Trying to find the weak link on my FX-57 OC

[damelon]

Basic components are FX-57, ASUS A8N-SLI Deluxe and Crucial Ballistix 2x1gb ram,
Dual SLI GTX cards, raptors at raid 0, all Watercooled

4x HyperTransport
220 x 14 = 3.08 ghz
ram at 3-3-3-8 1T but can be tightened at this speed

CPU Voltage at 1.5125, Ram at 2.8V
PCIE bus locked at 100, Synchro locked at 33mhz

The thing is, my temps are still just above 40oC on the processor at full load and the bus is usually stuck at 30oC. So I have plenty of room still on temps. There is still room to work on voltages too.

I plugged my SATA raid on ports 3&4 on the NVRaid controller and have two high end power supplies pushing the system components (PC P&C, Enermax)

So why can't I get a stable prime if I get the speed up any higher? I have read about some issues with my particular MB, and I would guess that is the issue (the 250 bus limit, the possible max-out limit on the processor voltage, etc) but I would think I could lower the HT to 3x, lower the multiplier and bump up the bus speed, but it always fails prime when I get it passed 3.1ghz. Is this more of a chip barrier, mb barrier, or am I just missing something. This is the first system I've tried to overclock so I may be ignorant of a few things, but I've read enough threads here to think I should be able to push this further.

 

[GuitarDaddy]

Definately a chip barrier, your not going to get any higher without phase change cooling.

I've been using the same mobo for over a year, and it will hit a bus speed of 320 no problem. The 250 limit is for ram speed at 1t, you can go higher with ram but you have to use 2t

 

[damelon]

Makes sense. I'll just tighten up my ram a bit then, sit back, and enjoy.

 

[GuitarDaddy]

That's a very nice overclock, and a monster rig by the way.

That beast should plow through most anything you throw at it, like a chainsaw through zombies

 

[secretanchitman]

wow that is an awesome OC and a rig. i think your FX-57 has hit the limit on watercooling...i agree with guitardaddy, go with phase change if you want more OC.

 

[VooDooAddict]

Originally posted by: secretanchitman
wow that is an awesome OC and a rig. i think your FX-57 has hit the limit on watercooling...i agree with guitardaddy, go with phase change if you want more OC.

(I'm curious ... not debating.)

Why would he get a better overclock on Phase change cooling? His temps look pretty good as is.

 

[theMan]

because he could set the voltage to like 1.7v or more, and it would still be below zero temps. that would make a higher oc.

try upping the voltage to 1.55v or more, especially with water cooling/

 

[damelon]

Yeah I could throw more voltage at it, but I remember reading somewhere the A8NSLI had issues above 1.5125. Could have been just one guy's problems though. I will try out a higher voltage. I had the same questions about phase cooling.. since my temps were still pretty low I thought I still had a lot of headroom to push it higher. I will let you know how it turns out.

I just posted some photos of the rig with more details.
http://damelon.photosite.com

 

[GuitarDaddy]

I haven't had any problems with voltage on the board up to 1.55v (the max on this board)

In my experience chips seem to have a sweet spot for voltage, and once you reach that sweet spot more voltage doesn't seem to help. And the odd thing about subzero cooling is that it seems to substantially increase a chips overclockability even at the same voltage.

 

[secretanchitman]

sweet case!

didnt know the mobo had issues above 1.5125 vcore. dont go too high though. you might just burn the CPU along with it! but you do have watercooling, so you should be okay.

 

[python023]

Just keep your temps below 55-60C(load) and raise the voltage to your hearts content. Nice rig.

 

[damelon]

Thanks. It's very helpfull to have a double-wide case when it comes to setting up a watercooling system. Lots of room for radiators + extra fans. Only minus was that I had to take out the 3.5 rack on the right side to fit the double-fan mounted radiator. The case is one of the best things I've ever purchased. I'm going to re-use it for my future machines too, but hopefully I don't need one for a good while yet.

I tried a couple changes today but no luck. I can crank the voltage all the way up to 1.55V and still not break 50oC maxed, but prime still fails. I've tried several combinations (240,13 - 250,12.5) they didn't work out. I tried my initial setup with 14,221 under the same volts and prime ran for 2.5 hours... so even 1 more on the bus seems to mess it up. Maybe if I try something like 221-225,14 at 1.55 volts it will work.

 

[Unkno]

i'd say you reached the chip's limit, be happy and play some games....don't forgot about benchmarking and posting some scores!