1 INFO-VAX	Sat, 27 Aug 2005	Volume 2005 : Issue 478       Contents: Re: Adding users RE: Adding users! Re: CHARON-VAX freeware available  Re: Copying jar on OpenVMS" DIAGNOSE INSTANCE CODE TRANSLATION7 Re: msa1000 - Connect to VMS host for arrary management 3 Re: Multiple email domains with TCPIP services SMTP 3 Re: Multiple email domains with TCPIP services SMTP 3 Re: Multiple email domains with TCPIP services SMTP $ Next project, C programming problem.) Re: Optimum cluster size for EVA storage? ) Re: Optimum cluster size for EVA storage?  Re: Q3CY205 arrived... Re: Q3CY205 arrived...! Re: Simh. How to triple the speed ! Re: Simh. How to triple the speed ! Re: Simh. How to triple the speed   F ----------------------------------------------------------------------  # Date: Sat, 27 Aug 2005 08:40:46 GMT   From: John Santos <john@egh.com> Subject: Re: Adding users ) Message-ID: <iGVPe.1195$um2.802@trnddc03>    JF Mezei wrote:  > Joseph Huber wrote:  > 7 >>I suggest to create a first account in group manually ? >>(e.g. for the group administrator), so the identifiers exist, C >>and the bulk procedure can generate all other UICs automatically.  >  > A > One thing to remember. By default, the UAF> command "ADD" takes  > "DEFAULT" as template. > G > (one should really be able to specify a template for user creation in  > the ADD command).   F You can!  (Well, it's a different command, but I don't think there are any practical differences...)   A Just set up a dummy username as your desired template, flag it as - DisUser, and copy it when creating new users:   H UAF> COPY MY_DEFAULT NEWUSER/UIC=[x,y]/pass=temp_password/flag=nodisuser  G (Put an explicit /UIC when ever you copy or add a user to avoid mucking = up the DEFAULT or source username's Identifier when you later  MODIFY/UIC NEWUSER.)  C As far as I can tell, UAF> ADD NEWUSER does the exact same thing as F UAF> COPY DEFAULT NEWUSER, except there is an implicit clearing of the$ DISUSER flag in the new user record.  I > ALL-IN-1 used its own account templates stored elsewhere and supplied a H > whome bunch of data for each account creation, bypassing the "DEFAULT"H > mechanism. It made it easy to have users from different profiles given8 > different UAF parameters (as well as ALL-IN-1 params).  B Why store them elsewhere when they can be stored right in the UAF, using existing mechanisms?  @ (Okay, the system mangler could mangle them using AUTHORIZE, but hey, that's his job!)   H > One way to do this in DCL would be to designate account template UICs.H > For instance for groupx X, the uic [X,1] would be an account template. > I > Then, your procedure to add accounts would use the COPY command copying B > from that group's template to the new user, and specify only the. > relevant fields (as well as /FLAG=NODISUSER)  B Rats!  I should have read the rest of your post before writing the above.  :-)    --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Sat, 27 Aug 2005 09:49:50 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com>  Subject: RE: Adding users R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB6B2248@tayexc19.americas.cpqcorp.net>   > -----Original Message-----9 > From: JF Mezei [mailto:jfmezei.spamnot@teksavvy.com]=20  > Sent: August 26, 2005 9:51 PM  > To: Info-VAX@Mvb.Saic.Com  > Subject: Re: Adding users  >=20 > "Main, Kerry" wrote:@ > > As much as some folks hate using gui's, OpenVMS users can=20 > also be added ; > > via point-click Windows based utility. Allows you to=20  > point-click-sort@ > > various user fields. Storage and print queues can also be=20 > managed with > > this Windows based utility.  >=20 >=20> > The OP needs to do mass additions. GUIs are ill suited to=20 > automate such tasks. >=20  C Bill also mentioned he wanted to maintain this list and while a DCL A routine can be used to do the initial load, I suspect the ongoing E maintenance of student ID's is a lot of work i.e. changes, additions,  deletions etc.=20   E Hence, the 10-15 minutes required to setup a Windows based utility to D manage the OpenVMS resources *might* be worth it - especially if youG have junior Operations and/or support staff that is used to Windows GUI  mgmt utilities.    