C,$(VAX_DIR)VAX_SYSDEV.C,\ >                $(VAX_DIR)VAX_SYSCM.C,$(VAX_DIR)VAX_SYSLIST.C,\<              $(PDP11_DIR)PDP11_RL.C,$(PDP11_DIR)PDP11_RQ.C,\<              $(PDP11_DIR)PDP11_TS.C,$(PDP11_DIR)PDP11_DZ.C,\<              $(PDP11_DIR)PDP11_LP.C,$(PDP11_DIR)PDP11_TQ.C,\%              $(PDP11_DIR)PDP11_XQ.C,\ :              $(PDP11_DIR)PDP11_RY.C,$(PDP11_DIR)PDP11_VH.C
 VAX_OPTIONS = : /INCLUDE=($(SIMH_DIR),$(VAX_DIR),$(PDP11_DIR)$(PCAP_INC))\;                 /DEFINE=($(CC_DEFS),"VM_VAX=1"$(PCAP_DEFS))     D Agree the point about the missing continuation char, but can't think8 where is was before I fixed it.  Fairly obvious however.  
 Kevin Murrell % for Computer Conservation Society, UK    ------------------------------  % Date: Sun, 24 Jul 2005 06:35:18 +0800  From: prep@prep.synonet.com % Subject: Re: Exabyte 220 robot blues. - Message-ID: <87irz16ti1.fsf@prep.synonet.com>   4 David J Dachtera <djesys.nospam@comcast.net> writes:   > prep@prep.synonet.com wrote: >>  A >> I have a feeling this is `user is an idiot', and I've forgoten $ >> something simple in the set up... >>  B >> I have just got a 220 robot with a Mammoth 8900 drive for home,? >> and hooked it to the VLC running 7.2. That's because the VLC E >> does not weigh 90lbs! Drive works fine, backup seems to be a happy % >> camper, but MRU, aka ROBOT is not.  >>  > >> Loaded the GK driver, defined the MRU_ROBOT logical, but... >>  A >> ROBOT GKA0 is not responding: Operating system specific error. 5 >> %ROBOT-E-OS_ERROR, Operating system specific error  > H > The robot in a TZ887, for example, is at LUN 1. If the drive's SCSI Id? > is 5, for example, as MKA500:, the robot can be found at 501:  > 0 > $ MC SYSMAN IO LOAD GKA501/DRIVER=SYS$GKDRIVER > F > Dunno if that gives any insight into what the Mammoth may offer, but > hope it helps, anyway...  E It was partial UII, after getting the `we have changed the uni number D message, I was silly enough to believe it... Yes, it had to be power cycled!    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Sun, 24 Jul 2005 14:28:45 +0200 ' From: "Klaus-D. Bohn" <info@it-bcsb.de> < Subject: Re: perfomance question about bool variables in cxx7 Message-ID: <42e37bde$0$24881$9b622d9e@news.freenet.de>   4 "Tom Linden" <tom@kednos.com> schrieb im Newsbeitrag" news:opsudfnzjezgicya@hyrrokkin...L > On Sat, 23 Jul 2005 12:57:10 +0200, Klaus-D. Bohn <info@it-bcsb.de> wrote: > 9 > > "Doc." <doc@openvms-rocks.com> schrieb im Newsbeitrag 5 > > news:Xns969BCE8448479dcovmsrox@212.100.160.126... + > >> %NEWS-I-NEWMSG, Klaus-D. Bohn wrote in 3 > >> news:42e118d0$0$15037$9b622d9e@news.freenet.de  > >>< > >> > "Doc." <doc@openvms-rocks.com> schrieb im Newsbeitrag8 > >> > news:Xns969BA8A931D17dcovmsrox@212.100.160.126... > >>K > >> >> Slight problem there - the compiler is so smart it optimises the 64  > >> > >> > Hello Doc,  > >> >E > >> > Yes, that's ok. The compiler is very intelligent. So, you must  compile E > >> > with the option "/nooptimize" to force the short program code.  > >>G > >> The issue here - at least for the denizens of this newsgroup - is,  "what I > >> problem are you trying to solve?".  If folks know that then they can 	 > >> make A > >> a contribution with more confidence of being helpful to you.  > >>H > >> You may or may not prove that in certain circumstances working withH > >> particular datatypes is more efficient, but what it will eventually comeK > >> down to is using the compiler optimizations and placing some degree of  > >> trust in them.  > >>E > >> I just pointed out that your code got optimised out of existence  because F > >> I found it mildly amusing.  What other folks in the newsgroup are > >> curiousJ > >> about is - as I said - "what problem are you trying to solve?".  SomeK > >> background and an indication of what your goals are might allow people  > >> to I > >> offer coding advice and other help.  If it is purely down to testing L > >> performance of specific operations, then I wouldn't be surprised if youH > >> can find people to discuss the A64 assembly code in detail.  WhilstK > >> there's lots of OT discussion here, it is certainly no indication that G > >> there are not very technically competent people lurking.  What you  > >> appear J > >> to be curious about is a topic that quite a few would welcome as very > >> on- > >> topic.  > >>L > >> Incidentally, considering I got really slow results on Deathrow's DS10L4 > >> for your test_64.c, what are you running it on? > >> > >>	 > >> Doc.  > >> -- L > >> OpenVMS:     Eight out of ten hackers prefer *other* operating systems.L > >> http://www.openvms-rocks.com    Deathrow Public-Access OpenVMS Cluster. > >  > > Hello Doc, > > G > > at first, I'm porting a very interesting application to the OpenVMS 9 > > plattform. It will be available in October this year. B > > The application is running. But im looking for to increase the > > performance.H > > So i tested with a short program code what is faster: 32bit or 64bit% > > variables for the Alpha platform. D > > Next step was to look what ist faster: bool-variables or __int64
