Hi all.
> > The two VEE applications don't share the COM port, it's the HP I/O
> > library control that controls the com port. The two VEE applications
>There you go! Of course you're exactly right. In the end, I think the
>only 'egoistic' behaviour that VEE would have to be criticised for is
>that it somehow claims a lot of the available system performance 'just
>in case', so that whatever NON-VEE-application gets horribly slow. As
Well, then again this port was a really *huge* task to pull off. I'm not
privvy to the details, but they did produce a Windows version that does work
- nevermind my nit-picky complaints about look and feel.
If I'm right about the technique Agilent used (again, I don't know the
details so that's a big "if"), then what they did was to create an X-like
environment inside Windows to host VEE in. That is no small feat. It may go
so far as emulating some of UNIX itself. If either of these assumptions are
correct, I understand why VEE needs to grab so much.
Oh! I just thought of a *great* example: do any of you guys remember
Desq-View? For those that don't, the short story is that it was an early
pre-emptive multitasker for the PC. It used to be possible to run Real or
Standard mode Windows *inside* DV, and naturally the DV fans got a real kick
out of that. Windows would cooperatively multitask it's applications while
DV preemptively multitasked everything else you cared to run.
The point is that I *think* (meaning I don't know) that's pretty much what
Agilent has accomplished here, only in reverse. They've taken a little piece
of their HP-UX and made it work in Windows. Any way you look at it, it's one
heck of an accomplishment.
-SHAWN-
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
> > The two VEE applications don't share the COM port, it's the HP I/O
> > library control that controls the com port. The two VEE applications
>There you go! Of course you're exactly right. In the end, I think the
>only 'egoistic' behaviour that VEE would have to be criticised for is
>that it somehow claims a lot of the available system performance 'just
>in case', so that whatever NON-VEE-application gets horribly slow. As
Well, then again this port was a really *huge* task to pull off. I'm not
privvy to the details, but they did produce a Windows version that does work
- nevermind my nit-picky complaints about look and feel.
If I'm right about the technique Agilent used (again, I don't know the
details so that's a big "if"), then what they did was to create an X-like
environment inside Windows to host VEE in. That is no small feat. It may go
so far as emulating some of UNIX itself. If either of these assumptions are
correct, I understand why VEE needs to grab so much.
Oh! I just thought of a *great* example: do any of you guys remember
Desq-View? For those that don't, the short story is that it was an early
pre-emptive multitasker for the PC. It used to be possible to run Real or
Standard mode Windows *inside* DV, and naturally the DV fans got a real kick
out of that. Windows would cooperatively multitask it's applications while
DV preemptively multitasked everything else you cared to run.
The point is that I *think* (meaning I don't know) that's pretty much what
Agilent has accomplished here, only in reverse. They've taken a little piece
of their HP-UX and made it work in Windows. Any way you look at it, it's one
heck of an accomplishment.
-SHAWN-
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
I am running Vee Pro6.0 on an HP Kayak XU800. As I am writing this I have 4
instances of the ForCount code that you mentioned running. I see no
degradation of computer performance at all. All four programs are humming
along with no problems. Perhaps I did not understand the test correctly?
Best Regards, Gary
Gary Nelson SPG Engineering Services
Agilent Technologies
Office: 408/435-4073 Fax: 408/435-5868
Pager: 408/589-6983
gary_nelson@agilent.com
-----Original Message-----
From: Shawn Fessenden [mailto:s_fessenden@hotmail.com]
Sent: Friday, December 14, 2001 2:31 PM
To: Thomas@Vauderwange.de; vrf@lvld.agilent.com
Subject: Re: vrf Multi-VEE-ing capability
Hallo Thomas!
>Indeed it is no problem at all
>to have several different VEE
>programs running on the same PC
>the same time.
I'm officially flabbergasted and will honor my pledge to remain silent on
the subject. Something is zinging right over my head. I can't imagine how
you're doing it. Let me set up a simple test and then I'm done.
ForCount to #H7fffffff->AlphanumericDisplay.
Start two instances of VEE and run this in each. Place them side by side and
watch. What I see is each one incrementing slowly. If I pop both VEEs down
and try to do something else... ANYTHING else, the computer has returned to
the days of trying to run Windows in Standard mode on a 286 with a 2M EMS
board: It takes forever to do anything at all and one gives up long before
there's any screen activity.
VEE is 5.01, disable debug is on, mode is standard. CPU activity is pegged
at 100% and shows no sign of letting up. Both VEEs are crawling and nothing
useful is possible. I just did it again and tried to start pview to check
working set size, but after a couple minutes pview still haddn't appeared
and I'm just too impatient to find out what the WSS is.
If I stop one, then the other picks up and flies like it should and my
computer is back to being only tollerably impared.
And that's a mystery too. I know Windows fairly well (I mean in detail, not
just it's operation but it's internals, quirks, trivia & sundry) and I'm at
a loss to explain how just running a program in even a single instance of
VEE can degrade performance to such a degree in the first place, yet it
does. When I'm running a VEE program, that's all I'm doing except maybe
typing in notepad. Even that's a hassle because it can't keep up with me.
