E6600 woes

cerebrex said:
The ram is actively cooled, case has a 25cm fan.. using my temperature probe (finger) the memory is case temp, remember Ballistix is 2.2v memory.. but it is running high, I will drop it down, thank you for the information, you've been an unbelievable help!

Also, I can't find ANYTHING on the net to configure speedfan for my board, is there a key anywhere to what everything is? temp1, temp2, temp3, etc... and the voltages?

btw Orthos STABLE 22min - load temp 49c ---eeek getting hot now.
Click to expand...

That's good that they're running cool. I've got d9's from ocz warrantied for 2.4-2.52v, but without active cooling above 2.1ish, they'll flake out eventually. Good to hear yours are staying cool.

As far as speedfan, there don't seem to ever be much updates to the database, but it's not a big deal, since most of the data is more or less explanatory. You just can't expect perfect results. For instance, in the case of the ds3, temp1 is system temp (somewhere on the motherboard, we're not sure), temp2 is cpu (although you have core1/2 as coretemps), hd temps are self explanatory, temp3 wasn't given a useful programming in the design of the motherboard, so just ignore it or choose to not show the reading at all.

vcore1 is vcore, vcore2 is vdimm, +3.3 and +5 are correct, +12, -12, and -5 are invalid readings. I believe you can get some of these readings from easytune's info page. The second +5v I believe is +5vsb, but don't quote me on that.
 
ziddey said:
That's good that they're running cool. I've got d9's from ocz warrantied for 2.4-2.52v, but without active cooling above 2.1ish, they'll flake out eventually. Good to hear yours are staying cool.

As far as speedfan, there don't seem to ever be much updates to the database, but it's not a big deal, since most of the data is more or less explanatory. You just can't expect perfect results. For instance, in the case of the ds3, temp1 is system temp (somewhere on the motherboard, we're not sure), temp2 is cpu (although you have core1/2 as coretemps), hd temps are self explanatory, temp3 wasn't given a useful programming in the design of the motherboard, so just ignore it or choose to not show the reading at all.

vcore1 is vcore, vcore2 is vdimm, +3.3 and +5 are correct, +12, -12, and -5 are invalid readings. I believe you can get some of these readings from easytune's info page. The second +5v I believe is +5vsb, but don't quote me on that.
Click to expand...

is there a north/south bridge temp?

oh btw Orthos FAIL 27min - peak load 50c

Do you think bringing down the ram voltage could lower my cpu voltage at all? I'd like to know if I can see whether or not to lower certain voltages
 
FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.49175
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.3v
vfsb: +0.2v

Orthos running
 
Keep the updates coming cerebrex, and 50c is nothing to these cpu's.......you are not overheating, even if you add 15c to the reported temps you are still OK. But I would not go over about 1.5 vcore.....:eek: :p
 
vdimm is independent of vcore, so no need to worry about that. week40's are known to not all overclock that well as an average, so I wouldn't be too surprised if it's the cpu holding you back and not the motherboard.

Have you tried removing the northbridge heatsink and changing the compound to asc/as5? The stock material isn't all too good, and +0.3v and 400fsb really burns the thing up. Even if it feels cool with the fan blowing down, it might not be transferring heat from the p965 to the heatsink well enough, although I don't think it'd be that big a difference from the stock compound.

Regardless, if it fails again, try +0.1vmch and +0.1vfsb. It may help it run cooler. Also, if it fails, give 401fsb a shot. It's not as good performance as 400, but it's a lot easier to achieve.
 
ziddey said:
Have you tried removing the northbridge heatsink and changing the compound to asc/as5? The stock material isn't all too good, and +0.3v and 400fsb really burns the thing up. Even if it feels cool with the fan blowing down, it might not be transferring heat from the p965 to the heatsink well enough, although I don't think it'd be that big a difference from the stock compound.
Click to expand...

Yes, I have lapped both the north and south bridges and polished them to a chrome finish, AS5'ed them and placed fans on each heatsink.