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: Sat, 27 Aug 2005 12:40:53 -0500 ( From: Wayne Sewell <wayne@tachysoft.com>* Subject: Re: CHARON-VAX freeware available/ Message-ID: <00A48E8D.069BDDD8.1@tachysoft.com>   , >From: Robert Boers <r_boers@softresint.com>9 >User-Agent: Mozilla Thunderbird 1.0.6 (Windows/20050716)  >X-Accept-Language: en-us, en  >MIME-Version: 1.0 >X-Newsgroups: comp.os.vms+ >Subject: Re: CHARON-VAX freeware available    >Robert Boers wrote:0 >> We posted downloads of CHARON-VAX freeware on: >> http://www.softresint.com/charon-vax/Tools_and_tips.htm >>  H >> These downloads are available for OpenVMS/Alpha, OpenVMS/Itanium and G >> Linux. They provide an emulated VAX 3600 with 2 disk drives, CDROM,  B >> floppy disk and Ethernet support. There are no time or runtime G >> limitations; these versions are for non-commercial use only and are   >> unsupported.  >>   >> Robert Boers. > 9 >The versions for OpenVMS are VAX model 3100/98 actually.   O Unfortunately, I found the kit unusable.  The files vaxsys.vdisk and user.vdisk O are provided as self-extracting executables for a *billybox*.  I decommissioned O my last physical billybox with great ceremony a couple of years ago.  I finally N managed to create the files on the billybox emulation in the mac but they wereO created in some state that prevents me from getting them out of the emulation.  L I get "no access" when I try to copy them to the common directory visible to5 the outer mac, and ftp opens the file and then hangs.   O For a kit destined for an alpha or itanium, wouldn't it make more sense for the L self-extracting zip to be for alpha or itanium?  Why should we have to screwG around with a billybox?  The rest of the kit is just a simple zip file.   J The fact that these files are required and I can't get them to the itanium makes the whole kit useless.   Wayne O =============================================================================== N Wayne Sewell, Tachyon Software Consulting  (281)812-0738   wayne@tachysoft.com; http://www.tachysoft.com/www/tachyon.html and wayne.html    O =============================================================================== P Jake Blues:"You traded the Caddy for a microphone? ...... Okay, I can buy that."   ------------------------------    Date: 27 Aug 2005 02:24:02 -07002 From: "Martin Stauber" <stauber@biomed.ee.ethz.ch># Subject: Re: Copying jar on OpenVMS B Message-ID: <1125134642.386201.95600@g43g2000cwa.googlegroups.com>   Arne Vajh=F8j schrieb:   > 2 > But you should get the image from within the JAR > file as a ressource. > . > And if you post your code I can explain how. >  > Arne   Thank's for this hint Arne, : That's what I actually wanted. I replaced now my code from  $ new ImageIcon("images/fileopen.png")   to  C new ImageIcon(ClassLoader.getSystemResource("images/fileopen.png"))    which seems to do the job. Martin   ------------------------------  % Date: Sat, 27 Aug 2005 09:54:47 -0500 + From: brandon@dalsemi.com (BRANDON, JOHN M) + Subject: DIAGNOSE INSTANCE CODE TRANSLATION 1 Message-ID: <05082709544726@dscis6-0.dalsemi.com>   L Are the instance codes posted by DIAGNOSE available anywhere?  Looking for a" table of IC and their description.       For example:N Instance Code             x43010064  Host Port Protocol component has detectedM                                      that the other controller has failed and L                                      that this controller has taken over theJ                                      units specified in the extended sense*                                      data.           John "REBOOT" Brandon  VMS Systems Administrator * firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  # Date: Sat, 27 Aug 2005 08:19:03 GMT   From: John Santos <john@egh.com>@ Subject: Re: msa1000 - Connect to VMS host for arrary management* Message-ID: <XlVPe.1977$Uz2.1010@trnddc02>   ksrich@bellsouth.net wrote: , > <flamingomn@hotmail.com> wrote in message ? > news:1125004454.018617.116060@g47g2000cwa.googlegroups.com...  >  >>OVMS 7-3.2 >>D >>We have a new ALPHA ES45 that has an MSA1000 in it. We're tryng toB >>connect to it from a VMS host via the seriel port.  There are no, >>errors, it just hangs and doesn't connect. >>I >>Unfortunatly, HP has not been a good resource for this problem.  We get B >>the feeling that we're the only people trying to do this in this	 >>manner.  >>E >>WE need to do it this way because the support will be in MN and the B >>machine will be in TX.  And I have no desire to move to TX !  :) >>8 >>Has anyone out there gotten this to work successfully? >>  F If got one (actually 4, but only one cable) connected to a serial portE on an ES40.  Works fine.  Have also experimented with connecting them B to a DS90.  Needed to construct an RJ45-DB9 adapter with the rightC pinning to make that work, but it does that as well.  (Advantage of E the DS90 is it has enough free ports on it to connect all 4 MSA1000's E at the same time, so we can manage it remotely.  Still waiting on the  other 3 cables, though.)   >  > M > I have it connected to a PC with no problems.  I use PC/Anywhere to get to   > the PC to do remote  > managment. > N > Sounds like it maybe a cable issue.  Also the MSA defaults to 19200 instead 