The laptop is a slightly decreped HP Omnibook XE2 (PII 233), 128M W2K. The
desktop is a little better off: AMD K-Flex 6-2 500 256M W98. The best
compter I tried this on was a Compaq PIII 7something-or-other (733? 740?) NT
4.0. In every case, the result is the same.
I offer sincere apologies to everyone. I thought everybody saw the same
exact behavior. I understand now that's not the case.
I've tried this at least a hundred different ways on different
installations, different hardware, different Windows versions, all
non-essential processes are killed, - everything I could think of and the
result is always the same. I can't explain it at all given that you don't
have this problem.
I would really appreciate it if somebody could figure this out, but I have
pledged to remain publicly silent and so I will. Private communications are,
however, encouraged
-SHAWN-
_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
>Indeed it is no problem at all
>to have several different VEE
>programs running on the same PC
>the same time.
I'm officially flabbergasted and will honor my pledge to remain silent on
the subject. Something is zinging right over my head. I can't imagine how
you're doing it. Let me set up a simple test and then I'm done.
ForCount to #H7fffffff->AlphanumericDisplay.
Start two instances of VEE and run this in each. Place them side by side and
watch. What I see is each one incrementing slowly. If I pop both VEEs down
and try to do something else... ANYTHING else, the computer has returned to
the days of trying to run Windows in Standard mode on a 286 with a 2M EMS
board: It takes forever to do anything at all and one gives up long before
there's any screen activity.
VEE is 5.01, disable debug is on, mode is standard. CPU activity is pegged
at 100% and shows no sign of letting up. Both VEEs are crawling and nothing
useful is possible. I just did it again and tried to start pview to check
working set size, but after a couple minutes pview still haddn't appeared
and I'm just too impatient to find out what the WSS is.
If I stop one, then the other picks up and flies like it should and my
computer is back to being only tollerably impared.
And that's a mystery too. I know Windows fairly well (I mean in detail, not
just it's operation but it's internals, quirks, trivia & sundry) and I'm at
a loss to explain how just running a program in even a single instance of
VEE can degrade performance to such a degree in the first place, yet it
does. When I'm running a VEE program, that's all I'm doing except maybe
typing in notepad. Even that's a hassle because it can't keep up with me.
The laptop is a slightly decreped HP Omnibook XE2 (PII 233), 128M W2K. The
desktop is a little better off: AMD K-Flex 6-2 500 256M W98. The best
compter I tried this on was a Compaq PIII 7something-or-other (733? 740?) NT
4.0. In every case, the result is the same.
I offer sincere apologies to everyone. I thought everybody saw the same
exact behavior. I understand now that's not the case.
I've tried this at least a hundred different ways on different
installations, different hardware, different Windows versions, all
non-essential processes are killed, - everything I could think of and the
result is always the same. I can't explain it at all given that you don't
have this problem.
I would really appreciate it if somebody could figure this out, but I have
pledged to remain publicly silent and so I will. Private communications are,
however, encouraged
-SHAWN-
_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
You have a very interesting problem and I can see how frustrating this can
be. I have three instances of VEE running now with a ForCount object as you
described and they are all running well. In fact, sometimes it looks as
though it's running so fast that I can't read the display very well. I also
surfed the web to see if any of the three instances would slow down and none
did.
Like Gary, I'm running VEE 6.0 but on an HP Kayak XU6/450, Windows NT 4.0.
Could the problem you're seeing be a bug with VEE 5.0? Did earlier versions
of VEE cause the same problem? I don't have a version 5.0 of VEE, otherwise
I would try it. If you ever find out what's causing it, I would be very
interested in reading about it.
Regards,
Andie
-----Original Message-----
From: Shawn Fessenden [mailto:s_fessenden@hotmail.com]
Sent: Friday, December 14, 2001 2:31 PM
To: Thomas@Vauderwange.de; vrf@lvld.agilent.com
Subject: Re: vrf Multi-VEE-ing capability
Hallo Thomas!
>Indeed it is no problem at all
>to have several different VEE
>programs running on the same PC
>the same time.
I'm officially flabbergasted and will honor my pledge to remain silent on
the subject. Something is zinging right over my head. I can't imagine how
you're doing it. Let me set up a simple test and then I'm done.
ForCount to #H7fffffff->AlphanumericDisplay.
Start two instances of VEE and run this in each. Place them side by side and
watch. What I see is each one incrementing slowly. If I pop both VEEs down
and try to do something else... ANYTHING else, the computer has returned to
the days of trying to run Windows in Standard mode on a 286 with a 2M EMS
board: It takes forever to do anything at all and one gives up long before
there's any screen activity.
VEE is 5.01, disable debug is on, mode is standard. CPU activity is pegged
at 100% and shows no sign of letting up. Both VEEs are crawling and nothing
useful is possible. I just did it again and tried to start pview to check
working set size, but after a couple minutes pview still haddn't appeared
and I'm just too impatient to find out what the WSS is.
If I stop one, then the other picks up and flies like it should and my
computer is back to being only tollerably impared.