 > > variables  > > which are doing the same. H > > I know my short program is not the same as the full application. But
 > > when iK > > see the 64bit are very much faster as 32bit the application must be run # > > faster. I hope that is correct.  > > I > > My develop system is a DS10L (617Mhz) with 1 GB memory and im working 	 > > alone 
 > > on it. > > 	 > > Klaus  > > I > Well, if it were me, I would step through the machine instructions with  the F > debugger.  After all, you only need to go through the loop one time. > 
 Hello Tom,  G :-). Yes, i know. If i have the counter only by 1, i can not see a time A difference. That's only the reason why i have the for  statement.    Klaus    ------------------------------  % Date: Sun, 24 Jul 2005 14:51:25 +0200 ' From: "Klaus-D. Bohn" <info@it-bcsb.de> < Subject: Re: perfomance question about bool variables in cxx7 Message-ID: <42e3812f$0$24881$9b622d9e@news.freenet.de>   : "Bob Gezelter" <gezelter@rlgsc.com> schrieb im Newsbeitrag< news:1122058234.888333.17740@f14g2000cwb.googlegroups.com... > Larry, > H > With all due respect, I disagree with your explanation.  If Klaus wereG > dealing with a large array of booleans, I would agree with you to the I > extent that stepping through 8 times as much data (quadwords instead of ' > bytes) would generate inefficiencies.  > G > In this case, the posted code generates a large number (42,949,672 to C > be precise) of references to a single variable (quadword or byte, F > respectively). In the case of the Alpha, Itanium, and almost all VAXG > models, presuming that the data is naturally aligned by the compiler, > > each reference should be a single cache or memory reference. > G > However, this simple experiment also illustrates the difficulties and  > challenges of benchmarking.  > @ > - The compiler will, in many cases, optimize out the invariantG > assignment code in the loop. Thus, with the correct compiler switches I > used, the nearly 43 million iterations of the loop will be optimized to . > a single instruction, executed exactly once. > I > - The fine level timing of this code, if it executes even a significant F > number of iterations will be influenced by the other activity on theF > CPU. Any event which affects cache hit rate, will effect the overallG > performance of the code. To do this experiment effectively requires a I > deeper analysis, and many test runs to get a significant certainty that C > the numbers are not being influenced by random events. Science is % > reproduceable, happenstance is not.  > G > - There is a certain amount of processing that is not attributable to H > the actual code under test. This includes image activation, the jacketI > code in the test program, and image rundown.  How much this is needs to I > be computed, and subtracted from the test results to obtain an accurate D > result. This overhead is the computational equivalent of the "tareC > weight", the weight of an empty container when placed on a scale.  > C > My apologies for running on a bit. I hope that this commentary is 
 > helpful. > & > - Bob Gezelter, http://www.rlgsc.com >   
 Hello Bob,   thank you very much.K I have a very interesting new application for the OpenVMS platform. It will " be available in october this year.B The application is ready, but I'm looking how i could increase theI perfomance. So i make some short test programs to look what is the really , difference between the variable definitions.J Normaly i could set the counter to 1. But the system is to fast to analyze the time difference.K I get some very interesting results. Now i chance my application to a clean  64bit application.K Next step is to analyze the application with some performance tools. I hope  i have a positive result.    Klaus    ------------------------------  % Date: Sun, 24 Jul 2005 06:21:13 -0700 # From: "Tom Linden" <tom@kednos.com> < Subject: Re: perfomance question about bool variables in cxx( Message-ID: <opsue51njezgicya@hyrrokkin>  : On 23 Jul 2005 18:06:59 +0100, Peter 'EPLAN' LANGSTOEGER   <peter@langstoeger.at> wrote:   J > In article <opsudfnzjezgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com>  	 > writes: L >> Well, if it were me, I would step through the machine instructions with   >> theG >> debugger.  