PDA

View Full Version : 3200+ Won't O/C


Riptide
12-11-04, 12:05 PM
Well trying to overclock this 3200+ might need to be considered a failed experiment perhaps. Can anyone think of something I might be doing wrong here?

I can't get the system to pass a SuperPi run at even 225x10. HTT is set to x4. I've tried increasing vcore all the way up to 1.6v from stock and it still fails even at 225x10. I don't see how it's the memory because it passed SuperPi at 250 just fine with my FX.

What in the hell?? Did I get another crappy overclocker here? Because if I did that is two CPUs in a row that didn't overclock well for me including a .09 which is supposedly very easy to O/C over stock. This is kinda frustrating...

I'm going to reapply the arctic silver again just to be sure the fact the heatsink seems barely warm isn't indicative of a problem there. But beyond that I really don't know what to do.

OCForums_Osirus
12-11-04, 12:18 PM
What board is this on?
Also, what kind of settings are you using?
like divider, ram volts, agp locks etc...

Riptide
12-11-04, 12:25 PM
Abit AV8 - the same one in the sig that has worked fine @ 250x10 w/my FX53.

2.5-3-3-8 timings (same as before) on the memory, vDIMM 2.7 or 2.8 (makes no difference)

The lock is engaged. I'm not using a RAM divider - shouldn't need to as it was OK @ 1:1 250 before.

245x9 and 225x10 fail SuperPi. I've tried raising vcore incrementally from 1.4-1.6v to no effect. HT 4x. SuperPi still fails.

I'm at a loss here.

Riptide
12-11-04, 12:44 PM
OK, I don't understand this now at all. :confused:

225x6 failed SuperPi. I left vcore @ default. I am confused at this point how this can be happening. The system passed SuperPi just fine @ 250 1:1 before... And this is the exact same set of DIMMs. What in the hell??

Just for the hell of it I did a stock SuperPi run (200x10). Passed just fine, even w/the RAM at 2,2,2,8. I am just flabbergasted that this memory would be fine with my FX but then freak out with this CPU. Why in God's name would it do that?

PaiN
12-11-04, 08:08 PM
Alright.....That's two bad A64 OCing experiences back to back....time to throw in the towel :surrender
Don't forget, you promised to return the 3200+ if it didn't OC well..
I'd try one more CPU...and if its a no go, then you can maybe blame it on something your doing wrong.

strk
12-11-04, 09:33 PM
Damn that sucks bro i have mine at 2.5 prime stable 11x228 with ram timings of 2.5-3-3-7-1T and this is my first overclock with any cpu.

Riptide
12-12-04, 02:33 AM
I just filled out the RMA request w/newegg so I'll send it back and eat the restocking fee.

I'm careful to take care of all the usual suspects like setting HTT back to 4x or 3x depending on HTT speed. The memory I'm using is known good @ 250 1:1 w/my FX. Yet this CPU won't even do 225 1:1? I tried vcore incrementally from stock up to 1.6 and it didn't help. Memory was set at the same timings/voltage it ran with the FX @ (2.5-3-3-8, 2.7v). So if it's something I'm doing... Noone has been able to tell me what that could possibly be. I'm covering all the bases and yet it doesn't matter - I just can't seem to get stable overclocks.

I think it's just a whole hell of a lot of bad luck. I understand people thinking I must be doing something wrong, but if so then what?? :confused:

Oh, and the FX53 is back in and passing SuperPi just fine at the same 250x10 1:1 settings I had it at previously. The 3200+ would not run 1:1 at even 225 and it's the same DIMMs and motherboard. That to me says CPU or mem. controller on the CPU is the problem, especially when you consider that I tried increasing vcore and that did not fix it.

PaiN
12-12-04, 07:55 AM
Rip..I've watched your OCing "adventures" closely :) ...and I have to agree I don't think its something you're doing. I believe, its unwilling CPUs, so keep plugging your luck is bound to turn :cool:

Riptide
12-12-04, 09:28 AM
I must admit there have been times I've wondered if I'm doing something wrong... But given the exchanges I've had in various forums, the reading I've done, etc. I just can't see how that could be it. So I have to chalk it up to bad luck and getting CPUs that won't cooperate. At least my FX53 gets up to 2.5 and it blows that 3200+ away in 3dmark01.

Oh well.. nothing ventured nothing gained right? I might give it another shot but I'm going to wait for the financial dust to settle from christmas now. This little project cost me $70 in restock fees and shipping. :)

Lfctony
12-12-04, 10:03 AM
Did you try dropping the HT multiplier to 3? Mine works fine at 4x220 but fails at 4x225.

Riptide
12-12-04, 10:10 AM
4x should be fine at 225. 3x should only be necessary at over 250. So if it fails at 225 for you I'd have to wonder what's going on there.

testing12342
12-18-04, 01:12 PM
is this any help:
http://www.theinquirer.net/?article=20262

"AMD says there's no problem with Winchester chips

Overclocking attempts come to naught

By INQUIRER staff: Thursday 16 December 2004, 11:07

CHIP FIRM AMD has denied a report on a forum that that it's got a problem with its 90 nanometre 3500+ "Winchester core" microprocessors.

The reports surfaced on this this forum, but AMD has told us there isn't a problem at all.

An AMD representative said that the firm's techies say that because the Athlon 64 chips are locked and cannot be overclocked, an attempt to increase the mobo's system clock is causing all the components to respond, rather than just the microprocessor.

He said: "A small increase in the signal clock speed - say 5MHz on the 200MHz signal clock - is probably within the tolerances of all of the components".

But, he continued: "These overclockers are trying to get all teh components to respond to a 25 per cent increase in the speed of the signal clock, which would cause multiple components on the board to become unstable".

FX Athlon 64 products aren't locked, and allow a fair amount of overclocking. But AMD doesn't endorse overclocking its chips. It's at your own risk. It will be happy to sell you another chip if it goes pfit. "

saturnotaku
12-18-04, 02:38 PM
That's all very well and good, but Rip had a 3200+ not the 3500 as indicated in that report.

testing12342
12-18-04, 03:06 PM
That's all very well and good, but Rip had a 3200+ not the 3500 as indicated in that report.

Which means nothing as this applies to all winchester core processors, it's possible he has a 90nm CPU.

Riptide
12-18-04, 04:17 PM
That doesn't seem to make a lot of sense. I would expect nobody would be overclocking winchesters if this were true. But some people seem to have no problems w/it.

Soylent
12-18-04, 07:09 PM
There's no guarantee that RAM that works at <some clock rate> in one motherboard will allways work at the same speed with another(processor in this case as the memory controller is on the CPU die). It usually does but memory controllers can be finicky about your particular sticks of RAM for no obvious reason.