And that's a mystery too. I know Windows fairly well (I mean in detail, not
just it's operation but it's internals, quirks, trivia & sundry) and I'm at
a loss to explain how just running a program in even a single instance of
VEE can degrade performance to such a degree in the first place, yet it
does. When I'm running a VEE program, that's all I'm doing except maybe
typing in notepad. Even that's a hassle because it can't keep up with me.
The laptop is a slightly decreped HP Omnibook XE2 (PII 233), 128M W2K. The
desktop is a little better off: AMD K-Flex 6-2 500 256M W98. The best
compter I tried this on was a Compaq PIII 7something-or-other (733? 740?) NT
4.0. In every case, the result is the same.
I offer sincere apologies to everyone. I thought everybody saw the same
exact behavior. I understand now that's not the case.
I've tried this at least a hundred different ways on different
installations, different hardware, different Windows versions, all
non-essential processes are killed, - everything I could think of and the
result is always the same. I can't explain it at all given that you don't
have this problem.
I would really appreciate it if somebody could figure this out, but I have
pledged to remain publicly silent and so I will. Private communications are,
however, encouraged
-SHAWN-
_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
Started one program (background), then started the other (foreground) and
timed with a stopwatch for 20 seconds
Result display: 33000
Then I stopped the background VEE, and re-ran the experiment with the
stopwatch on the foreground VEE
Result display: 39000
Then I quit the background VEE process and timed the foreground one for 20
seconds
Result display: 39000
The numbers are rough because the experiment is rough.
Another somewhat subjective point I noted: the foreground (window focus) VEE
process went fast, while the background one went noticeably slower. Not what
I'd called dead slow or stalled, but maybe "sluggish". When I clicked the
background process, it sped up as soon as it popped to the front, and seemed
to run at about the 39000 speed (subjective estimate)
So, it looks like two VEEs share the processor "reasonably", but not
"fairly". Putting a processor intensive VEE program in the background while
another processor intensive VEE program runs in foreground probably
decreases the speed of the background considerably, possibly as much as an
order of magnitude.
Disclaimers (but not legal ones, just commonsense ones)
I'd hesitate to extrapolate to 3 or more VEEs without further experiments.
I'd want to check with other programs than VEE as the background process, if
that's what my production system did. Better experiments could easily be
devised (hey, it's Friday afternoon!). This experiment attempts to
characterize only a processor-intensive VEE programs, with a
processor-intensive background VEE. I/O or other "shared resources" could
change results considerably. You'd need to test with your own proposed mix
or a reasonable simulation before you could be confident that results
predicted your performance. More complex programs, say with lots of threads,
for example, might show wildly different results.
But running more than one VEE at a time for many kinds of tasks is not
necessarily infeasible.
Sincerely,
Scott Bayes
Software Technical Support
Agilent Technologies, Inc.
815 14th Street S.W.
Loveland, CO, U.S.A. 80537
970 679 3799 Tel
970 635 6867 Fax
> -----Original Message-----
> From: Shawn Fessenden [mailto:s_fessenden@hotmail.com]
> Sent: Friday, December 14, 2001 3:31 PM
> To: Thomas@Vauderwange.de; vrf@lvld.agilent.com
> Subject: Re: vrf Multi-VEE-ing capability
>
>
> Hallo Thomas!
>
> >Indeed it is no problem at all
> >to have several different VEE
> >programs running on the same PC
> >the same time.
>
> I'm officially flabbergasted and will honor my pledge to
> remain silent on
> the subject. Something is zinging right over my head. I can't
> imagine how
> you're doing it. Let me set up a simple test and then I'm done.
>
> ForCount to #H7fffffff->AlphanumericDisplay.
>
> Start two instances of VEE and run this in each. Place them
> side by side and
> watch. What I see is each one incrementing slowly. If I pop
> both VEEs down
> and try to do something else... ANYTHING else, the computer
> has returned to
> the days of trying to run Windows in Standard mode on a 286
> with a 2M EMS
> board: It takes forever to do anything at all and one gives
> up long before
> there's any screen activity.
>
> VEE is 5.01, disable debug is on, mode is standard. CPU
> activity is pegged
> at 100% and shows no sign of letting up. Both VEEs are
> crawling and nothing
> useful is possible. I just did it again and tried to start
> pview to check
> working set size, but after a couple minutes pview still
> haddn't appeared
> and I'm just too impatient to find out what the WSS is.
> If I stop one, then the other picks up and flies like it
> should and my
> computer is back to being only tollerably impared.
>
> And that's a mystery too. I know Windows fairly well (I mean
> in detail, not
> just it's operation but it's internals, quirks, trivia &
> sundry) and I'm at
> a loss to explain how just running a program in even a single
> instance of
> VEE can degrade performance to such a degree in the first
> place, yet it
> does. When I'm running a VEE program, that's all I'm doing
> except maybe
> typing in notepad. Even that's a hassle because it can't keep
> up with me.
