Thanks Capt Caveman.
I missed that option in the flash utility.
Anyways, the system was not any more stable with the beta 2 version.
I did 2 things. I ran Memtest86 for about 6 hours on the default setting (I did not feel like figuring out the details of the advanced settings, but if anyone has a suggestion of an important option, please tell me). Anyways it did not find any erros.
Then I kept running the Microsoft Developers debugger, to see what system files were causing the crashes. And they seemed to be a mixture of 5 different files being the source. But still they were 0x0000001D and Ox0000008E errors. So that means it could be the software or it could be the hardware that was corrupting the kernel memory space the software utilizes. But to have 5 different sources of kernel software causing these errors...........seems like hardware really could be the problem.
But finally a 0x0000007F error happened. I caught it out the corner of my eye before the system did an autopower restart. And no minidump took place from the errror. I could not read fast enought to get the sub error -- but that is a detail. So it is a hardware issue only with the 7F error -- Not a software error.
So I downloaded some nice version of a Prime program SP2004 and ran it. And with no overclocking at all, at pc3200 speed 1T setting dual channel (2 sticks), it it kept stopping with its rounding error report alert, after either 10 minutes or 1hours (ran it many times and that was the range) on its blend test that stresses the CPU and the Memory. I then backed the system down to 2T memory setting and it went for like 5 hours and I turned it off.
So I used Everest to look up the exact timings the memory reports for running pc2700 speed instead of pc3200. And , I manually entered those into the bios, but moved the setting back to 1T. And the system worked dual channel at 1t pc2700.
So, now I am looking at each stick of memory. I pulled one of them out, and am running a single one in the slot the manual specifies (3rd slot away from the processor). (When I was running 2 sticks I had the 1st one nearest the processor, and the other in the 3rd slot away from the processor, just like the manual specifies). The 1st stick failed to run at pc3200 1T . It gave a system freeze during the screen saver, after about 20 minutes. And a high level program error, and the SP2004 stops after about 10 minutes with rounding errors with the program option selected to stress the memory rather than the processor ("Type: Large, in-place FFTs - stress RAM"). I did not test it at pc2700 1t or pc3200 2t as I assume it works , as it the system could work in dual channel this way.
So I pulled that stick out and have put the other in. And neither the kernel system, the high level programs, nor the SP2004 prime program set to stress the memory have failed in 80 minutes. So it looks like the 2nd stick of 1GB memory is more stable than the 1st at pc3200 1T. I will report back here if that stick fails too after a longer time period of running the SP2004.
_____________
_____________
So here is my question.
Assuming that one stick of memory passes the SP2004 test at pc3200 1T for a long period of time, while the other does not, (and they both are run in the same slot of the motherboard), does that really mean that the stick that failed is a subpar stick? All that I see that I can conclude is that it is more sensitive to whatever information the motherboard is delivering it from the CPU's memory controller. Could there be a problem with the signal quality or strength coming from the CPU? Could there be a problem with the motherboard keeping that signal clean? So I guess what I mean is that, does the stick of memory that failed , well did it fail because the signal quality getting to it is not up to par, or that it itself is not up to par to deal with the 1T command rate at pc3200? Doesn't the 1T command rate have to do a lot with the CPU's onboard memory controller?
Finally, how do these prime like programs actually work differently on the memory that the Memtest86 program. I don't get it.
Thanks.
PS. By the way , once I popped each stick of memory out , and put each back in (for dual channel operation), and the system failed to boot , giving me a memory test failed. I know for sure. that each stick was seated all the way down, with the locks over. I had to pop them out and push them back in, and the system booted up fine then. So the motherboard seems really sensitive to installing the memory.
And of course I take anti-static grounding precautions!