 > of 9600.  H You have to use the special RJ???-DB9 cable that comes with the MSA1000.G The RJ end looks like an RJ45, but has 10 pins instead of the normal 8. D The DB9 end of the cable is plugged directly into the serial port onA the Alpha.  (It has PC-compatible pin arrangement, unlike the DB9 C console port on some older MicroVAXes.)  Maybe you got a bad cable?   4 And Shael is absolutely right about the 19200 speed.   What I do is this:   In SYSTARTUP_VMS.COM,   )   $ set term/perm/speed=19200/notypeahead   = This prevents the line from chattering, and trying to log in.   $ When I need to access the MSA1000, I     $ allocate tta0:   $ set term/typeahead tta0:G (No /perm, so it reverts to /notypeahead when I deallocate or log off.)    $ set host/dte tta0:
 (Do my stuff)    ctrl-\ to exit   $ deallocate tta0:   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  + Date: Sat, 27 Aug 2005 07:11:53 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)< Subject: Re: Multiple email domains with TCPIP services SMTP$ Message-ID: <dep3np$bs4$1@online.de>  5 In article <430FC5B6.AE2550B8@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:    G > With a single node, because the TCPIP Services rely on VMSmail, there ; > isn't a way to differentiate between an email adressed to D > support@chocolate.com and support@vanilla.org, as both would go to% > username "support" on the one node.   D Assuming that support is a username on the VMS system.  However, one	 could do    A    MAIL> SET FORWARD/USER=support@chocolate.com chocolate_support =    MAIL> SET FORWARD/USER=support@vanilla.org vanilla_support   K The "SET FORWARD/USER=foo bla" syntax is quite flexible.  For example, foo  A does not have to be a valid VMS username---it can be essentially  I anything.  Also, bar can be a local VMS username, or an email address on   a remote system.   Perhaps you could do    1    MAIL> SET FORWARD/USER=user@chocolate.com null   C For each user.  Then, email to user@vanilla.org would be processed  H normally, and mail to user@chocolate.com would go to null, which should = be a VMS username whose mailbox fills up with unwanted stuff.   ! Not pretty, but perhaps workable.    ------------------------------  % Date: Sat, 27 Aug 2005 03:22:00 -0400 , From: JF Mezei <jfmezei.spamnot@tekavvy.com>< Subject: Re: Multiple email domains with TCPIP services SMTP+ Message-ID: <43101483.97756863@tekavvy.com>   / Phillip Helbig---remove CLOTHES to reply wrote: F > Assuming that support is a username on the VMS system.  However, one
 > could do > C >    MAIL> SET FORWARD/USER=support@chocolate.com chocolate_support ? >    MAIL> SET FORWARD/USER=support@vanilla.org vanilla_support    Are you sure this works ?   A Doesn't the SMTP symbiont just use the callable mail interface to G deliver to a VMS username ? Wouldn't VMSmail see only "support" and not 9 the domain name when it is asked to deliver the message ?   G (Remember that callable mail sets multiple individual enveloppe TO:s in F deliverable format, and then just one line of text for the content TO:G (displayed when you open the message). The enveloppe TOs wouldn't be in 7 smtp format since they would loop back to the symbiont.    ------------------------------  + Date: Sat, 27 Aug 2005 11:42:39 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)< Subject: Re: Multiple email domains with TCPIP services SMTP$ Message-ID: <depjjf$69d$1@online.de>  4 In article <43101483.97756863@tekavvy.com>, JF Mezei& <jfmezei.spamnot@tekavvy.com> writes:   1 > Phillip Helbig---remove CLOTHES to reply wrote: H > > Assuming that support is a username on the VMS system.  However, one > > could do > > E > >    MAIL> SET FORWARD/USER=support@chocolate.com chocolate_support A > >    MAIL> SET FORWARD/USER=support@vanilla.org vanilla_support  >  > Are you sure this works ?  > C > Doesn't the SMTP symbiont just use the callable mail interface to I > deliver to a VMS username ? Wouldn't VMSmail see only "support" and not ; > the domain name when it is asked to deliver the message ?  > I > (Remember that callable mail sets multiple individual enveloppe TO:s in H > deliverable format, and then just one line of text for the content TO:I > (displayed when you open the message). The enveloppe TOs wouldn't be in 9 > smtp format since they would loop back to the symbiont.   I You're right; it doesn't work!  The emails don't seem to be delivered to   anywhere, as a matter of fact!  G Nevertheless, with SET FORWARD/USER=FOO BAR, FOO does not have to be a  B VMS username and does not even have to be a potentially valid VMS E username.  Also, BAR can definitely be an address on a remote system.   F Originally, of course, FOO and BAR were both intended to be local VMS F usernames.  BAR can be a remote address; I was thinking FOO could too.   ------------------------------   Date: 21 Aug 2005 15:09:10 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)- Subject: Next project, C programming problem. , Message-ID: <3mrjomF1888arU1@individual.net>  B OK, now that I am ready to turn the students  loose on the VAX forB programming projects I have another problem to solve.  While SAMBAE now makes it trivial to move files between the VAX and out fileserver B the files will not be in a proper VMS text format and the compilerA won't deal with them.  No big thing.  I merely wrote 2 utilities, D "unix2vms" and "vms2unix".  Simple on the unix side, but the idea isF to have the students work on the VMS side.  If they  have to be loggedD into the Unix server to convert every file before moving it, nothingC is gained in trying to get them to work in the VMS environment. So, 7 I need these utilities to work on the VMS side as well.   A Now the problem.  I convert the file by writing it to a temp file D while playing with the "lf"/"lf cr" combinations and then rename theH tempfile to the original files name.  This (according to the descriptionG of "rename") guarantees that the file can not be lost even in the event E of a system crash.  But the problem is the tempfile is not created as E StreamLF.  According to the CC docs I have found on the web it should D have been as that is the default for "fopen(filename, "w")".  AnyoneD here tell me what I need to do to make sure the new file I create isD StreamLF so it doesn't end out with garbage padding at the end which5 upsets the compiler even more  thanthe lack of "cr"s.    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Sat, 27 Aug 2005 17:41:43 +0800  From: prep@prep.synonet.com 2 Subject: Re: Optimum cluster size for EVA storage?- Message-ID: <87irxr7m3s.fsf@prep.synonet.com>   & "Hein" <hein.nomail@hp.nomail> writes:  A > Yes, they should set the VMS CLUSTER SIZE to mutliples of 4 for  > Vraid5 units on EVA's   C > This is just a first step to ensure that significant IOs start as E > LBN (Logical Block Numbers) which are a multiple of 4.  The Cluster C > size setting garantuees that files at least start out evenly.  If F > the applications then continues on reading in 'even' chunks (and RDB# > will) then all is well (optimal).   E But CLUSTER SIZE only affects *allocation* of blocks, not IOs, so all C you can say is that the 1st block of the file will start on a mod 4 $ LBN. Or am I missing something here?   --  < 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: Sat, 27 Aug 2005 06:39:08 -0600 4 From: Norman Lastovica <norman.lastovica@oracle.com>2 Subject: Re: Optimum cluster size for EVA storage?* Message-ID: <43105EEC.BCD98A3D@oracle.com>   prep@prep.synonet.com wrote: > ( > "Hein" <hein.nomail@hp.nomail> writes: > C > > Yes, they should set the VMS CLUSTER SIZE to mutliples of 4 for  > > Vraid5 units on EVA's  > E > > This is just a first step to ensure that significant IOs start as G > > LBN (Logical Block Numbers) which are a multiple of 4.  The Cluster E > > size setting garantuees that files at least start out evenly.  If H > > the applications then continues on reading in 'even' chunks (and RDB% > > will) then all is well (optimal).  > G > But CLUSTER SIZE only affects *allocation* of blocks, not IOs, so all E > you can say is that the 1st block of the file will start on a mod 4 & > LBN. Or am I missing something here?  A Hein's point ("...continues on reading in 'even' chunks") is that B if the starting LBN is at a multiple-of-4 block bounary *and* the B application accesses at multiple-of-4 block boundaries, then every I/O would be 'aligned'.   