>
> The laptop is a slightly decreped HP Omnibook XE2 (PII 233),
> 128M W2K. The
> desktop is a little better off: AMD K-Flex 6-2 500 256M W98. The best
> compter I tried this on was a Compaq PIII 7something-or-other
> (733? 740?) NT
> 4.0. In every case, the result is the same.
>
> I offer sincere apologies to everyone. I thought everybody
> saw the same
> exact behavior. I understand now that's not the case.
>
> I've tried this at least a hundred different ways on different
> installations, different hardware, different Windows versions, all
> non-essential processes are killed, - everything I could
> think of and the
> result is always the same. I can't explain it at all given
> that you don't
> have this problem.
>
> I would really appreciate it if somebody could figure this
> out, but I have
> pledged to remain publicly silent and so I will. Private
> communications are,
> however, encouraged
> -SHAWN-
>
>
> _________________________________________________________________
> Send and receive Hotmail on your mobile device: http://mobile.msn.com
>
> ---------------------------------------------------------------------
> This is the "vrf" maillist, managed by Majordomo. To send messages to
> this maillist, just email to "vrf@lvld.agilent.com".
> Subscriptions and
> unsubscriptions are done through the address
> "vrf-request@lvld.agilent.com".
> If you need details, just send a message containing the text "help"
> to "vrf-request@lvld.agilent.com".
> ---------------------------------------------------------------------
>
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
quick, no, it's not a bug in VEE, it's a bug in me somewhere
I first noticed in 3.12, however it went away for everybody but me by VEE 4.
If and when I figure out exactly what's going on, I'll make sure to let
everybody know.
-SHAWN-
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp. is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
comment:
>So, it looks like two VEEs share
>the processor "reasonably", but not
>"fairly".
This is exactly how it should work. What we see here is Windows' task
scheduler divying up time to applications. On W2K systems, these ratios can
be altered in the Control Panel-> System applette: on the Advanced tab,
click the Performance button. There you'll find two radio buttons that
change how the task scheduler works. On NT 4 systems, the option was in the
same general location but there was a slider called "Foreground Application
Boost". This slider performed basically the same function.
-SHAWN-
_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
Indeed it is no problem at all to have several different VEE programs running on
the same PC the same time. Even tricky stuff, like RS232-comms etc. My customers
and the folks down in the production hall do this all the time with my software
(Win98, NT4, W2K), and everybody's happy. This happened when I changed from good
ol' VEE3-interpreter (where it did not work) to the one that came with VEE5.
And: someone found out by coincidence by just mistakenly starting several
processes. Even sharing one and the same COM-port (!!) worked without having
been specified!
All I had to do to make this a real feature is to specify different working
directories where different config-files could be placed. And, of course, to
allow for an identifier string in the left top of the temperature graph, because
otehrwise nobody could tell the processes apart...
Regards,
Thomas
Shawn Fessenden schrieb:
> Well thank you everybody for your comments!
>
> There are some real eye-openers there, especially Jens. While I strongly
> disagree with VEE's affinity for "structured" code, I suspect that's simply
> because we have different ideas of what structured code is but, "that's
> probably a discussion more suited to a comp.programming group," (he said,
> trying to stay on topic).
>
> On the other side of the fence, VEE's innate flowchart-like diagramatic /
> execution model is definitely a godsend. In fact, whenever customers
> explicitly request VEE this model is almost always the reason why. When I
> first started using VEE, I just assumed customers wanted VEE because it was
> an HP product.
>
> Anyway Jens, I found this particularly interesting:
>
> >2 VEE clients and the VEE-controller
> >can run on the same PC or on different PCs.
>
> If I understand this correctly you can run *three* instances of VEE on *one*
> computer? If this is the case, then my hat is off to you and I will
> henceforth keep my mouth shut!
>
> Using NT 4.0 I've had terrible problems just keeping *one* instance going on
> anything less than a PII 233! At the risk of boring everybody, let me
> outline two of these problems. I really need to understand why it is I have
> such awful problems and others do not. At the time, we had HP field
> application engineers crawling all over what I did and their conclusion was
> "It's fine, that's just the way it works". Ever since then I've been saying,
> "well then it doesn't work right".
>
> These two problems are actually just different aspects of the same root
> cause, CPU hogging:
>
> 1. Printing. Customer wants a simplistic graph of readings over time. Using
> the Print Screen object, it took over 45 minutes to print one graph. This is
> over a 100MB network. Spooler priority was set highest. I abandoned the
> spooler and of course that speeded up the printing to a minute or two, but
> the UI completely stops responding (of course) and that's unacceptable. I
> wound up using Microsoft's DIBAPI to capture the screen myself & print it.
> The catch was that this utility had to shift itself to HIGH_PRIORITY_CLASS
> to produce a graph in under 5 minutes. This slowed down the UI some, but the
> customer accepted the result as a necessary trade-off.
>
> 2. Browsing. The test scenario is a room full of some 100 UUTs. Test time
> varies from 4 hours to three or four days sometimes. Data points consist of
> samples of 6 channels every 30 seconds. The DAC box is an Agilent 34970, one
> per UUT. Loading the test line is pretty simple. One plug for power, one for
> 4 data channels. The other two channels are from sensors placed inside the
> UUT by test personnel. A bar code is scanned and the test cycle begins.
