The right board can make your quad fly

[Brolloks]

I just PM'd a good friend here with an experience I had today and thought I'd post it here so others could maybe benefit from it. I must say I was surprised...never too old to learn

I picked up a Q9950 up at Microcenter yesterday to do a review on LGA 775 quads vs P55 quad...the 1st one I got was a L845B batch and I read at XS after I got home that it is a crappy batch, luckily I did not open the box, so I drove 30 minutes back and pick up a new L847C batch, yes the "C's" are out. the VID is a nice low 1.2125 v

Now the strange thing is I put it on two boards, X38 and the DDR3 Giga EPT45-UD3P and it failed the RAM every time if I go above 450 FSB. Now I know the DDR2 Giga EP45-UD3P is a killer when it comes to clocking 45nm quads, so I took another drive to Frys to get one....and wholla, I'm priming while I type at 471 x 8.5 (4 G's) with only 1.25v in CPUz and temps just touch 50C on air with and old crappy Thermalright 120 (not the extreme version)

Now the moral of the story is do not in the first instance blame the CPU if you are having trouble with your OC, your board could just not be the best fit to yield a good OC

Then those who are looking still to get a good Q9550, look out for the "C" batches

 

[ghost_recon88]

Looking good man, that chip reminds me of my old Q9550 E0 I used to have

 

[Evilsizer]

your only at 4ghz and havent broken 60c yet... that just means to push it more! make it scream!

also ram does play a role as well, just throwing that out there. that way no one thinks they can get away with DDR2-667 and tring for 4ghz, LOL

 

[Brolloks]

I funny enough used RAM that was proven to run at 1600 (DDR3) and 940 (DDR2), I also thought it was the RAM at 1st, even tried a set of double sided Ballistix 1066 goodness, failed at 32M on the X38 whereas it runs fine at the same speeds with an E8400. I suspect the NB on both boards cannot handle the quad at FSB > say 450.

 

[doz]

Yes, Ive figured this out as well. As previously posted numerous times, I went from a P5Q-Pro to another one (when my power supply went out, it took the board). The next board I got was total garbage. First board I could hit 512 FSB (stable w/ 8.5 multi) and hit 4.6ghz. Next board I could NOT get above 478FSB stable.

I had read something that the earlier P5Q's were alot better than the older batches. Not sure how much truth there is to that, but motherboards DO make a HUGE difference. They are just as variable as CPU's IMO.

 

[SuperDave1685]

Very good point. Sometimes the difference between a poor overclock and a good one is just simply the motherboard you are using. My good friend has a similar experience when running his quad on an older 975 board and switched to a P35 board.

 

[hank123]

Brolloks that board has FSB holes and divider issues. I know. LOL

 

[ChinStrap]

4GHz @ 1.25 vcore.

also, p45 Vs p55?

sweet dude

 

[Brolloks]

Brolloks that board has FSB holes and divider issues. I know. LOL

Actually this UD3P is more consistent than the EPT-45 UD3P which has huge holes and is very RAM finicky.

 

[ghost_recon88]

EP45T-UD3P has some major BIOS problems, only Gigabyte P45 DDR3 worth considering is the EP45T-Extreme.

 

[Neuromancer]

Dunno what the difference is with the T (I think on other manufacturers it means DDR3 version). Very nice results though. low 50s loaded, 4GHz on less than 1.3 volts.

Excellent job

For 45nm quads, p45 is definitely the way to go for high bus speeds Weird that two similar boards produced such a radical difference though. Did you try resetting CMOS or updating BIOS on the older board?

 

[madhatter256]

Motherboard is like a foundation. Without a strong foundation the building (OC) can fall apart.

The RAM is like the steel beam. Without a strong enough steel (high OC capability) the smaller the building will be.

The PSU.... well the PSU helps in determining how long the OC will last.

 

[Brolloks]

Weird that two similar boards produced such a radical difference though. Did you try resetting CMOS or updating BIOS on the older board?

Yep, another interesting titbit is that this particuliar X38 (Abit IX38) is the same board that took many of my Q9650 close to 4.5 v, maybe the board is getting tited, I hammered it enough in the past,

 

[Craxbax]

I would agree that the individual mobo is a key to great OCs. My GA-P35-DS3L has a NB problem with quads also. I can run the G.Skill PC6400 memory at close to or over 1100mhz but a 400mhz NB setting is a bust. I am going to stick a new E6300 on it and see if this problem continues.