Tight while loops, Athlon 6.1 reboot

[Joseph Oravec]

Hi,

I was programming with LV 6.1 the other day and ran into an unusual problem.
I had 2 tight while-loops (without the Wait X ms) running in parallel: one
with a Sound Write Output and the other with a single channel analog
continuous read. Its the work machine, so its not in front of me right now,
but I know it was LV 6.1, Windows 2k with all service packs, Athlon CPU,
ASUS mobo, SB Live, and I forget the NI card right now.

Anyway, the machine consistently hard rebooted when these two while loops
ran in parallel. Adding a wait X ms fixed the problem and I have never seen
the machine reboot when there is a wait 500ms in each loop.

This is a little worrisome because even though the wait is a good idea for
thread fairness, I see no
good reason why the machine should reboot itself
just because of a tight loop. Has anybody else experienced this behavior
with LV 6.1?

-joey

[Aaron Marks]

I know that on some systems, this is triggered by holding the CPU at 100% too long causing it to overheat. The system reboots as a safety mechanism. I think there might actually be some safety feature in 2K that also reboots if it detects a power draw due to heavy communications traffic on the PCI bus. I haven't seen it with DAQ products though. Something to do with the ACPI (Advanced Configuration and Power Interface) settings.

[T. Schumacher]

I have had this happen with LV RT 6.1 on the PXI 8156B (AMD 333Mhz RT
controller). I have 3 loops running in parallel (each on a different
thread) and the RT controller would lock up if there are any 0 wait loops
or if any of the queue or notifier VI's have a -1 wired to the timeout
input. I placed a 5ms wait in the loops to solve the problem although
the loops run on the order of 150 ms per iteration. Thought that this
was an RT related problem, but it must be a general 6.1 and AMD issue.




"Joey Oravec" wrote in news:3c6f805e@newsgroups.:

> Hi,
>
>...
> Anyway, the machine consistently hard rebooted when these two while
> loops ran in parallel. Adding a wait X ms fixed the problem and I have
> never seen the machine reboot w
hen there is a wait 500ms in each loop.
>
> This is a little worrisome because even though the wait is a good idea
> for thread fairness, I see no good reason why the machine should
> reboot itself just because of a tight loop. Has anybody else
> experienced this behavior with LV 6.1?
>
> -joey
>
>

[Joseph Oravec]

"T. Schumacher" wrote in message
news:Xns91BAA758AEB1nonenospamcom@204.127.68.17...
> I have had this happen with LV RT 6.1 on the PXI 8156B (AMD 333Mhz RT
> controller). I have 3 loops running in parallel (each on a different
> thread) and the RT controller would lock up if there are any 0 wait loops
> or if any of the queue or notifier VI's have a -1 wired to the timeout
> input. I placed a 5ms wait in the loops to solve the problem although
> the loops run on the order of 150 ms per iteration. Thought that this
> was an RT related problem, but it must be a general 6.1 and AMD issue.

Oh no!! I hadn't noticed queue/notifier VI's locking me up with a -1 wired
to the timeout input yet, but I'm going to watch out for that now. Thank
s
for mentioning that one.

I'll chalk it up as a bug and watch out in the future. I was very
unconvinced that it was a 100% load overheating problem, it really seems to
be a general 6.1/AMD problem given people's comments so far.

-joey