>
> Now, the people who run the test system are union laborers, not engineers.
> There are usually three people per test room. It takes all of about an hour
> to get a full room under way and then they sit back and wait. It's no fun
> sitting there staring at a graph or a summary of all running tests, so they
> usually start playing solitare or browsing the web, looking for girly-pix.
> Now granted that this isn't the ideal scenario, but realistically it's to be
> expected. As soon as VEE is put in the background, problems develop.
>
> In the first place, after about 40 UUTs VEE starts having a hard time
> keeping up with sampling data, updating 34970 displays, checking for card
> swaps or entire 34970 swaps, verifying 34970 configuration, analyzing data
> to decide on shifts in test phases, testing com ports for proper
> communication, watchdog pinging other test computers to make sure they're
> ok, etc. This is even with VEE in the bloody FOREGROUND. When it gets put in
> the background, look out!
>
> Now I realize full well that I'm no VEE expert. Like I've said, 30 years of
> programming doesn't mean a thing when it comes to VEE so we called in HP
> experts for expert advice. After over half a year of tweaking and some
> truely extraordinary efforts on the part of various HP engineers, the
> customer was still dissatisfied and started threatening to pull out, and for
> that I don't blame them. They dumped *huge* amounts of cash on hardware and
> they couldn't even run at 50% capacity.
>
> Even then, memory leaks caused by the infamous error pins would, after only
> a few days of continuous operation, cause VEE to reset it's heap (or
> something) and all in-progress tests were lost. Granted, there was always a
> current status backup stored on the drive, but this was updated every time
> data changed on any DAC box and the moment VEE lost it's variables this file
> was overwritten with emptyness. The individual data files themselves
> remained, but without status information (current test phase, phase test
> instructions, current machine state, state machine registers, etc.) the
> state machine couldn't possibly continue.
>
> It was finally decided to move to a compiled language so I locked myself in
> a closet for 14 weeks and re-wrote the entire thing in C++. And when I say
> re-wrote, I mean exactly that. The C++ version is almost an exact duplicate
> of the VEE version. VEE's failures in this instance were definitely not due
> to software architecture. This application is currently buzzing happily
> away, testing like crazy. There are no issues with printing graphs, playing
> solitare or browsing the web.
>
> The *opinions* I express here are based largely on this one extraordinarily
> negative experience. This was definitely the most detailed analysis I've
> ever made of VEE on Windows in the wild, though I've had plenty of somewhat
> negative experiences with it before. The icing on the cake was HP's own
> conclusion that "yes, everything is fine. That's the way it works". Since
> then I've recommended to those customers who want to use VEE that they then
> should stick with HP-UX. Sometimes that's fine, and sometimes it's not
> because Windows is a requirement. In that case, we use "Windows specific"
> tools like Visual Basic, Visual C++, Delphi or C++ Builder.
>
> Now I challenge any of you to put yourself in my shoes: Put your life on
> hold for 14 weeks of consecutive 16-20 hour days WITHOUT PAY, bear the whip
> from management AND the customer, watch HP go over everything you've done
> with a fine-tooth comb and shrug their shoulders, watch five years of your
> life helping to build a company from a two-man operation to a small global
> enterprise go down the drain as law suits and bankruptcy ensue and then have
> anything positive at all to say about VEE. Here it is, on public record:
>
> IN SPITE OF ALL OF THIS I STILL LIKE VEE!!!
>
> So please, if anybody has anything to say to me personally about VEE bashing
> then lets keep it off the list as it probably won't be productive at all and
> I guarantee I'll respond to all flames with a blow-torch.
> -SHAWN-
>
> _________________________________________________________________
> Join the worlds largest e-mail service with MSN Hotmail.
> http://www.hotmail.com
>
> ---------------------------------------------------------------------
> This is the "vrf" maillist, managed by Majordomo. To send messages to
> this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
> unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
> If you need details, just send a message containing the text "help"
> to "vrf-request@lvld.agilent.com".
