EE - Have I reached my max oc?

Hi,

I have a 478 3.2ee @ 218 fsb stock voltage totally stable, dual prime stable for 30 hours etc etc. The thing is I can't get it stable at 219fsb even with a voltage increase, I've tried up to 1.65v and it still fails prime quite quickly.

I have set a divider on the RAM to eliminate that. I have a P4P800e-dlx.

Does anyone have any ideas? Have I just hit the limit of this chip?

Thanks

we are going to need to know more about the system to help out.

Motherboard:
CPU:
Ram: speed timings brand IC if known (Hynix, Samsung, Micron....)
PSU:
GFX: JIK

^^ like he said
if your on high or even water dont ever run a S478 pressy above 1.6v, it can/will kill that chip.

Hi,

The RAM is running with a divider as I said, so its not that, its corsair 3500ll anyway, I wrote what mobo I have as well, p4p800e-dlx, psu is enermax liberty 500w. gfx is a 7800gs.

Its not a prescott, its a gallatin chip.

Anyone got any ideas?

what about your temps what hsf or watercooling are you using?
yes the voltage still does matter, you never want to go over 1.6v even with the new lga775 p4's/p-d's.
is the pci/agp speed locked?
divider which one, just saying you are using one does not tell us which one.
we are not mind readers.

i have the LGA version of your chip the 3.4 gallatin chip and it is not much of an overclocker. Most of them only do about 3.6 on air, maybe 3.8 on good water with a heavy overvolt.

my temps are 22 idle and 33 load.

The stock voltage is 1.55, so I don't think 1.6 is that bad.

Thanks for the info about your 3.4, sounds like 218 is the max of my chip, I just thought it wierd that I hit a wall there with stock voltage and 219 just isn't possible even with increased volts.

well just because my mobo and chip suck dont give up it sounds like you have really good cooling on your chip because 33c is awsome. give it like 1.65-1.7v and see what it can do. You will want to drop your memory just to make sure its not the bottleneck.

I sometimes wonder if the temperature readings are correct! But I did have a prescott in there and had to work really hard to keep the temperatures in the safe zone. With the same setup, the gallatin runs nearly 30 degrees cooler than the Prescott!!

I have 2 80mm in fans at the front, 1 out at the back. 1 50cfm fan on the side blowing onto the ram/chipset/cpu and an xp-90 with a 50cfm fan. This was required to keep the prescott at 60 degrees under load!!

chipset temp never gets more than about 30.

I have dropped the memory (even though its totally fine at 218) just to eliminate it as a problem. I'll try near 1.7v just to see if that helps but I'd be a bit scared of running that 24/7 with the SND syndrome..

running 1.7v even for a little while, stressing the cpu can cause permant damamge. i warned you once about cpu voltage now you just asking for it.

ok but you also said that 1.6 is dangerous but it runs that stock....

I don't think you know much about the gallatin chip. Also I said near 1.7, not 1.7 or over.

1.6v is stock voltage for a few Gallatin CPU's. My particular one is 1.55v stock.
Not an easy CPU to overclock stable (easy enough to get a few benchs and thats it), they hate heat and like heaps of voltage just to move up anywhere, that in turn creates even more heat, so making it near impossible to get anywhere at all unless you have super cooling.
I am no expert, but if I were to guess, my money would be on the extra transistors it has to feed, 2MB of L3 cache increases the die size.
You may have reached it, every chip behaves slightly different.

yeah, I'm sure its that cache that is hampering the overclocks. What have your EE to?

Mine is also 1.55v stock, but cpuz reports it as being 1.6 most of the time and even a bit higher.

I wonder, as the stock voltage is that much higher than a 3.2 northwood, is going over 1.7v still risking Northwood sudden death? No way am I gunna be testing this though

Sounds like unless I get some phase change 218 is as high as I'll get. Not too bad, 3.2 -> 3.49.

It's 1.55v stock, but my mobo overvolts it to bang on 1.6v
I have throw every voltage from 1.1v using AI booster software up to the max where the chip will fail to even respond let alone post.
Around 1.8v it will not post. I have posted at 4GHz and seen the magic number on the screen for about 1 second. Have at times made it into windows between 3.8-3.9GHz. Have been able to run full benchmarks bang on 3.8GHz and play games briefly, though all in all it really is turely stable at 3.6GHz.
Lots of volts = overheat and crash (even with water). I found the CPU responds best at it's stock voltage upto its stable max speed where the true wall is 3.6GHz.
To be on the safe side so as not to have an system on the balance of its last usable Vcore that it can tap into just to stay operating I often run at 3.47GHz (217FSB).
Now and then I like to use AI booster and run low voltages. I have found the chip responds very well to low voltages at low freq, part in fact the temp barely moves. Can run 1.1v @ 2.8GHz

Right now as I am typing, I have it set to stock 3.2GHz (voltage to ~1.3v) and I am forcing the 4:5 ratio with some Mushkin 3500 LV2 @ DDR 500 timings 2-2-2-6.
It's running very nice so far. Super PI 512k = 17.125s (which is only one second slower than 3.6GHz 1:1)

me 218, you 217 at stock voltage. Sounds very much like the limit to me, thats all great info, thanks.

I was really interested to see you're running DDR500. I have tried desperately to get that working but I get artifacts in games and prime failures.

I got hold of some ocz 4800 (DDR600) because I thought the corsair stuff was causing my problems @ 273mhz (still on 218fsb) but the artifacts were worse!

I can't remember if I tried running at 200fsb and the 4:5 divider for DDR500..... if its a bit slower as you say, I wouldn't go for it, but sounds like there must be a crossover point somewhere around 210fsb...

Yeah I have 3.7v going through these 2x512MB pair rigth now. I want to hit 210FSB as you say, that will produce 3360MHz with DDR 525 at 2-2-2-6, which would be insane if I can make it. Might go for 3.9v in to these sticks.
I can tell ya, something is not a bit slower at all. For some reason the EE is loving the low CPU freq with the high memory freq and fast timings. My minimum frame rate is holding up well in games where they other wise drop. I don't have any numbers, though I can tell something has improved.

Normal: 3472MHz / 217FSB / 1:1 DDR 434

Now playing with 3200MHz / 200FSB / 4:5 DDR 500

Check here for a screen shot in my last post where I hit DDR 508:
http://www.ocforums.com/showthread.php?t=453038&page=2
Overclocking on a budget, there no way I could go over 250 with the CPU and I always wanted to go over 250 with the ram

did you do the basics like disable spread spectrum and lock the pci bus?

yeah I've done that what have you got yor 3.4ee to veryhumid?

microfire, the thing you said about low voltage is interesting. I've never loaded that AI program but I might just do that this evening and see if lowering the voltage helps me go higher, from everything I've seen with this chip so far (including yours and my experience) higher voltage = bad and unstable, maybe lower voltage is the way forward? We shall see.....

I might try the 500 divider again although the pain I went through trying to get it stable and then finding gfx artifacts in games (can't explain that at all, must be something to do with AGP) has possibly put me off doing that again!

I actually have never clocked it for any long period of time. Sometimes, I put the fsb up to 215, and it seemed to do that stable with stock volts, at least for the benchmarks I did.

There is another forum member "Electron Chaser" who also has a 3.4 and much more experience clocking it. He might have some advice for you.

I've heard the 3.4 can do 3.8-3.9 on air, but that's about it.

Nice to see another Saints fan on the site