FSB: 400MHz
PCI-E: 110MHz
1:1
vcore: 1.49175
vmch: +0.3v
vpci-e: +0.1v
vdimm: +0.3v
vfsb: +0.2v

Orthos STABLE 1 hour 17 minutes - max load 48c SUCCESS

http://jadefalcon.shackspace.com/Pics/Computer/32_boot_004.JPG

Just as I took this pic it bursted to 51c during one of the orthos tests... highest I've ever seen... now, I can still bring these temps down about 8c if I have to, I have 2 fans that are throttled to 33% at the moment.. but I would rather not have to make the noise.


yay... looks like bringing down the vdimm, for whatever reason, has in fact brought down peak load temps 1.5c Looks like this clock is stable completely.

Thank you very much ziddey, I really appreciate your insight. I will save these settings, write them in my manual, and later after a few more hours to verify, I'll work my way to 3.4 with some higher bus settings.. what should I expect to encounter? What ratios would you expect need raising to achieve this? It would appear that my goal of 3.6GHz might be within reach given the mild temperature @ 3.2GHz load. What do you presume?

Thanks for the support Rapture! I thought I was a madman there for a little while, haha
 
good to hear you're seeing stability. I wonder if it'll fade in and out though like your tests before. I don't know about 3.6ghz though. It looks like 3.2ghz is even requiring a huge amount of voltage scaling, so I don't know if 3.6 is within grasp. Just because temps are good doesn't mean it's a green light to increase volts. There's still the concern of electromigration.

That said, I suppose
 
ziddey said:
good to hear you're seeing stability. I wonder if it'll fade in and out though like your tests before. I don't know about 3.6ghz though. It looks like 3.2ghz is even requiring a huge amount of voltage scaling, so I don't know if 3.6 is within grasp. Just because temps are good doesn't mean it's a green light to increase volts. There's still the concern of electromigration.

That said, I suppose
Click to expand...

Time will tell... this has proven stable however. I do have a few questions, what can be done to try and lower vcore? I do plan on proceeding.. a full out blasting from orthos for 3 hours and I peak at 50c... no game I'll be playing will ever cycle the cores that much, I never expect to see temperatures like that outside of orthos.

Cooler temps are on tap if I throttle up the rad fans:
http://jadefalcon.shackspace.com/Pics/Computer/32_boot_005.JPG

Well, there you go 3.2GHz stable with the [H]elpful forum members... I hope the system stays stable... but like I've said before, I'm at the mercy of it, so we'll see. Ziddey, I hope you'll continue to stop by the thread later on when I start breaking stuff again :)

What voltages should I expect to increase to make it to 500fsb 1:1... I'd like to get the memory to it's designed speed if at all possible.. if temps/vcore get scary I might decide to back it down.

Is there any way to monitor northbridge temps beside my temperature probe (finger)?

Also, as for the L640 batch... it would appear I would have the first verified OC to make the c2d chart for the batch (disgustingly).
 
nice. good to see that you're getting success now. the voltage you're using for 3.2ghz isn't bad at all in terms of being too much. What I was trying to suggest is that if it took that much vcore to hit 3.2, it'd probably require close to 1.6 if not more to get 3.6, which may very well be too much, when the longevity of the cpu is taken into consideration.
 
ziddey said:
nice. good to see that you're getting success now. the voltage you're using for 3.2ghz isn't bad at all in terms of being too much. What I was trying to suggest is that if it took that much vcore to hit 3.2, it'd probably require close to 1.6 if not more to get 3.6, which may very well be too much, when the longevity of the cpu is taken into consideration.
Click to expand...


Yes, that is definitely a concern I see also, your advice is well received - They didn't call it SNDS for nothing. The northwoods (I had a 2.4C @ 3.6) were 90nm technology at it's razors edge.. they ran much too hot - that's why they bit it so fast. However, electromigration is a derivative of ion bombardment causing micro vibrations which result directly in joule heating. So in essence, remove the heat, significantly slow the process.. it's actually one of the biggest variables in the equation (I actually wrote a thesis on this -super nerd). If I can keep the processor cool, the life of the CPU would probably certainly be decreased regardless, however to a degree which would not ever effect me...I mean, who cares if the CPU went from a life of 8 years down to 5? In 3 years it's most certainly going to be in a static bag in my closet... or maybe hanging in a frame with a engraved plate reading "4.0GHz beast" (alright, alright, I'm dreaming). However, I don't think I can cool this cpu that effectively at a much higher voltage... water cooling is a logarithmic curve, and I'm approaching the beginning of heat soak - my loop is very conservative.

