Help with new memory

[Matt2]

Ok, so I finally went 2x1GB Patriot PC3200 LL.

Timings are stock @ 2-3-2-5

This memory will not OC at all at the stock timings.

IN fact @ 289 HTT using the 7/10 divider yields 200mhz RAM speed on the dot according to Everest, but the system is unstable at DD400 2-3-2-7!!!

It memtested fine for about 3 hours this morning, but I get random reboots. These reboots happen while the system is starting up, when windows first boots up, about half way through 3dmark '05 and when I start the P95 torture test.

Right now the system is running absolutely fine in 3dmark, P95 and HL2 stress test @ 216mhz 2.5-3-3-7

Can these sticks just not handle the tight timings?? Even though that's what I bought them for and what their SPD is???

ANY enlightening thoughts would be appreciated.

Thanks.

 

[NesuD]

My first thought would be to not exclude something other than the memory. If everything works stable at the stock settings but then gets random reboots when you overclock the HTT bus I wouldn't jump to the conclusion that the ram is the problem especially since it is runnign at it's rated settings with the divider you are using. at those settings the ram isn't over clocked but the HTT bus is as well as the cpu unless you set the cpu to a lower multiplier. you need to isolate the major areas that are being overclocked one at a time and test them for stability. In this case I would first reduce the cpu multiplier low enough to get it as close as possible to the chips rated clock speed. If the instability goes away then I would lean toward the cpu overclock as the root of the problem, If the instability is still there then I would be looking at the motherboards ability to run that HTT bus setting stable. One more thing make sure you have lowered the Hyper Transport Multiplier setting down to 4x or probably even better 3x. If it is still at 5x then the HTT is actually running at over 1400 and it is only speced for 1000. You may have done some or all of these things already so I guessed at some things to look at just in case you didn't.

 

[Matt2]

Originally posted by: NesuD
My first thought would be to not exclude something other than the memory. If everything works stable at the stock settings but then gets random reboots when you overclock the HTT bus I wouldn't jump to the conclusion that the ram is the problem especially since it is runnign at it's rated settings with the divider you are using. at those settings the ram isn't over clocked but the HTT bus is as well as the cpu unless you set the cpu to a lower multiplier. you need to isolate the major areas that are being overclocked one at a time and test them for stability. In this case I would first reduce the cpu multiplier low enough to get it as close as possible to the chips rated clock speed. If the instability goes away then I would lean toward the cpu overclock as the root of the problem, If the instability is still there then I would be looking at the motherboards ability to run that HTT bus setting stable. One more thing make sure you have lowered the Hyper Transport Multiplier setting down to 4x or probably even better 3x. If it is still at 5x then the HTT is actually running at over 1400 and it is only speced for 1000. You may have done some or all of these things already so I guessed at some things to look at just in case you didn't.

I've been running the system in my sig for about 3 months now. 100% stable. This instability just started last night when I popped the new RAM in.

 

[Matt2]

Also, when I loosen the timings to 2.5-3-3-7 and raise the divider to 3/4 which gives me 217mhz, the system runs fine. Video Stress Test, 3DMark run fine and Prime has been running for about an hour now no problem??

From what I've seen it seems like the RAM cannot handle CL2, but that just doesnt make sense.

 

[imported_Kensai]

More voltage to the Patriot?

 

[Matt2]

Well my DFI board already overvolts the RAM by about .1, @ 2.6v the BIOS registers 2.7 plus i enabled the booster.

But I shouldnt have to, these sticks should run CL2 at their rated speeds without a voltage bump.