e6300 poor overclocking

Lupus

Lupus

Lupus

Associate
Joined
13 Aug 2003
Posts
100
Seems I got a lemon :(

I put it into a DS4 with the matched pair of 1Gb Geil PC6400 with a Scythe Ninja. Initially it was about 52 C idle. So I took the heatsink off and redid the paste (I'd initially put the paste on the cpu, not heatsink, when I removed the sink there was only a little paste on it ie. poor contact).

Anyway, afterwards the idle temp was about 34 C with load (prime95 running on each core) about 48 C.

I just started overclocking, following the Gigabyte board guide on this forum, but so far the results are awful. Upping the vcore to between 1.4 and 1.45 ends up with a load temp of 60-65 C. The computer has crashed on the higher clocks I've tried, such as 400x7, 350x7, 400x6.

Does anyone have any advice please? I was really hoping to get some decent overclocking from this (like almost everyone else has managed), but I'm not sure what to try now.

I bought all these components (and more) on Monday. Do you think I might be able to exchange it for another even though it's not broken?

I'd appreciate any advice (relevant to this issue; I already know about avoiding yellow snow) :)
 
Cheers.

Just tried those suggestions(there is no option about hyperthreading for me). I already updated the bios to F6 and installed the latest chipset drivers, which I forgot to mention before.

I've got it running at 340x7 (2380 MHz). I had to up the vcore a bit to get it to boot to windows with this (1.35v).
To get it to boot to windows at 350 required upping the vcore a few notches more, but wasn't stable, so I moved back down. Also tried going up to 400x6, but couldn't get it to boot to windows.

Not sure what else to try now, was really hoping to hit 2.8GHz (at the least; was aiming for 3GHz+) :(

Any thoughts?
 
I've set the ram system multiplier to 2 and the ddr2 overvoltage to +0.3v. Set the MCH and FSB overvoltages to 0.1V (also tried OCing before upping them).

That's all I've done to change the ram settings.

Both cores are pretty much the same temperature, with a 1 C degree difference at most, both idling about 32 C stock idle and 43 C stock load. The heatsink seems to be on properly as I've tried pushing it down a bit while watching the temperatures and there was no change. Also tried it without the side of the case on and there was no noticable difference.

I set the bios to load optimized defaults before I'd started OCing again. Using a Lian Li PC7 with the 120mm fans that came with it (one in one out). The airflow is fine I think. Got the 120mm Ninja fan blowing through the heatsink towards the fan on the back of the case.

Each time I go to the bios, all the settings seem to have been saved correctly, couldn't spot any discrepancies. I'll try a CMOS reset and start again, hopefully that will help.
 
Just reset the CMOS (left the jumper on for over 10 mins) and loaded optimized defaults. Booted up, all was ok.

Then disabled the stuff previously mentioned, upped the voltages as said and set the ram timing to 4-4-4-12 with multiplier of 2. And changed the settings to those suggested in this thread.

At stock voltage the maximum I can get is 345MHz FBS (which I've got it at now). Going to 350 causes the comp to either not boot into windows, or to reset soon after. Upping the vcore a notch didn't help with that.

The code on the box the cpu came in (OEM) is *CP-135-IN

Been awhile since I've overclocked (barton 2500), but I had good results on my last system. So I'm not totally new to it :)

I'm open to other ideas.
 
Good call pieman.

I set the timings to 5-5-5-12 and ran memtest86, letting it run through all the tests, which resulted in 3 errors. Then removed one stick, tried 350 MHz and it rebooted again, so took that stick out. After gradually working my way up, I had it boot into windows at 490 MHz, so a bit of an improvement there :). But I've got it at 430x7 at the moment.

Unfortunately though, it seems the heatsink isn't sitting as snuggly as it was: the idle temps are now about 47-50 (vcore is still at stock) and pushing down on the heatsink will drop it a few degrees. Guess I'll have to reseat it again (sigh), which is difficult due to the size of the thing. Probably just take the whole mobo out again.

So it seems it was one stick of ram that was the problem, I'll see if I can get that exchanged.
 
Just put everything back in. I suspect I just pushed the clips in last time till they clicked, and didn't turn them to lock afterwards. My idle temp is now about 39 C and load is about 54 C. Still to do longer stability tests, but I'm feeling better about things now. If it proves stable here I'll see if I can push it further, since I've got more room experiment now.

Thanks for the help all, much appreciated :)
 
It's the Scythe Ninja. Turning them puts part of the black bit under a ridge on the white part, so it can't move upwards.
 
It could just be that the two parts weren't lined up properly for me. I guess when it clicks it's the black part slipping over the white ridge, locking it. I'd had them still twisted from when I removed it, so just turned them back to make sure they were locked under the ridge.
 
Got the RMA stuff sorted for the RAM, which I'll post tomorow. While that's away, would it be possible to use my old PC3200 ram from my last computer, if I clock the cpu back down a bit?#

It's just so I can use my comp while the Geil is away.
 
Update:

New ram is back, got the comp running stable at 3.22 GHz, 1.35 vcore 5-5-5-12 ram timing.

It's a week 29 processor. It can go slightly higher, but not much before second core starts producing errors in prime95. But I'm happy with what I've got :)
 
You must log in or register to reply here.
Back
Top Bottom