• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Is my memory corrupting Win XP files?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

v8440

Member
Joined
Sep 5, 2001
Hi all,


I've just installed XP Home ed on a duron 950 machine, running a Shuttle AK31 Rev 3 board, and 256 MB of Kingston PC2700 DDR. Here's my situation: Last night I was attempting to overclock the system, and it seems that every time I bumped the FSB past 150, the system would hang or get some type of sporadic error. What surprises me is that I have successfully run 512 MB of DDR at 157 MHz on a AK31 Rev 2 board, using an Athlon 1.33 266 fsb CPU. One stick was regular Micron, and the other was Crucial. See my confusion? Two presumably mismatched sticks can run at 157 MHz on an older motherboard with an inferior northbridge chipset, but 1 stick of pc 2700 can't make it past 150 on a KT266A-based board. (on both setups, all memory settings were at their most aggressive except for the command rate-it was set to 2T in both instances. Additionally, on the duron/AK31 Rev 3 setup, the que depth (something like that) was set to 4. This setting is unavailable on my Rev 2 board) Some more system info: The graphics card on the duron/rev 3 setup is a radeon 8500. Surely that should tolerate fsb's well past 150? The only other difference between the two systems is a home pna network adapter card I installed on the duron system that didn't exist on the athlon/rev 2 system. Could this be causing data corruption?

Finally, the OS: Is XP tremendously more sensitive to memory errors (assuming that memory errors are in fact my problem) than other OS's? After one overclocking-induced crash, XP suddenly demanded that I activate the software. That installation was about 20 minutes old. I noticed that the system clock was way off then, so I'm assuming that the incorrect time/date is what made XP demand to be activated, but why would overclocking make the time/date settings change?

(in case anyone is wondering, the reason my XP is not activated is that I'm about to change all sorts of hardware in the computer, and I don't want to lock myself into a configuration that may not be the same in a few weeks.)
 
My friends memory went bad on his Asus A7V133 (it wouldn't post). We replaced the memory, but now WinXP won't finish booting even in safe mode. Strange stuff.. Maybe something else got fried. We're going to try starting him from scratch with Win2k.
 
The problem is when you OC over 150FSB your data tends to get scrambled, and show corrupted data on your HD. It isn't a Ram issue but rather a OC issue. Check out TOmsharware.com on your MOBO and see how well they rate it for OC. SOme boards can OC better and you just might have a bad OC. Or maybe even need a Updated Bios.
 
Usually if you don't have 1/4 dividers and your running at 150fsb that typically means your PCI bus is like 50 and HDD data corruption starts at 41pci bus.

So try turning it down to stock speeds and see if it happens.
 
Clarification needed

Ok, in response to some of the replies above, I should probably emphasize that I ran the FSB at 157 without problems under windows 98 with an AK31 rev 2 and two sticks of pc2100 Micron 256MB DDR. To me, this would indicate that there was no data corruption occuring because of the high FSB speed. Also, I note that some people are hitting FSB speeds as high as 166. Now, when I ran it at 157, that was with an athlon. Now I'm running a duron. I noticed that when I bump the FSB up to 150 and reboot, the DRAM speed is reported as 266 instead of 200 on the screen that flashes up during the boot process. I'm guessing that this part of the display changes from 200 to 266 when the FSB hits 133 or higher. Since a duron has a 200 mhz bus speed, could this be causing problems that an athlon wouldn't have?

Also, what is the difference between setting the DRAM clock to "host clk" and setting it to "by spd?"
 
hdd

Could be your hard drive. I had this happen once. Try switching from ATA100 to ATA66 in the bios if you can. Some components just can't handle overclocking.

P.S. Could be the ATA100 cable, too. Had similar problems with a non-oc'd system, replaced the cable, and all was A-OK.
 
<Could be your hard drive. I had this happen once. Try switching from ATA100 to ATA66 in the bios if you can. Some components just can't handle overclocking.

P.S. Could be the ATA100 cable, too. Had similar problems with a non-oc'd system, replaced the cable, and all was A-OK.>


I should re-emphasize that everything is just fine at 168 MHz fsb with only one stick of ram, and 165 with two sticks. Same hard drive. The drive only clicks off and on when high fsb's AND high chip speeds are involved. NOT when only one of the two situations exists. If I underclock the cpu, while running a high fsb, everything is fine. I ran over 221,000 iterator tests on desktop dyno 2000 last night with no probem, fsb is set at 165. But, if I then crank the cpu voltage up (even with it still underclocked) to add more of a power drain to the system, the hard drive will then begin to switch off and on as previously described. I can only assume that even a 500 W power supply can't keep up with running 2 pumps, 4 80 mm fans (including the power supply fan), the high fsb speeds, and a SEVERELY overvolted cpu.

Does anybody have an idea how much current the pumps that koolance uses draw?
 
Back