Confusing memory O/C problem

[AirGibson]

Here's what I've got:

P4 2.4c
Abit IS7
512 PC 2700 ram (1 stick of generic, one stick of Mushkin)

I only use one of the sticks at a time for testing.

I'm noticing that when I run the processor at default 2.4 and set the generic memory to run at 166 with totally relaxed timings, it all runs fine. When I start O/Cing the processor, but keep the memory at 166, the system fails Prime95.

When I put in the mushkin memory, prime 95 passes. I don't get it. I'm not pushing the generic memory any harder, yet it's failing when the processor is O/Ced. Any ideas as to why that happens when I'm not pushing the memory any harder?



Second question is that even with the Mushkin 2700 memory, I can only get to around 2.8 GHz stable. When I go up to 3 GHz with the same settings for memory (totally relaxed), I again start getting prime 95 errors after the O/C. Based on my previous experiment where the memory was causing errors even though it was running at the same speed & settings, I'm wondering if buying some PC 3500 memory is going to allow me to o/c better or if it's more likely that I've hit the ceiling for my processor.

 

[Novaric]

i dunno about your first question. I'm on an IS7 with 1 GB of Geil value 3200 memory anda 2.8c. So far i've only gone up to 220 FSB in 1/1 divider. That = RAM running at 440mhz and CPU at 3.08 ghz. I think with a 600 mhz overclock you are probably reaching your CPU limit, but its probable that better ram would let you squeeze out another 1-200 mhz.

My ram seems to be holding me back as well. So far i'm planning to buy 512mb of Corsair and see if i can't OC higher. Hopefully it won't be a waste of 40 bucks and a downgrade

 

[AirGibson]

One bump for good measure.

 

[mroleg]

Also, not all RAM will work in async mode stable. This is what I think your issue is. Generic ram doesnt like running async with cpu.

 

[jjyiz28]

Originally posted by: AirGibson
Here's what I've got:

P4 2.4c
Abit IS7
512 PC 2700 ram (1 stick of generic, one stick of Mushkin)

I only use one of the sticks at a time for testing.

I'm noticing that when I run the processor at default 2.4 and set the generic memory to run at 166 with totally relaxed timings, it all runs fine. When I start O/Cing the processor, but keep the memory at 166, the system fails Prime95.

When I put in the mushkin memory, prime 95 passes. I don't get it. I'm not pushing the generic memory any harder, yet it's failing when the processor is O/Ced. Any ideas as to why that happens when I'm not pushing the memory any harder?



Second question is that even with the Mushkin 2700 memory, I can only get to around 2.8 GHz stable. When I go up to 3 GHz with the same settings for memory (totally relaxed), I again start getting prime 95 errors after the O/C. Based on my previous experiment where the memory was causing errors even though it was running at the same speed & settings, I'm wondering if buying some PC 3500 memory is going to allow me to o/c better or if it's more likely that I've hit the ceiling for my processor.

so you're saying , lets say you overclock your fsb to 250, you get prime95 errors using your generic memory running at 166. and then you overclock again to 250, but this time only using muskin memory running at 166, you DONT get prime95 errors??
if thats true, then did you try testing prime95 to your generic memory when you are not overclocking? maybe your generic ram is just bad?