e6300 + p5b deluxe help

[catchsomezz]

hey y'all

haven't been here for awhile but long story short, thought i had my e6300 @ 3.5ghz stable, but i thought wrong anyways back to the drawing board...

right now i have:
asus p5b deluxe
e6300 @ 450x7
fsb = 450
vcore = 1.2500 (in bios)
vnb = 1.25
vsb = (i dont quite remember right now (@ work), but it set at the lowest manual setting. i want to say 1.35?)
fsb termination = 1.300
vich = auto
crucial ballistix 2x2gb ram (rated @ 2.0v, the "new" kind)
4-4-4-12-6-42, 1:1 running @ 900mhz

idle temps: ~45C
load temps: ~66C

i dont know if this is applicable, but my DRAM TRRD, Rank Write to Read Delay, Read to Precharge Delay, Write to Precharge Delay are all set @ 8

as of right now, im 12 hr dual prime95 stable, 10 hr memtest stable. i want to get back to the good old days of 3.5ghz (or higher? )

the biggest limiting factor, i think, is my load temp. from my personal consensus, it seems around 70 is the uh-oh point. if that is true, if i bump up the vcore and continue increasing the fsb, it wont be long until my load temps go from 66 to 70.

what do y'all think? can i bust through this or what?

 

[inkfx]

You are definitely limited by your temps. Until you can get those temps down, you are stuck around where you are now.

 

[x2s3w4]

Can you try reseating your Zalmann? Maybe you put on too much thermal compound. Many times when temps are too high it's a result of too much thermal compound that acts like an insulator instead of a conductor of heat. You only need to fill in the microscopic voids between heat sink and CPU. That seems like a lot of heat for that vcore. My Q6600 wasn't that high on my P5B deluxe Wifi at that voltage.

 

[erricrice]

ooh, yeah, somethins up with your temps, i've got a 6320 right now(same as yours, but with 4 mb cache) at 440 x 7 and 1.425v in bios and i'm running load temps at 40C and i've only got the zalman 9500

+1 reseating the heatsink

 

[catchsomezz]

!$*(&!#!.....

*summoning will*

 

[inkfx]

As previously stated, you could try remounting the heatsink. What method did you use to apply the thermal paste?

 

[catchsomezz]

As previously stated, you could try remounting the heatsink. What method did you use to apply the thermal paste?

its been over 2 years but i think the way i did it was small amt in the center, seat it, move it around a bit.

im thinking of just being content with this oc/temp. never re-seated. i need to take out the mb right? or have relaly really small hands.

 

[inkfx]

I've never gone that long without remounting. It's possible the paste is drying out and not working as well as it used to. As far as taking the motherboard out, you may or may not have to.

 

[catchsomezz]

improved! after some hardcore cable strapping and compressed air cleaning, my new temps are 40C/60C

i tinkered with the fsb (still @ 450), but even a modest increase to 460 causes errors in the memtest. anyone know anything about the new ballistix. they are rated to run @ 4-4-4-12 @ 2.0v as opposed to the "old" ballistixs which were rated to run @ 2.2v. can i safely bump the voltage of the memory from 2.0 to 2.2?

 

[inkfx]

Try upping your FSB termination and vNB a notch first. If increasing either of those doesn't help, then you could up the memory voltage. At 2.2v the sticks might start to get pretty warm. So if you go there, I'd look at slapping a fan on top of the sticks. A little airflow over the sticks will do wonders.

If your vCore is really at 1.25v, then you may be needing more voltage there and not to the board or memory.

 

[catchsomezz]

Try upping your FSB termination and vNB a notch first. If increasing either of those doesn't help, then you could up the memory voltage. At 2.2v the sticks might start to get pretty warm. So if you go there, I'd look at slapping a fan on top of the sticks. A little airflow over the sticks will do wonders.

If your vCore is really at 1.25v, then you may be needing more voltage there and not to the board or memory.

would an unstable vCore affect memtest?
also, i dont have any active cooling on the nb, is it safe to bump the voltage up to the next level (i think its 1.55? (im at the lowest setting, 1.25v))

 

[inkfx]

I was kind of tired when I wrote that and must have missed the part on memtest errors. Give the memory some more voltage and see if it's stable.

