Overclocking an FX-55 with water

[Bobby74213]

Right now I'm just trying to find my limit on the this processor in my system. It's an FX-55 (San Diego) on a DFI NF4 Ultra D board. It's watercooled with swiftech parts. I think I recall getting it Prime95 stable at 2.95 GHz once but now that I'm attempting it again it doesn't seem to be working very well.

Here's what I've done.

Locked Multi at 13x
Locked FSB/LDT ratio at 3x (to prevent as possible cause of error)
Locked FSB/DRAM ratio at 1/02 (100 mhz) (to prevent as possible cause of error)

With these settings I test various voltages and HTTs.

1.62v at 228 (2964) failed in 2 minutes
1.62v at 227 (2951) failed in 2 hours 5 minutes
1.6v at 227 (2964) failed in 7 minutes
1.62v at 226 (2938) failed in 0 minutes

These results are really confusing me. If anyone has had any experience overlclocking an FX-55 San Diego, please share your results. Or if anyone has any tips or insight, I'd appreciate that too.

[brigby]

Re: Overclocking an FX-55 with water

I personally can't help you out with any advice, but I can give you some direction. In case you haven't already done so, check out the following sites:

www.dfi-street.com (DFI tech support forum...LOADS of technical know how)

https://www.extremeoverclocking.com

Those two sites are very helpful to the overclocker.

[Yawgm0th]

Re: Overclocking an FX-55 with water

How much voltage does the RAM have? What wattage is that power supply?

[Bobby74213]

Re: Overclocking an FX-55 with water

The ram is at 2.6v. The power supply is 510 Watts.

Also, there are a few options for the torture test in Prime95. I've been running the default one (blend). Is that the one I should be running?

[Bobby74213]

Re: Overclocking an FX-55 with water

I just ran Prime95 at all stock settings and it failed in less than a minute. Then I ran it again and it failed in 14 minutes. Does this mean there's something wrong with the hardware? I did a quick run through all the tests in memtest and it passed, no problem. I'm going to do some more extensive testing on it in a bit though.

[Yawgm0th]

Re: Overclocking an FX-55 with water

That RAM won't be stable at 2.6V regardless of what speeds you have it on. VX is notorious for needing lots of voltage just to run at stock. Try at least 2.8v and then run memtest. It's in my signature. Don't ever waste time in Prime 95 until memtest has cleared your RAM.

[Bobby74213]

Re: Overclocking an FX-55 with water

k. good thinking. I'll try that. But what do you mean pass memtest. I ran through all ten of their tests and it passed but it was just once through. I've heard you usually want to run test #5 a few times then run through the whole thing.

[Bobby74213]

Re: Overclocking an FX-55 with water

I'm really starting to wonder about this board... This is the second Lanparty I've owned and the first one found its way to the dumpster pretty quick.

With default settings (or even if I up the voltage to 2.9 on the RAM) it doesn't always boot. If I pull one powercord out of anything, say a cd-rom drive, or change one jumper, it hangs at detecting drives. I have to clear CMOS and powercycle it a few times til it comes back. It doesn't recognize my USB keyboard or mouse quite often and it won't give me the option to up my voltage past 3.2v even though I've changed the jumper.

I really haven't done much to this board. I let everything run stock for about 3 months and now that I try overclocking it I find I can't even make it to an hour in prime95 with stock settings.

Is it even worth troubleshooting at this point. Because it doesn't seem like a problem that's going to go away.

[Yawgm0th]

Re: Overclocking an FX-55 with water

Hard to say, but it looks like you could just be really unlucky. Might be time for another RMA...

[casecutter]

Re: Overclocking an FX-55 with water

Every DFI board I have ever had has given me trouble. That is not to say it wasn't user error also. Those boards are very tought to set up. I suggest going on OCZ's web site for some guidance.

What bios are you running? That is the single most problematic area with DFI boards. On the OCZ site, they have what is considered the best bios for that board and VX ram. I would suggest a bios flash before you give up on that board.

As for running VX at low voltages, I have had 3 sets and they all ran flawless @ 2.6V @ 2.5-3-3-7-T1. I also did extensive testing @ 2.85V and all 3 sets ran perfect @ 2-3-3-6-T1. Any tighter at all, 3.3V or more is needed. Don't forget to cool that ram at any Vdimm higher than 2.9V.