After all, you only need to go through the loop one time.  > % > OTOH, this is why DEC made the PCA.  >   / I guess that would also include the shared libs    ------------------------------    Date: 24 Jul 2005 08:58:39 +01006 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)2 Subject: Re: Platform Support vs. Business Support, Message-ID: <42e3583f$1@news.langstoeger.at>  \ In article <42E2DADB.C1848025@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:! >Peter 'EPLAN' LANGSTOEGER wrote: G >> The right answer to this bullshit is "Since when is Windows an OS ?"  > O >Olsen made a grave mistake when he underestimated Sun and then refused to even - >consider wintel to be competing against VMS.   ' Talking and doing are different things. J He must have talked about SUN and WINTEL being no competitives while doingK all that they are not becoming one. As we know the latter part was missing.   I >It is wrong to have the attitude "Since when is Windows an OS". You must O >respect your competitor, especially when you competitor has 90% of the market.   E No. The answer to bullshit is another bullshit. When they admin it is H bullshit, you can also step back and admit and 'respect the competitor'.M It was 'not fighting against the bullshit' which made it '90% of the market'.    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Sun, 24 Jul 2005 11:05:38 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: Platform Support vs. Business Support0 Message-ID: <11e7bbnqr01309f@corp.supernews.com>   JF Mezei wrote:  > Dave Froble wrote: > H >>The engines are manufactured by a company in West Virginia.  They haveJ >>multiple uses, models, target drones, and such.  Each produces 45 lbs of	 >>thrust.  >  > 8 > Are they pure jet, or do they have a bypass section ?  >  >  > H >>Neat things.  About 12 inches in length, 6-7 inches in diameter.  Very? >>light.  Expensive, $5,000 per engine, with a TBO of 50 hours.  >  >  > What does "TBO" mean ?   Time between overhaul.  F >>My friend Jack McCornack produced the thingies that were in the last >>James Bond movie.  >  >  > @ >>Stupid, still have to deploy a parachute to land.  Not for me. >  > P > But with those engines on it, it might be useful to transfer a person from one > cargo plane to another.   F Well I guess there is a stuntman somewhere stupid enough to try that. H Some things aren't quite as simple as the movies make them seem.  there I have been occurances when a light aircraft on takeoff or landing got too  C close behind a large aircraft, and the wingtip vorticies flip them  F inverted.  At low speed is when they are the worst, but still, flying ? right up behind a big heavy cargo aircraft wouldn't be trivial.    --  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: Sun, 24 Jul 2005 11:09:55 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: Platform Support vs. Business Support0 Message-ID: <11e7bjni48li34b@corp.supernews.com>   Keith Cayemberg wrote: > JF Mezei wrote:  >  >> Dave Froble wrote:  >>J >>> it.  There's 2 jet engines on my back porch, and I'm putting them on aF >>> 50 MPH Ultralight aircraft this afternoon.  Not only have I joinedG >>> 'them', but now I'm leading them.  But at least I'll have some fun.  >> >> >> >>D >> Since your ultralight isn't very aerodynamic, make sure your jet  >> engines areC >> overpowered to compensate. I hope you got at least 2 GE90s  jet   >> engines forC >> your ultralight :-) :-)  Oh, and be careful when you start your  
 >> engines to 6 >> make sure the exhaust isn't aimed at your house ... >> >>E >> (For those who don't know, the GE90 is the engine used in the 777   >> aircraft,D >> with the nacelles almost as wide as the fuselage of a 737 :-) It  >> generatesJ >> over 90,000 pounds of thrust in its early vintage and has been uprated  >> since  >> (I thing it now does 115,000) >> > A > Na, a transonic turbofan developed using OpenVMS would be more  K > appropriate. How about the Pratt & Whitney F119-PW-100 used in the F/22A  ! > at 35,000 pounds of thrust? :-) % > http://www.adaic.org/atwork/pw.html 0 > http://www.pratt-whitney.com/prod_mil_f119.asp  H I can't afford enough fuel to just start the thing.  I'd like it.  If I + was given one, I couldn't afford to run it.    --  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 2005.410 ************************                                                                                                                                                                                              