> ---------------------------------------------------------------------
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
<font size=-1>Staff Engineer</font>
<font size=-1>Test Engineering</font>
<font size=-1>Ericsson Amplifier Technologies, Inc.</font>
<font size=-1>49 Wireless Blvd.</font>
<font size=-1>Hauppaugh, New York 11788</font>
<font size=-1>Phone: (631) 357-8513</font>
<font size=-1>E-Mail: paul.rubin@ericsson.com</font>
<p><font size=-1>-----Original Message-----</font>
<font size=-1>From: Thomas Vauderwange [<a href="mailto:Thomas@Vauderwange.de">mailto:Thomas@Vauderwange.de</a>]</font>
<font size=-1>Sent: Friday, December 14, 2001 4:04 PM</font>
<font size=-1>To: vrf@lvld.agilent.com</font>
<font size=-1>Subject: vrf Multi-VEE-ing capability</font><p><font size=-1>Hi Shawn,</font><p><font size=-1>Indeed it is no problem at all to have several differentVEE programs running on</font>
<font size=-1>the same PC the same time. Even tricky stuff, like RS232-commsetc. My customers</font>
<font size=-1>and the folks down in the production hall do this allthe time with my software</font>
<font size=-1>(Win98, NT4, W2K), and everybody's happy. This happenedwhen I changed from good</font>
<font size=-1>ol' VEE3-interpreter (where it did not work) to the onethat came with VEE5.</font>
<font size=-1>And: someone found out by coincidence by just mistakenlystarting several</font>
<font size=-1>processes. Even sharing one and the same COM-port (!!)worked without having</font>
<font size=-1>been specified!</font>
<font size=-1>All I had to do to make this a real feature is to specifydifferent working</font>
<font size=-1>directories where different config-files could be placed.And, of course, to</font>
<font size=-1>allow for an identifier string in the left top of thetemperature graph, because</font>
<font size=-1>otehrwise nobody could tell the processes apart...</font><p><font size=-1>Regards,</font><p><font size=-1>Thomas</font><p><font size=-1>Shawn Fessenden schrieb:</font><p><font size=-1>> Well thank you everybody for your comments!</font>
<font size=-1>></font>
<font size=-1>> There are some real eye-openers there, especially Jens.While I strongly</font>
<font size=-1>> disagree with VEE's affinity for "structured" code,I suspect that's simply</font>
<font size=-1>> because we have different ideas of what structuredcode is but, "that's</font>
<font size=-1>> probably a discussion more suited to a comp.programminggroup," (he said,</font>
<font size=-1>> trying to stay on topic).</font>
<font size=-1>></font>
<font size=-1>> On the other side of the fence, VEE's innate flowchart-likediagramatic /</font>
<font size=-1>> execution model is definitely a godsend. In fact, whenevercustomers</font>
<font size=-1>> explicitly request VEE this model is almost alwaysthe reason why. When I</font>
<font size=-1>> first started using VEE, I just assumed customers wantedVEE because it was</font>
<font size=-1>> an HP product.</font>
<font size=-1>></font>
<font size=-1>> Anyway Jens, I found this particularly interesting:</font>
<font size=-1>></font>
<font size=-1>> >2 VEE clients and the VEE-controller</font>
<font size=-1>> >can run on the same PC or on different PCs.</font>
<font size=-1>></font>
<font size=-1>> If I understand this correctly you can run *three*instances of VEE on *one*</font>
<font size=-1>> computer? If this is the case, then my hat is off toyou and I will</font>
<font size=-1>> henceforth keep my mouth shut!</font>
<font size=-1>></font>
<font size=-1>> Using NT 4.0 I've had terrible problems just keeping*one* instance going on</font>
<font size=-1>> anything less than a PII 233! At the risk of boringeverybody, let me</font>
<font size=-1>> outline two of these problems. I really need to understandwhy it is I have</font>
<font size=-1>> such awful problems and others do not. At the time,we had HP field</font>
<font size=-1>> application engineers crawling all over what I didand their conclusion was</font>
<font size=-1>> "It's fine, that's just the way it works". Ever sincethen I've been saying,</font>
<font size=-1>> "well then it doesn't work right".</font>
<font size=-1>></font>
<font size=-1>> These two problems are actually just different aspectsof the same root</font>
<font size=-1>> cause, CPU hogging:</font>
<font size=-1>></font>
<font size=-1>> 1. Printing. Customer wants a simplistic graph of readingsover time. Using</font>
<font size=-1>> the Print Screen object, it took over 45 minutes toprint one graph. This is</font>
<font size=-1>> over a 100MB network. Spooler priority was set highest.I abandoned the</font>
<font size=-1>> spooler and of course that speeded up the printingto a minute or two, but</font>
<font size=-1>> the UI completely stops responding (of course) andthat's unacceptable. I</font>
<font size=-1>> wound up using Microsoft's DIBAPI to capture the screenmyself & print it.</font>
<font size=-1>> The catch was that this utility had to shift itselfto HIGH_PRIORITY_CLASS</font>
<font size=-1>> to produce a graph in under 5 minutes. This sloweddown the UI some, but the</font>
<font size=-1>> customer accepted the result as a necessary trade-off.</font>
<font size=-1>></font>
<font size=-1>> 2. Browsing. The test scenario is a room full of some100 UUTs. Test time</font>
<font size=-1>> varies from 4 hours to three or four days sometimes.Data points consist of</font>
<font size=-1>> samples of 6 channels every 30 seconds. The DAC boxis an Agilent 34970, one</font>
