Fast CPU cause error starting my program

[flippered]

Hi All

I have a new computer at work, AMD Dual Core 4800+, which is kinda fast... real fast... so fast that my program get confused and cause an error at start.
(I just get the boring windows message saying that my program will be closed, error at Offset: 00004192).

I used to have the same problem with my other machine, but I could bypass it by creating load on the CPU (like starting a complex task that requires 100% cpu), then once a while my program would start and run for the day(Stock market tracker).

With my new machine, nothing I can do...

Anyone experienced similar problems?

(Program was built on Delphi 7, compiled on AMD machine running XP SP2.)


thanks a million for any tips on this.

Dom

 

[TonHu]

Do you have all servicepacks of D7 installed?

HTH
TonHu

 

[flippered]

7.1 is the very latest?

if so, then I'm up-to-date.


thanks for the reply!


Dom

 

[TonHu]

In that case I'd look at the code that gets executed during startup. I expect that some database is beeing opened, and the sequence of things is probably time-critical. Maybe some extra checks whether each step is ready before proceeding could help?

I used to have the same problem with my other machine, but I could bypass it by ...

This should have warned you, back then... :~/

HTH
TonHu

 

[flippered]

LOL yah, fast cpu should not be an option

thanks for your time dude.

I Think I will attempt to recompile in Delphi 2005...but many components will go busted.


cheers!

Dom

 

[Glenn9999]

Curious: With this being such a faster setup, I wonder if there isn't a repeat of a similar error that caused the RTE 200 issues in Turbo Pascal 7? Is there any CPU cycle dependent code that is getting overflowed by the CPU's speed?

 

[flippered]

It always ran fine on my dual AMP 1800MP,

Would start only if I would overload the CPU on a Pentium 4 2.8Ghz.

And the AMD 4800+ X2... no can do. way too powerfull =)

I guess that Intel wouln't like to know that


And I hope it will not happend every 4 years :x

 

[whosrdaddy]

seems to me you are experiencing a race condition between multiple threads. are you using threads (or are there components that are using it?)?

-----------------------------------------------------
What You See Is What You Get
Never underestimate tha powah of tha google!

 

[flippered]

Maybe the AdvancedStringGrid component is, i'm really not sure, but there are so many things in that program that it would take a while to figure out.

i'm a trial'n'error programmer, I just drop component and fight until they work =)

 

[Glenn9999]

Yeah the next questions would be can you isolate the problem to a specific thing, and if so, can you produce a small program that duplicates the error.

Note on what RTE 200 was: When you used a specific Borland-provided unit, the code would run fine up to a certain speed (2000Mhz I believe?). Then when the code was run at anything equivalent or above, it would produce a Run Time Error 200. This occurred on any program that used the unit, because the CPU would cycle so fast that it would overflow a variable used by a routine within this unit. Long story short, a small internet industry cropped up on providing fixes for this error.

It wouldn't surprise me if it reoccurred.

 

[Glenn9999]

Whoops, extra 0. Should be 200Mhz.