As for memtest, it needs to run error free for at least 12 hours before it can be considered stable. 24Hrs is considered 100% stable if it is error free. Prime stresses the entire system and needs to run the same times for a true test of stability. You most certainly have an issue. I would get a newer bios, then test each piece exactly the way you have been. You are using a great test method and if it just won't get stable, time to isolate. I cannot recommend enough to go to the OCZ Warranty support forum and make a post there. Do some searches too. They have done extensive testing with DFI boards and VX ram. Good luck.

[Bobby74213]

Re: Overclocking an FX-55 with water

Thanks for the tip. OCZ's forum had a lot of helpful info on there. Upping the voltage on the RAM seemed to make a difference. I just got home from work and I've had prime95 running stable for about 8 hours now with these settings:

1.63v, 224 x 13 = 2912, 48 C
LDT is at 2.5x with 1.26v
chipset is 1.57v at 46 C
RAM is 112 mhz, 2.5-3-2-5 @ 3.22v

I'm probably going to try a new BIOS, later and see what happens.

I'm a little apprehensive though. I mean if this can't run stable at stock settings, will it ever be stable?

Also, an interesting thing that happened before all this. I was burning some files to a CD and it locked up. This happens every now and then with my Lite-On burner. The only way to get out of it was to restart my computer. I did that and it wouldn't start. Now if I have to molex powercord plugged into the cdrom the computer won't even post. unplug it, starts right up.

I don't think it has anything to do with the motherboard though, seeing how something similar happened before with my first Lite-On burner in my old rig.

[Bobby74213]

Re: Overclocking an FX-55 with water

After some tips at the OCZ forum, I flashed with a new BIOS and got it stable at 2925 @1.63v.

I really don't want to take the voltage any higherso I may leave it here. My temp seems a little high though.

How often should you change the water in your system?

[casecutter]

Re: Overclocking an FX-55 with water

The water really only needs to be changed to prevent gunk buildup. Every 6-12 months is fine. As for heat, what is the temps you are worried about. If you are running the VX slower than CAS 2, it takes less juice. CAS 2 requires at least 3.3V to run right. I run my VX @ 3.3V 24/7 @ 2-2-2-6-T1. My CPU Heatsink/fan (XP-90 w/120MM fan) overlaps it nicely and the temps of the ram are never eally more than slightly warm. THey are usually cool to the touch. When I run benchmarks, I have upped the ram to 3.8-3.9V and not had issues. DO NOT go over 3.5V on my advice. I have run both sets of my VX that high with no effects, but I do have excellent case airflow and cooling. As long as it is cool, it doesn't care. If you are looking for ram that doesn't make the heat with timings like you are running, get some TCCD based ram. It will only need 2.75-2.85V to run at the speeds you are running now. Other than that, I am glad it is getting more stable.

As for the burner, stay away from LiteOn. They are crap. Get a NEC DVD burner or even a BENQ would be better than Liteon. If it won't post with it attached, it be dead. Junk it.

[Bobby74213]

Re: Overclocking an FX-55 with water

I think I've reached the max on the cpu with 2938.5 @ 1.63v. I'm about 8 hours deep in prime95 so it's lookin good. I'm about to test the RAM after this which is gonna be the tough part. I got two 20mm fans directly on top of the sticks. I'll probably change them to the yellow slots, too, cause I hear those ones are better for high voltage. What's your HTT on that RAM, casecutter?

[RenascentMisanthropy]

Re: Overclocking an FX-55 with water

I just ran Prime95 at all stock settings and it failed in less than a minute. Then I ran it again and it failed in 14 minutes. Does this mean there's something wrong with the hardware? I did a quick run through all the tests in memtest and it passed, no problem. I'm going to do some more extensive testing on it in a bit though.

(Taken from some Prime95 text)

FREQUENTLY ASKED QUESTIONS
--------------------------

Q) My machine is not overclocked. If I'm getting an error, then there must
be a bug in the program, right?

A) The torture test is comparing your machines results against
KNOWN CORRECT RESULTS. If your machine cannot generate correct
results, you have a hardware problem. HOWEVER, if you are failing
the torture test in the SAME SPOT with the SAME ERROR MESSAGE
every time, then ask for help at http://mersenneforum.org - it is
possible that a recent change to the torture test code may have
introduced a software bug.