I do have a phase system in that same closet also, but I don't want to battle the octopus anytime soon. I think I'll fire up a game and see what the 8800 thinks of that fsb... try and enjoy it a while or something before I break it again.
 
I ended up running 3dmark 06

3.0GHz - 9206 3DMarks

3.2GHz - 9539 3DMarks - peak load temp - 38c

Rather impressive increase for such a small jump.. especially since 3dmark is almost entirely videocard based.
 
cerebrex said:
I ended up running 3dmark 06

3.0GHz - 9206 3DMarks

3.2GHz - 9539 3DMarks - peak load temp - 38c

Rather impressive increase for such a small jump.. especially since 3dmark is almost entirely videocard based.
Click to expand...


Yeah nice isn't it? At my 3.2ghz I am getting 9647 in 3dm06, but that is in Vista and with an overclock of 600/925 on the 8800GTS. At stock GTS speeds it gets about 150-200 lower scores, right in line with your score. In XP it will get a 150 or so more than in Vista. I expect that gap to improve once NV gets some better drivers out for the 8800's.
 
TheRapture said:
Yeah nice isn't it? At my 3.2ghz I am getting 9647 in 3dm06, but that is in Vista and with an overclock of 600/925 on the 8800GTS. At stock GTS speeds it gets about 150-200 lower scores, right in line with your score. In XP it will get a 150 or so more than in Vista. I expect that gap to improve once NV gets some better drivers out for the 8800's.
Click to expand...

one more reason I don't run vista at the time being. my 8800 is clocked at 576/850
 
just a couple seconds after I posted that, I clicked "find optimal overclock" and it did exactly what it did to me last time in vista.. except in vista I got a nice sky blue screen, in xp I got a black screen.. so I rebooted...

... and the board started fast beeping at me (memory error or power error) - cleared the cmos to 2.13GHz - I reloaded my 3.2GHz OC and it booted...I think the board is in fact screwed with the dreaded cold boot bug... is there any solution to this?

ntune auto-tune causes an instant lock on multi-processor systems according to nvidia... great feature guys.
 
cerebrex said:
I reloaded my 3.2GHz OC and it booted...I think the board is in fact screwed with the dreaded cold boot bug... is there any solution to this?
Click to expand...


If you have the F10 bios, and it still exists....maybe go ahead and RMA the board? My cold boot bug was terrible with the F9 bios, but F10 took me from a shitty no booting board to a nice stable board. And I agree using ntune sucks....I don't use it except to monitor my card temps.
 
Yeah, I was in the middle of another bench after manually overclocking a little too high on the vid card... I was well on my way to squeezing over over 10k bench, and it locked.. I rebooted and decided to aggressively tighten up my memory timings.. and it went gay again, kept shutting down, so I rebooted and loaded my 3.2 settings.. wouldn't post... so I did a small 333fsb 3.0 oc, which did.. rebooted 3 times without it shutting down, and then went to the 3.2GHz setting... had to fool it.
 
cerebrex said:
Yeah, I was in the middle of another bench after manually overclocking a little too high on the vid card... I was well on my way to squeezing over over 10k bench, and it locked.. I rebooted and decided to aggressively tighten up my memory timings.. and it went gay again, kept shutting down, so I rebooted and loaded my 3.2 settings.. wouldn't post... so I did a small 333fsb 3.0 oc, which did.. rebooted 3 times without it shutting down, and then went to the 3.2GHz setting... had to fool it.
Click to expand...

man. Switching to the p5b from the ds3, personally, I've seen the new quirks of the p5b, and am not liking it. Almost made me want to try my ds3 again. But after I read your accounts on it, it reminded me of the reasons I decided to ditch that board. When it works, it's fantastic. But it fails to work like it should too often :-\. Hope you the best of luck
 
The DS3 seems to have no middle ground. Kind of like the FPD2485W display, you either get a great one, or you get a shitty one, no average ones.

My DS3 has treated me very well. I might go ahead and see if she will do a bit more tomorrow...I know my ram is good for another 60mhz over default, so that gives me something to work up from 400fsb I am at now. 460x8 = 3680mhz....hmmmmmmmmmm....doubtful. 3400mhz would be cool though!
 
ziddey said:
man. Switching to the p5b from the ds3, personally, I've seen the new quirks of the p5b, and am not liking it. Almost made me want to try my ds3 again. But after I read your accounts on it, it reminded me of the reasons I decided to ditch that board. When it works, it's fantastic. But it fails to work like it should too often :-\. Hope you the best of luck
Click to expand...