<font size=-1>> per UUT. Loading the test line is pretty simple. Oneplug for power, one for</font>
<font size=-1>> 4 data channels. The other two channels are from sensorsplaced inside the</font>
<font size=-1>> UUT by test personnel. A bar code is scanned and thetest cycle begins.</font>
<font size=-1>></font>
<font size=-1>> Now, the people who run the test system are union laborers,not engineers.</font>
<font size=-1>> There are usually three people per test room. It takesall of about an hour</font>
<font size=-1>> to get a full room under way and then they sit backand wait. It's no fun</font>
<font size=-1>> sitting there staring at a graph or a summary of allrunning tests, so they</font>
<font size=-1>> usually start playing solitare or browsing the web,looking for girly-pix.</font>
<font size=-1>> Now granted that this isn't the ideal scenario, butrealistically it's to be</font>
<font size=-1>> expected. As soon as VEE is put in the background,problems develop.</font>
<font size=-1>></font>
<font size=-1>> In the first place, after about 40 UUTs VEE startshaving a hard time</font>
<font size=-1>> keeping up with sampling data, updating 34970 displays,checking for card</font>
<font size=-1>> swaps or entire 34970 swaps, verifying 34970 configuration,analyzing data</font>
<font size=-1>> to decide on shifts in test phases, testing com portsfor proper</font>
<font size=-1>> communication, watchdog pinging other test computersto make sure they're</font>
<font size=-1>> ok, etc. This is even with VEE in the bloody FOREGROUND.When it gets put in</font>
<font size=-1>> the background, look out!</font>
<font size=-1>></font>
<font size=-1>> Now I realize full well that I'm no VEE expert. LikeI've said, 30 years of</font>
<font size=-1>> programming doesn't mean a thing when it comes to VEEso we called in HP</font>
<font size=-1>> experts for expert advice. After over half a year oftweaking and some</font>
<font size=-1>> truely extraordinary efforts on the part of variousHP engineers, the</font>
<font size=-1>> customer was still dissatisfied and started threateningto pull out, and for</font>
<font size=-1>> that I don't blame them. They dumped *huge* amountsof cash on hardware and</font>
<font size=-1>> they couldn't even run at 50% capacity.</font>
<font size=-1>></font>
<font size=-1>> Even then, memory leaks caused by the infamous errorpins would, after only</font>
<font size=-1>> a few days of continuous operation, cause VEE to resetit's heap (or</font>
<font size=-1>> something) and all in-progress tests were lost. Granted,there was always a</font>
<font size=-1>> current status backup stored on the drive, but thiswas updated every time</font>
<font size=-1>> data changed on any DAC box and the moment VEE lostit's variables this file</font>
<font size=-1>> was overwritten with emptyness. The individual datafiles themselves</font>
<font size=-1>> remained, but without status information (current testphase, phase test</font>
<font size=-1>> instructions, current machine state, state machineregisters, etc.) the</font>
<font size=-1>> state machine couldn't possibly continue.</font>
<font size=-1>></font>
<font size=-1>> It was finally decided to move to a compiled languageso I locked myself in</font>
<font size=-1>> a closet for 14 weeks and re-wrote the entire thingin C++. And when I say</font>
<font size=-1>> re-wrote, I mean exactly that. The C++ version is almostan exact duplicate</font>
<font size=-1>> of the VEE version. VEE's failures in this instancewere definitely not due</font>
<font size=-1>> to software architecture. This application is currentlybuzzing happily</font>
<font size=-1>> away, testing like crazy. There are no issues withprinting graphs, playing</font>
<font size=-1>> solitare or browsing the web.</font>
<font size=-1>></font>
<font size=-1>> The *opinions* I express here are based largely onthis one extraordinarily</font>
<font size=-1>> negative experience. This was definitely the most detailedanalysis I've</font>
<font size=-1>> ever made of VEE on Windows in the wild, though I'vehad plenty of somewhat</font>
<font size=-1>> negative experiences with it before. The icing on thecake was HP's own</font>
<font size=-1>> conclusion that "yes, everything is fine. That's theway it works". Since</font>
<font size=-1>> then I've recommended to those customers who want touse VEE that they then</font>
<font size=-1>> should stick with HP-UX. Sometimes that's fine, andsometimes it's not</font>
<font size=-1>> because Windows is a requirement. In that case, weuse "Windows specific"</font>
<font size=-1>> tools like Visual Basic, Visual C++, Delphi or C++Builder.</font>
<font size=-1>></font>
<font size=-1>> Now I challenge any of you to put yourself in my shoes:Put your life on</font>
<font size=-1>> hold for 14 weeks of consecutive 16-20 hour days WITHOUTPAY, bear the whip</font>
<font size=-1>> from management AND the customer, watch HP go overeverything you've done</font>
<font size=-1>> with a fine-tooth comb and shrug their shoulders, watchfive years of your</font>
<font size=-1>> life helping to build a company from a two-man operationto a small global</font>
<font size=-1>> enterprise go down the drain as law suits and bankruptcyensue and then have</font>
<font size=-1>> anything positive at all to say about VEE. Here itis, on public record:</font>
<font size=-1>></font>
<font size=-1>> IN SPITE OF ALL OF THIS I STILL LIKE VEE!!!</font>