B Consider the case of a utility that does 127 block sequential I/O.A If the disk cluster size is a multiple-of-4 as suggested, writing ? to the file will do an aligned transfer only once every 4 I/Os.    ------------------------------  # Date: Sat, 27 Aug 2005 08:19:19 GMT   From: John Santos <john@egh.com> Subject: Re: Q3CY205 arrived... * Message-ID: <bmVPe.1978$Uz2.1069@trnddc02>    VAXman- @SendSpamHere.ORG wrote:k > In article <1125075402.995797.48120@g14g2000cwa.googlegroups.com>, "Ed Wilts" <ewilts@ewilts.org> writes:  > F >>>Q3CY2005 SPL, SDL and an Alpha F/W update just arrived.  Before Q3! >  > ( > That was supposed to read Q4 not Q3... >  >  >  > : >>You sure you're not getting your's by Canada Post?   :-) >>- >>The rest of us had the Q3 condist for ages.  >  > N > Bugger if I know.  The box had a UPS label.  Perhaps that means Unbelievably > Pathetically S-L-O-W.  > @ > Strangely, the Q2CY2005 came only about 2 weeks prior to this.  D Brian, I think you were right the 1st time, if you meant "September"  when you said "Q3" the 2nd time.  A I also got Q2 (June) about 3 weeks ago, and Q3 (September) today.   6 Ed, I think your a quarter ahead of the rest of us :-)  A Seriously, I think HP dropped the ball on us DSPP people sometime ? between March (Q1) and August.  I usually get the SPL's late in B the previous month or early in the month printed on the CD labels,> so late August is just about right for the Sept. distribution.; Early August is about 2 months late for the June (Q2) dist.   = Due to the ancillary emails the ordering process now seems to B produce, I think the DSPP program is now ordering SDK kits throughC the normal ordering channel, but charging them back to HP.  For the D last two kits (but never before), I have received 2 emails for each,D one saying they were sending the new kits, but to ignore any billingE email I might get and a second email saying I had been billed $135.44 D for the kit, which I ignored...  (If invoices go to a separate place? in your company, then you should make sure they don't pay it by 	 mistake.)      --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  # Date: Sat, 27 Aug 2005 10:29:59 GMT " From:   VAXman-  @SendSpamHere.ORG Subject: Re: Q3CY205 arrived... 0 Message-ID: <00A48E59.3675559E@SendSpamHere.ORG>  M In article <bmVPe.1978$Uz2.1069@trnddc02>, John Santos <john@egh.com> writes: ! >VAXman- @SendSpamHere.ORG wrote: l >> In article <1125075402.995797.48120@g14g2000cwa.googlegroups.com>, "Ed Wilts" <ewilts@ewilts.org> writes: >>  G >>>>Q3CY2005 SPL, SDL and an Alpha F/W update just arrived.  Before Q3!  >>   >>  ) >> That was supposed to read Q4 not Q3...  >>   >>   >>   >>  ; >>>You sure you're not getting your's by Canada Post?   :-)  >>> . >>>The rest of us had the Q3 condist for ages. >>   >>  O >> Bugger if I know.  The box had a UPS label.  Perhaps that means Unbelievably  >> Pathetically S-L-O-W. >>  A >> Strangely, the Q2CY2005 came only about 2 weeks prior to this.  > E >Brian, I think you were right the 1st time, if you meant "September" ! >when you said "Q3" the 2nd time.    Q1: Jan/Feb/Mar  Q2: Apr/May/Jun  Q3: Jul/Aug/Sep  Q4: Oct/Nov/Dec   F Q4 was right as I wouldn't have expected to see this until much later.      B >I also got Q2 (June) about 3 weeks ago, and Q3 (September) today.  B I see your confusion.  June and Sept are the ends of the quarters.B If I see them before the end of the month of the quarter, then I'd consider them on time.        B >Seriously, I think HP dropped the ball on us DSPP people sometime@ >between March (Q1) and August.  I usually get the SPL's late inC >the previous month or early in the month printed on the CD labels, ? >so late August is just about right for the Sept. distribution. < >Early August is about 2 months late for the June (Q2) dist. > > >Due to the ancillary emails the ordering process now seems toC >produce, I think the DSPP program is now ordering SDK kits through D >the normal ordering channel, but charging them back to HP.  For theE >last two kits (but never before), I have received 2 emails for each, E >one saying they were sending the new kits, but to ignore any billing F >email I might get and a second email saying I had been billed $135.44E >for the kit, which I ignored...  (If invoices go to a separate place @ >in your company, then you should make sure they don't pay it by
