CC Hangs in my Dual Xeon system ...

[cp10932]

CC works very well in my uni-P4 system and, in my opinion, a better software than LCDMax. However, I have not been able to get CC to run reliably on my dual Xeon system; it would hang within a few hours. When it hangs, the LCD would freeze and I would not be able to bring up CC to change/restart. The only way out of this hang is to kill the CC process and restart CC. When CC is started, I also lose all the screens that I created. Could CCAdmin help me narrow down what's wrong? Is there a way to save the CC configuration so that I can reload it?

Thanks.

Crystalfontz LCD Blog

Looking for additional LCD resources? Check out our LCD blog for the latest developments in LCD technology.

 

[CF Tech]

Hmmm... i cant see why a MP system would cause CC to break.
The OS might have somthing to do with it crashing... never can tell

Settings are saved on closing of CC.
If you set CC up the way you like, then close it, the settings will be there next time.

 

[cp10932]

Thanks. I am also trying a few different things to see if I can narrow down the cause ... it's kind of a long shot though.

 

[cp10932]

CCAdmin, this is definitely just a CC problem on my SMP rig. I ran Win Test with fan speed monitoring on for hours and it runs fine; no hang. CC would hang within, at most, a couple of hours, if not quicker.

 

[cp10932]

I am still not making much headway on this ... may be I will try using the USB-serial converter and see if the USB path is better ...

 

[cp10932]

Well, I tried with the IOGEAR USB-Serial Adapter and it didn't work. It's actually worst as it hung my system when CC tried to access the "Serial" port.

 

[CF Mark]

Ok.
Well I'll ask around to see if there are any know problems with the compiler or anything like that.

 

[2cool4u]

Im running CC on a Dual AMD MP System. I haven't any Problem with freezes or anything. I run 0.95.

There must be something wrong with your System

 

[CF Mark]

Im running CC on a Dual AMD MP System. I haven't any Problem with freezes or anything. I run 0.95.

This is good to hear.
Using a Serial or HD44780 LCD?
Not that it should make a difference though.

 

[2cool4u]

It's a serial 4x20 from Crystalfontz

 

[cp10932]

Well ... the weird thing is that Wintest runs fine while CC just keeps hanging ...

 

[CF Mark]

CC is MUCH more complex than WinTest, just because wintest works, it doesnt mean your system is ok.

2cool4u has proved that CC does run fine on MP systems, so the problem shouldnt be caused by that.

I bet if you pulled one CPU out of that system, it would still hang.

The only thing i can recommend is that you reinstall windows.

 

[cp10932]

FIrst of all, nowhere did I say that my system is "ok". All I was saying is that CC freezes when I run it on my dual Xeon system. All I want is CC admin (or someone) to help me/us debug this problem.

Isn't it ironic that CC admin also jump to conclusion by declaring CC is fine simply because there is one known dual MP system (don't even know what OS runs on that, what mobo, etc) that runs CC fine? If this is how Crystalfontz define reliability, I am speechless.

The fact that CC is MUCH MORE complex (yeah I can at the very least guess that!!) than Wintest doesn't make it right to not be as reliable as Wintest. Or is it "More complex stuffs MUST break"??? CC Admin still has not adequately address why Wintest runs fine while CC does not. The fact that CC is more complex is NOT an adequate and satisfactory answer; to me and, I hope, Crystalfontz.

I can live with Crystalfontz telling me that they don't have the resource and exact configuration to debug this problem effectively and it may be a while before my problem can be resolved and they can't commit to a time line to fix it. Or something to that effect.

But jumping to a quick exit door on the earliest opportunity just isn't the right way to deal with problems like this.

 

[CF Mark]

There isn't a "right" way to deal with a problem like this.

The fact is that now we know that CC does run on a MP system ok (thanks 2cool4u) so its not strictly an MP problem. And I have not heard of CC having this problem before on any machine.

To me this simply means that there a very good chance that there is something wrong with your PC which causing CC to crash.

As you can probably understand, I'm not going to sit here for hours looking for a problem with CC's code, when it may not exist.

This is why i recommended you try reinstalling windows.

