I have a box with an AMD Athlon 64 X2 dual core 3800+ processor with 2GB memory with Award Modular BIOS V6.00PG. NVIDIA is installed. I am running Windows XP SP2

I am running an application that querries information from a USB dongle through a USB device driver. The application performs eight consecutive calls to the device driver in each loop of my code. The program is set to perform continuous looping in order to try forcing the 'hanging' event. I can run the identical application on an Intel dual-core 64-bit processor for a week (tens of millions of consecutive tests without incident). However, the program periodically hangs when I run it on the Athlon processor.

If I run the program on the Athon with maximum USB device activity the hanging event occurs very seldom. If I add a one second delay at the end of each loop in my program the hanging event occurs every hour on average. If I replace the one second delay with very brief delays of 1-20 milliseconds after each call to the device driver (i.e. eight times per loop) the hanging event will occur within a few minutes.

There are no programmed delays in the device driver (no calls to KeStallExecutionProcessor), although there are wait events.

The device driver does not have any variables declared as static.

The AMD Dual Core Optimizer utility and current driver for the processor from AMD are installed (V 1.3.2.0 5/27/06).

I have tried altering the power management settings of powercfg.cpl, without any change to the behavior.

With SoftIce installed the debugger does not trap on any problematic event when the system hangs. After the hang, stop of the system from operating by SoftIce always traps at a call to HAL!KeQueryPerformanceCounter.

The activity occuring within the device driver immediately prior to the hanging event are the issuance of a call to IoCallDriver, followed by a KeWaitForSingleObject. The KeWait... was altered to change the call from 'wait forever' to 'wait for a maximum of 3 seconds', however the OS still does not return from this call.

It appears that the problem is somehow related to the handling of time delay events issued by the application, that perhaps are conflicting/colliding with other time related events issued by the OS or the device driver.

The processor is not actually hanging when the application/driver fails to respond. It is possible to start other unrelated applications without incident.

I would appreciate any suggestions for resolving this problem or at least identifying the source of the problem (i.e. a bug in the Athlon processor, conflicting device drivers, an issue within the USB driver, etc).

HuH /wacko.gif" style="vertical-align:middle" emoid=":wacko:" border="0" alt="wacko.gif" />

-------------------------
Amd 4400x2
asus a8n-sli premium
pc power 510
corsair twinx2048-3500LLpro
70gb wd 10,000x2
evga 7800gtx X-Fi Platinum

Hopefully, some competent soul will have some idea what the [content edited] you're talking about and help you out.

In the meantime...
If I understand correctly, both the optimizer and CPU (Cool 'n' Quiet) driver "mess" with timing at a fairly low level. Have you tried running without either? What other hardware is installed? Post the rest of the system specs to give that competent soul an idea of what else may be involved.

-------------------------
939Dual-SATA2 v1.05, BIOS v2.31
Opteron 165 (CCBWE 0550UPMW) @2.7GHz, 1.35V
TWINX2048-3200C2PT 2.5-3-3-6 1T
Thermaltake Big Typhoon
ATI X1800GTO