PC2-6400 Ballistix (4x1GB) Trying for 417FSB- Help Please

[CreePinG_DeatH]

Hello,

I am trying to finally OC my old system with the goal of attaining a 417FSB, starting with determining the limits of the ram. The 417FSB would allow me to be at ddr2-1000 speeds, and thus hit 500FSB with a 2.0 multiplier. My system:
e6550 C2D
Gigabyte GA-P35-DS3R
4 x 1Gb Crucial Ballistix Tracers PC2-6400

I have dropped the multiplier to 6, and have my FSB currently at 390mhz (over 8 hours into p95 torture test) with memory multiplier at 2.4. Ram timings are currently at 5-5-5-15 @ 2.3v (I believe, I might have left it at 2.2v and I'll edit later if that's the case). I have not changed the NB voltage or vcore.

Assuming that everything I have reported is accurate, can I do anything more to further push my ram? Should I do anything differently? Thank you.

 

[Bobnova]

Is the ram getting warm?
Do the ram heatspreaders have clips on them?
Have you tried with just two sticks?

 

[Woomack]

I would try with higher NB voltage especially when you are running 4 memory sticks.
As Bobnova said check with 2 sticks on the same settings and if you make it run then bump NB voltage +0.1-0.2V for 4 sticks.
I would also set RAM:FSB 1:1 and try if FSB will work higher this way. Some boards are clocking better while 1:1 than on higher ratios. In this case I doubt that's memory fault.

 

[CreePinG_DeatH]

I re-read my post and found myself confused. Just to clarify:
I have a 390mhz FSB with the CPU multi @6 and MemMulti@ 2.4. So, my ram is currently running stable at ddr2-936, or 468 FSB, at 2.2v confirmed. My goal is to hit 500mhz with the ram, or at least come closer to it.

I don't have the time to try the steps advised as I am off to work. I will try later today when I come home and report back. Thanks again for your help, Bob and Woomack.

 

[CreePinG_DeatH]

Just a quick update:

I'm trying two sticks now in P95@400mhz FSB/DDR-960 (only at the 10-minute mark so far). When I removed the other two sticks, they were quite warm to the touch. I've never touched freshly used ram before, so I have no frame of reference.

These tracers do have little metal clips on the heat spreaders, small clips that easily come off and don't really to do anything. I'll update later.

 

[Bobnova]

I strongly recommend putting a fan on that ram.
The good news is that it's D9 chips in there. They should do 1200-5-5-5 without issue if you give 'em enough volts and keep 'em cool.
The bad news is D9s really hate being warm to hot.

 

[CreePinG_DeatH]

Interesting, do you recommend and specific coolers? Or just any fan blowing down on them? I ordered 2 x 120mm gentle typhoons to improve the case air flow. Browsing the current selection of replacement heatsink/fan combos didn't leave much to consider since the ram sticks are so close together as it is. I was wondering if it would be more effective to purchase 2 x 2gb sticks of ddr2-1066 just do give the sticks some space.

Edit: p95 still going

 

[CreePinG_DeatH]

Well after 4+ hours of Prime95, I bumped the fsb to 410, so testing now at DDR-984. How can I tell if this is an issue of heat, using all four slots, or both? I am probably jumping the gun, but I guess I can order something like THIS to see if it makes any difference with using all four slots. Adding more voltage to the memory to hit this speed is only going to make them hotter.

Are there better coolers to use? TIA

 

[Bobnova]

If bumping NB voltage helps it, it's running all four slots.
If adding a fan helps it, it was heat

I would expect that ram to do 1100-5-5-5 at 2.3v with ease. Odds are it's the NB.

 

[CreePinG_DeatH]

Well dang it, I've got prime 95 running with all four sticks but the side of the case is off and haven't changed the NB voltage yet. I'll let this run all night and see how long it lasts. I don't know how I am going to cool the ram as I have a TRUE with 25mm fans on each side blowing towards the back. I don't know if a ram cooler will clear the front of the fans...

 

[Bobnova]

I like AMD heatsink fans, the non-heatpipe heatsink fans.
They're 50mm or 60mm x 12mm. I'd put one at the top of the ram blowing down along it.
Alternatively you can take the push fan off the TRUE and see how much of a difference in temp it makes, it may only be a degree or two for the CPU.

 

[CreePinG_DeatH]

Thanks for the feedback; I'll keep that in mind and check out the AMD hsf's.

Just an update: P95 lasted about 6 hours last night then conked out. This morning I restarted and upped the NB voltage by 0.1v and ran the tests again- only lasted about 6 minutes. I guess the next step would be to reduce NB voltage and try upping the ddr voltage, but i'm not optimistic about that. Who knows, I may be surprised and will report back.

 

[CreePinG_DeatH]

Well I DID actually surprise myself, and then felt a bit stupid. The DDR failed with 2.3v after 5 minutes, both sticks. I dropped the FSB to 415 (ddr2-996) kept the same voltage, same result. I then figured what the heck and lowered the voltage. I dropped it 0.2v and tested the ram at 2.0v. Well, after 10 minutes it hadn't failed, so I figured that I should see how low I can go. I am currently testing 1.9v and I'll update.

While I do think that this is a heat issue, I don't feel that I cooked my ram yet. I will most likely see how far I can push my cpu before I order cooling for my ram.

 

[CreePinG_DeatH]

After 3.5 hours and still going without a hiccup, I'm trying 417 FSB for the magic DDR2-1000... It's odd; the further I push this ram at the lower voltage, the more inept I feel. Quite humbling

 

[CreePinG_DeatH]

Ok, I got up to 430 FSB (ddr-1032) before p95 failed immediately. I upped the NB voltage, then the tests lasted 5 minutes. I upped the ddr voltage and the tests lasted 7 minutes. I lowered the FSB to 420 and the NB back to normal, DDR to 1.9v and I'm going to test this speed for several hours, perhaps overnight. Next, I'll test the CPU while I contemplate how to cool the ram.

 

[CreePinG_DeatH]

Well at that setting, the ram lasts 2-3 hours then the worker stops. I guess I'll rig a way to get some airflow on these chips. Because of lasting a few hours and then stopping, heat would be the logical guess, correct?

 

[Bobnova]

It's certainly not helping.

 

[CreePinG_DeatH]

Well, the ram is stable at 960 (400*2.4) for 24 hours, so I think I'll test my CPU now until I receive the active cooler.

Edit: still at 1.9v