I've never seen an E6300 at 1.25v (with droop its probably closer to 1.2) @ 3.5Ghz, so you must have an incredible chip there.

 

[doublejack]

I've never gone that long without remounting. It's possible the paste is drying out and not working as well as it used to. As far as taking the motherboard out, you may or may not have to.

Actually, in my experience a good thermal paste never needs to be replaced. Once it goes through several heat cycles and settles well, it should be good to go for the life of the CPU.

improved! after some hardcore cable strapping and compressed air cleaning, my new temps are 40C/60C

i tinkered with the fsb (still @ 450), but even a modest increase to 460 causes errors in the memtest. anyone know anything about the new ballistix. they are rated to run @ 4-4-4-12 @ 2.0v as opposed to the "old" ballistixs which were rated to run @ 2.2v. can i safely bump the voltage of the memory from 2.0 to 2.2?

Please be advised that "old" ballistix was very different memory than the "new" stuff. Old crucial utilized D9GMH Micron chips. They overclocked like mad and responded well to voltage up to the 2.3-2.4v range, but had a tendency to die. The high RMA rate is the reason crucial doesn't use D9GMH in their memory any more.

The "new" ballistix is not nearly as overclockable, nor does it respond to voltage the same way. In short, you may be reaching the limit of your new ram, and if so there's nothing you can do about it short of buying different memory. You can try using more voltage but I have my doubts that will help any, and it may in fact kill your new ram.

would an unstable vCore affect memtest?
also, i dont have any active cooling on the nb, is it safe to bump the voltage up to the next level (i think its 1.55? (im at the lowest setting, 1.25v))

I have a p5b deluxe and have set the nb voltage and the sb voltage to higher levels. Even without active cooling there is no problem.

However, I don't suggest going higher than 1.3v on the termination voltage, as that can outright kill a CPU.

 

[catchsomezz]

darn...okay. i'll play with the nb settings.

as of right now, i set my ballistix with 2.1v. back down or keep as is?
does nb voltage have anything to do with ram stability?

update: i also just noticed that according to JEDEC the ballistix can run 5-5-5-15 @ 500mhz as opposed to 4-4-4-12 (currently) @ 450mhz. i'll probably give it a shot, and see if my cpu can even handle that

 

[liftedcj7on44s]

My E6300 seems to top out at 3.4ghz now matter how much voltage i feed it.

 

[catchsomezz]

I was kind of tired when I wrote that and must have missed the part on memtest errors. Give the memory some more voltage and see if it's stable.

I've never seen an E6300 at 1.25v (with droop its probably closer to 1.2) @ 3.5Ghz, so you must have an incredible chip there.

to be honest, i dont believe it either. with vdroop its 1.19. i've just finished 12 hrs on blend. im currently on the 3-4? hr of small ffts...trying to find a way to make it error out

 

[doublejack]

darn...okay. i'll play with the nb settings.

as of right now, i set my ballistix with 2.1v. back down or keep as is?
does nb voltage have anything to do with ram stability?

update: i also just noticed that according to JEDEC the ballistix can run 5-5-5-15 @ 500mhz as opposed to 4-4-4-12 (currently) @ 450mhz. i'll probably give it a shot, and see if my cpu can even handle that

Yup. The NB is where the memory controller is located.

Also, your plan of loosing the timings to 5-5-5-15 is a good one. That's a way to get more bus speed by trading latency.

 

[catchsomezz]

thanks for all your help. heres an update...

i tried 5-5-5-18 timings @ 2.1v, tried increasing the vnb to 1.45, fsb termin = 1.200
after a quick 5 passes of test#5 @ 475fsb, i ran smallffts, kept getting errors within 10 mins. tried vcore of 1.250 to 1.275. any more than that, and i'd be temperature limited (load temps were at 66C with 475fsb 1.275vcore)
i was hoping that it was an fsb wall (i dunno, i can dream right?) bumped to 485fsb @ 1.265vcore. error in smallffts.

right now...
460fsb
ram timing: 4-4-4-16 @ 2.1v
vcore: 1.250
fsb termination: 1.200
vnb: 1.45
same vsb
vich: auto

still testing stability, but 10 passes of test#5 and running smallffts(so far 5hrs stable)