 >mistake.)  B I got one of these email invoices too.  I received a followup from! DSPP to ignore the email invoice.   C It's difficult to develope software for HP's platform (which is the B purpose of the "D" in the DSPP program) without the tools which weC have paid for.  This late delivery of these tools only parlays into B a later delivery of Solutions (the "S" in the DSPP program) for HP
 platforms.   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Sat, 27 Aug 2005 04:35:58 -0400 ' From: Dave Froble <davef@tsoft-inc.com> * Subject: Re: Simh. How to triple the speed0 Message-ID: <11h0980jutss373@corp.supernews.com>   Lee Witten wrote: / > "Stanley F. Quayle" <stan@stanq.com> wrote in > > news:1124848870.716917.115110@z14g2000cwz.googlegroups.com:  >  > , >>On 23 Aug 2005 at 23:21, Dan Foster wrote: >>A >>>Just ask.  It's called "Accelerated CPU Emulation".  It caches  >>A >>frequent sequences of instructions and pre-compiles them.  It's 2 >>available in the "Plus" versions of the product. >  > L > Sounds like something I read about in the DTJ around 10 or so years ago... >  > E >>The biggest issue is knowing where the idle loop is in an operating F >>system.  Or even knowing what operating system is in use (we support! >>VAXeln and Digital Unix, too).   >  > K > One of my perverse hobbies was reading the source code for the idle loop  O > (called idle_thread()) in DU, from the very first drops from OSF/RI (back in  O > the MIPS PMAX days) through DU 5.  Boy, did that code grow!  At first it was  M > the simple spin loop that only broke when the cpu got an interrupt (either  G > due to I/O completion, timer expiry, or an inter-processor interrupt  P > triggered by another cpu).  Then, various hackers added more and more garbage K > to it.  I imagine detecting what truly is the idle loop part of the idle   > thread is very hard to do! >  > " >>Since CHARON-VAX is intended forD >>server applications, most customers don't care that 1 processor is >>running 100%.  >  > 
 > I agree. >  > --lw--  D Actually, with some of the newer processors, with the capability of C cutting back on power consumption, heat, and such when there is no  L demand for the processor, running the processor 100% isn't such a good idea.   --  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: Sat, 27 Aug 2005 10:45:07 GMT ( From: Alan Greig <greigaln@netscape.net>* Subject: Re: Simh. How to triple the speed= Message-ID: <TuXPe.33564$jr4.23913@fe3.news.blueyonder.co.uk>    Lee Witten wrote:   + > Dan Foster <usenet@evilphb.org> wrote in  / > news:slrndgnifc.omk.usenet@zappy.catbert.org:  >  >  >>Thanks for VUPS.COM. >> >>My system: >>G >>     3.0 GHz Pentium 4 + 800 MHz FSB + 1 MB L2 cache + Hyperthreading  >>          ("Northwood") < >>     OS: Linux (2.6.12 kernel with SMP enabled for HT use)+ >>     Load avg of 0.60 before running SIMH  >  > N > This doesn't sound right.  Maybe you should put the system into single user E > mode to see what performance can be achieved on a more idle system?   F Why doesn't it sound right? 12.2 VUPs is in line with what I'd expect C from simh with a 3Ghz Pentium 4 - maybe a little more squeezed out  I single user. What figures are you seeing? I get 7.8 (Windows XP gcc) out  I of an AMD XP 1800+ compiling -O3 and -march=athlon. I get about 7.5 with  F -O2 and no architecture tuning so lets use that as the base. Now 1800 G +33% = 3000 (3Ghz Pentium equiv AMD rating) and 7.5 + 33% = 12.5. Ok a  1 little faster than 12.2 but in the same ballpark.   H It's the gamers with their overclocking motherboards up at 4000+ rating E who would be heading towards 20VUPS. If you know a trick to increase   performance I'd love to know.    --  
 Alan Greig   ------------------------------   Date: 27 Aug 2005 15:17:29 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)* Subject: Re: Simh. How to triple the speed* Message-ID: <3nbeg9ForkdU1@individual.net>  0 In article <11h0980jutss373@corp.supernews.com>,* 	Dave Froble <davef@tsoft-inc.com> writes: > F > Actually, with some of the newer processors, with the capability of E > cutting back on power consumption, heat, and such when there is no  N > demand for the processor, running the processor 100% isn't such a good idea.  I Why does this sound like a good emulation of the old Motorola instruction  HCF (Halt & Catch Fire)?  :-)    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   End of INFO-VAX 2005.478 ************************