<font size=-1>></font>
<font size=-1>> So please, if anybody has anything to say to me personallyabout VEE bashing</font>
<font size=-1>> then lets keep it off the list as it probably won'tbe productive at all and</font>
<font size=-1>> I guarantee I'll respond to all flames with a blow-torch.</font>
<font size=-1>> -SHAWN-</font>
<font size=-1>></font>
<font size=-1>> _________________________________________________________________</font>
<font size=-1>> Join the world's largest e-mail service with MSN Hotmail.</font>
<font size=-1>> <a href="http://www.hotmail.com" TARGET="_blank">http://www.hotmail.com</a></font>
<font size=-1>></font>
<font size=-1>> ---------------------------------------------------------------------</font>
<font size=-1>> This is the "vrf" maillist, managed by Majordomo. To send messages to</font>
<font size=-1>> this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and</font>
<font size=-1>> unsubscriptions are done through the address "vrf-request@lvld.agilent.com".</font>
<font size=-1>> If you need details, just send a message containingthe text "help"</font>
<font size=-1>> to "vrf-request@lvld.agilent.com".</font>
<font size=-1>> ---------------------------------------------------------------------</font><p><font size=-1>---------------------------------------------------------------------</font>
<font size=-1>This is the "vrf" maillist, managed by Majordomo. To send messages to</font>
<font size=-1>this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and</font>
<font size=-1>unsubscriptions are done through the address "vrf-request@lvld.agilent.com".</font>
<font size=-1>If you need details, just send a message containing thetext "help"</font>
<font size=-1>to "vrf-request@lvld.agilent.com".</font>
<font size=-1>---------------------------------------------------------------------</font></blockquote></html>
To answer Paul's question: No, I don't remember that actually morethan two VEE-applications did share the same COM-port. So, in spite thatI'd expect it to work, I think Scott makes a point in his statement thatthe extrapolation of that idea to more than two instances would be up toan experiment.<p>Regards,<p>Thomas Vauderwange
<p>Gerard Geurts schrieb:<blockquote TYPE=CITE><style></style><font face="Arial"><font size=-1>Thetwo VEE applications don't share the COM port, it's the HP I/O librarycontrol that controls the com port. The two VEE applications are both talkingto HP I/O library control, which in turns talks to the com port. Ergo,only one application actually accesses the com port.</font></font> <font face="Arial"><font size=-1>Cheers</font></font> <font face="Arial"><font size=-1>GerardGeurts</font></font><font face="Arial"><font size=-1>G Squared Ltd</font></font> -----Original Message -----<blockquote style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"><div style="BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: black"><b>From:</b><a href="mailto:Paul.Rubin@am1.ericsson.se" title="Paul.Rubin@am1.ericsson.se">PaulRubin (AMT)</a></div><div style="FONT: 10pt arial"><b>To:</b> <a href="mailto:'Thomas@Vauderwange.de'" title="Thomas@Vauderwange.de">'Thomas@Vauderwange.de'</a>; <a href="mailto:vrf@lvld.agilent.com" title="vrf@lvld.agilent.com">vrf@lvld.agilent.com</a></div><div style="FONT: 10pt arial"><b>Sent:</b> Friday, December 14, 2001 9:53PM</div><div style="FONT: 10pt arial"><b>Subject:</b> RE: vrf Multi-VEE-ing capability</div> <span class=181394921-14122001><font face="Arial"><font color="#0000FF"><font size=-1>Thomas,</font></font></font></span><span class=181394921-14122001></span><span class=181394921-14122001><font face="Arial"><font color="#0000FF"><font size=-1>Apparentlyyou are only claiming that 2 _VEE_ programs can "share" a Comport. Correct? Have you ever succeeded in having VEE share a Com port withany other type application?</font></font></font></span><b><font face="Arial"><font size=-1>PaulH. Rubin</font></font></b>
<font face="Arial"><font size=-1>Staff Engineer</font></font>
<font face="Arial"><font size=-1>Test Engineering</font></font>
<font face="Arial"><font size=-1>Ericsson Amplifier Technologies, Inc.</font></font>
<font face="Arial"><font size=-1>49 Wireless Blvd.</font></font>
<font face="Arial"><font size=-1>Hauppauge, New York 11788</font></font>
<font face="Arial"><font size=-1>Phone: (631) 357-8513</font></font>
<font face="Arial"><font size=-1>E-Mail: <a href="mailto:paul.rubin@ericsson.com">paul.rubin@ericsson.com</a></font></font></blockquote></blockquote></body></html>
Just fired up 4 VEE 6.01 sessions on my PII 450, NT4 SP6 system
and with all 4 running the loop indicated I see 80% load factor
in task manager , fairly evenly shared at 20% and other applications
are a trifle slow but nothing serious.
When I am using VEE I almost _always_ have at least two
copies running since I deal with lots of libraries.
Are you perhaps running into swap? I could certainly see
slowdown in that case.
Or if you were running one of the consumer-grade OS's
(9x, ME).
If not- and you are staying in RAM and aren't running funny
in the background (like a resource hog like Netscape)
then sorry but I have no idea why you would see a slowdown
like you mention.
I have various systems to play with. If you can specify
what sort of OS/processor/ram/swap/patch levels/etc you have I'll
see if I can reproduce what you are seeing.
regards
Stan
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------