Item Subject: cc:Mail Text
Hi--
I have noticed the following behaviors with VEE 4.0/NT 3.51 SP4, and
was wondering if anyone else had noticed the same, or if the VEE team
had any work-arounds:
1. The Instrument Manager drops instruments from its list.
There are 8 instruments on the HPIB bus (not incl. computer), 4 of
which are energized. If I run the VEE Development Environment, and
open the Instrument Manager, "refreshing" the HPIB card causes all
of the energized instruments to appear. If I "Save Config", exit
VEE, and then restart VEE, open the Inst. Man., the last instrument
on the list will be gone, so that only 3 appear. If I go through
this exit/restart cycle again, there will still be 3 instruments
listed. If I add the missing instrument (so that there are 4 in the
instrument manager list), exit VEE, and re-start, the last
instrument on the list will be gone again (back down to 3).
This problem is particularly vexing when trying to run programs
with the Run-Time environment, since an error message is generated
(Parse Error) and the program exits when it can't find the needed
instrument in the configuration.
2. An "On Cycle" loop hogs all of the CPU cycles on the computer.
I have written a program that reads a temperature from a
GPIB-based temperature controller at 1 second intervals. A
strip-chart of the temperature is displayed, and it is auto-scaled
after each reading. In parallel with this thread there is a
"Confirm" button executing. The CPU is a Pentium 90, Vectra XU/90C
32MB RAM. When this program is running, however, the computer's
processor time is listed as ~100%. I cannot reasonably work on any
other programs in the background, and downloading small files off
of this computer over the network while the VEE program is
executing is *very* slow. Changing the cycle interval to longer
times does not alleviate the problem.
Or is the problem the Confirm button? If so, what is the best way
to give a user the option to "abort" a program that is running?
3. In an effort to combat problem (2), I installed a second Pentium
90 MHz CPU in the Vectra, and upgraded the NT software to the
multiprocessor version, hoping that an additional processor would
provide some more horsepower for doing background tasks. Instead,
VEE programs no longer execute reliably, particularly when doing
HPIB instrument I/O. The program described in (2) no longer hogs
all of the CPU time, but instead will crash at random intervals
because of I/O time-outs. Re-installing the single-processor
version of the NT kernel restored the original behavior (reliable
instrument I/O, sluggish computer).
Thanks for any assistance you can offer,
Nate Gardner
Hewlett-Packard Company
Optoelectronics Division
370 W. Trimble Rd. MS 91-ML
San Jose, CA 95131
Hi--
I have noticed the following behaviors with VEE 4.0/NT 3.51 SP4, and
was wondering if anyone else had noticed the same, or if the VEE team
had any work-arounds:
1. The Instrument Manager drops instruments from its list.
There are 8 instruments on the HPIB bus (not incl. computer), 4 of
which are energized. If I run the VEE Development Environment, and
open the Instrument Manager, "refreshing" the HPIB card causes all
of the energized instruments to appear. If I "Save Config", exit
VEE, and then restart VEE, open the Inst. Man., the last instrument
on the list will be gone, so that only 3 appear. If I go through
this exit/restart cycle again, there will still be 3 instruments
listed. If I add the missing instrument (so that there are 4 in the
instrument manager list), exit VEE, and re-start, the last
instrument on the list will be gone again (back down to 3).
This problem is particularly vexing when trying to run programs
with the Run-Time environment, since an error message is generated
(Parse Error) and the program exits when it can't find the needed
instrument in the configuration.
2. An "On Cycle" loop hogs all of the CPU cycles on the computer.
I have written a program that reads a temperature from a
GPIB-based temperature controller at 1 second intervals. A
strip-chart of the temperature is displayed, and it is auto-scaled
after each reading. In parallel with this thread there is a
"Confirm" button executing. The CPU is a Pentium 90, Vectra XU/90C
32MB RAM. When this program is running, however, the computer's
processor time is listed as ~100%. I cannot reasonably work on any
other programs in the background, and downloading small files off
of this computer over the network while the VEE program is
executing is *very* slow. Changing the cycle interval to longer
times does not alleviate the problem.
Or is the problem the Confirm button? If so, what is the best way
to give a user the option to "abort" a program that is running?
3. In an effort to combat problem (2), I installed a second Pentium
90 MHz CPU in the Vectra, and upgraded the NT software to the
multiprocessor version, hoping that an additional processor would
provide some more horsepower for doing background tasks. Instead,
VEE programs no longer execute reliably, particularly when doing
HPIB instrument I/O. The program described in (2) no longer hogs
all of the CPU time, but instead will crash at random intervals
because of I/O time-outs. Re-installing the single-processor
version of the NT kernel restored the original behavior (reliable
instrument I/O, sluggish computer).
Thanks for any assistance you can offer,
Nate Gardner
Hewlett-Packard Company
Optoelectronics Division
370 W. Trimble Rd. MS 91-ML
San Jose, CA 95131