If it still has the same problem, then i would look into it further.

 

[cp10932]

I give up. I think CC admin just does selective reading. Somehow CC admin just doesn't exercise the same type of attitude and approach to achieve high level of customer satisfaction that the front end people I dealt with at Crystalfontz do.

I REPEAT. I am NOT asking Crystalfontz (or in this case, CC admin) to do whatever it takes to solve my problem. READ IT AGAIN, PLEASE. And the declaration of "Oh there is ONE MP system that CC runs fine means CC is ok" just isn't a logical one. And are you going to address my questions in my previous post at all?

May be it's time to give Matrix Orbital a try .....

 

[CF Mark]

Ill admit im not the best at tech support (and never want to be).

But i am giving you tips on how to fix the problem... what more can i do?

 

[CF Tech]

Well ... the weird thing is that Wintest runs fine while CC just keeps hanging ...

The code that has trouble may be a part of CC that does not exist in WinTest . . . like the part that interfaces with MBM for example. So I do not know how relevant it is that WinTest runs OK.

"Dual Xeon" is probably quite a bit different than "Dual AMD MP". It is significant that CC does run on "Dual AMD MP" because it does show that dual processors alone is not the problem.

It is not the case that more complex programs should be less reliable. We hope that CC is a very reliable and bug-free program. However, as a machine gets more complex, there become more potential places for it to be broken. Generally a good beta testing cycle will flush most of those out, and I think that CC is quite reliable considering its complexity and that it is designed to run under all the current Windows OS versions.

It is true that we do not have a dual Xeon system to test with. Due to the apparently limited number of these systems that are running CC and experiencing trouble, we may well not pursue this problem aggressively. CC is a free program, which we produced with the main intention of making some LCD software that is easy to install, easy to use, and stable. We do not guarantee that it will be useful or even usable. Most people find it to be both however

Do you have access to a debugger? If so you could run CC under the debugger on your system and see where it falls down. CCadmin may then be able to track down the trouble. Some other things you could try that may help:

A) try to boot with the cleanest configuration possible, in case it is a conflict between some other program and CC. If CC is stable on a clean configuration, slowly add programs back in to see which one makes CC go South.

B) try to set CC up in the simplest form: like two screens that are text only, no announcements. If this is stable, then slowly add new screens until you find the one that gives CC grief.

C) Do you know of another dual Xeon system that you could try it on? That would help us to see if it is generic to all dual Xeon systems or specific to your particular configuration.

D) Use CC with the "Sim LCD" only (no real display). That would let us know if it is somehow related to the communications port.

 

[cp10932]

Thanks for your response. This is what I needed.

My system is actually a pretty *clean* Win2k install because I set this up no more than 6 weeks ago. Of course, it's not as *minimal* as would be preferred. A clean install of Win2k is probably not the first few things I will try. Also I don't have debugger ... so that's out of the question.

Your suggestions are very helpful and I will try as many of them as possible. Results will follow ....

Thanks.

 

[2cool4u]

Maybe I was to fast in saying no problem with crystal control in my system.

When I let it run for more than 5 hours without reboot, sometimes cc hangs, like discribed before.

this seems to happen since I upgraded to v0.96beta

 

[cp10932]

OK ... Finally had some time to try out a few things again. Here are my observations.

1. CC hangs much faster (in a matter of minutes) when I have SetiDriver running (i.e. utilizing ~100% of both CPU). This is very repeatable. I exited SetiDriver and run CC again. CC ran much longer (a few hours) before it hung.

2. There doesn't seem to be any particular screen that hung more often than others.

3. I set CC up in the simplest form with 0 screen and it hung when I was adding text.

4. I set CC up with just 1 screen; just text. Still hung like before. It didn't matter if it were stationary text or scrolling text or bouncing text. They all hung CC.

5. I set up CC with the "Sim LCD" only (no real display). Many screens were added and CC didn't hang after many hours of running in that mode; even with SetiDriver running at ~100% CPU utilization.

From what I observed so far, it seems likely that CC's communication routine is hung by something like handshaking timeout that worsen as the CPU is busy and not processing communication traffic quick enough.