I hope this thread serves as a warning to future prospective DS3 owners, it seems more people have chimed in to share problems, than successes. Rapture seems to be in the minority here... it might be the later model processor batches just require a board with higher tolerances than the DS3 offers -- might seen like not a big issue at the moment, but when the market is flooded with more of these later model processors.. it will make this board much less appealing for the overclockers.
 
TheRapture said:
The DS3 seems to have no middle ground. Kind of like the FPD2485W display, you either get a great one, or you get a shitty one, no average ones.

My DS3 has treated me very well. I might go ahead and see if she will do a bit more tomorrow...I know my ram is good for another 60mhz over default, so that gives me something to work up from 400fsb I am at now. 460x8 = 3680mhz....hmmmmmmmmmm....doubtful. 3400mhz would be cool though!
Click to expand...

I don't think I'm going to push it further for now, I need a bigger rad to control 3.4-3.6GHz - Orthos stable for 5 hours 55c peak - left my door shut in a small room... I woke up and thought my house was on fire.
 
I noticed your vdimm at +.3 which im assuming is 2.1v..... If you have ballistix then you should never go below 2.2v, even at stock speed....
 
cerebrex said:
430w thermaltake - system is e6600 gigabyte DS3, 8800GTS, 2 500gb sata drives, 1 dvd burner, watercooling system.
Click to expand...

Lets's see.
"430w thermaltake w/ a 8800GTS and many cold cathodes, 2 500gb sata drives, 1 dvd burner, watercooling setup, etc.."

And you say:
"I see in CPU-Z that my core voltage fluctuates from 1.328 to 1.344 when it's set to 1.375 in the bios... and when I go to high voltages.. the system doesn't even start... is that symptoms of a weak psu?"

Is this really not adding up to you?

You're running today's hardware on yesterday's PSU. ;)
 
KJY9 said:
Lets's see.
"430w thermaltake w/ a 8800GTS and many cold cathodes, 2 500gb sata drives, 1 dvd burner, watercooling setup, etc.."

And you say:
"I see in CPU-Z that my core voltage fluctuates from 1.328 to 1.344 when it's set to 1.375 in the bios... and when I go to high voltages.. the system doesn't even start... is that symptoms of a weak psu?"

Is this really not adding up to you?

You're running today's hardware on yesterday's PSU. ;)
Click to expand...
lol you're writing today's date on yesterday's post

it's been done :)
 
KJY9 said:
Lets's see.
"430w thermaltake w/ a 8800GTS and many cold cathodes, 2 500gb sata drives, 1 dvd burner, watercooling setup, etc.."

And you say:
"I see in CPU-Z that my core voltage fluctuates from 1.328 to 1.344 when it's set to 1.375 in the bios... and when I go to high voltages.. the system doesn't even start... is that symptoms of a weak psu?"

Is this really not adding up to you?

You're running today's hardware on yesterday's PSU. ;)
Click to expand...


Yeah that could be an issue...I run a 500w Antec on 3 sata drives, a sata dvd drive, overclocked 8800GTS, and overclocked cpu. My voltages are stable but I don't think I will try pushing any more.
 
KJY9 said:
Lets's see.
"430w thermaltake w/ a 8800GTS and many cold cathodes, 2 500gb sata drives, 1 dvd burner, watercooling setup, etc.."

And you say:
"I see in CPU-Z that my core voltage fluctuates from 1.328 to 1.344 when it's set to 1.375 in the bios... and when I go to high voltages.. the system doesn't even start... is that symptoms of a weak psu?"

Is this really not adding up to you?

You're running today's hardware on yesterday's PSU. ;)
Click to expand...

you people have got to read the posts, lol. I did purchase an Ultra 600w 37a dual rail PSU - and it's installed - thats no longer an issue, but yes it was I believe.
 
luihed said:
I noticed your vdimm at +.3 which im assuming is 2.1v..... If you have ballistix then you should never go below 2.2v, even at stock speed....
Click to expand...

You think so? thing is, at *0.4v I get 2.31vdimm at 0.3v I get 2.13 .... nice gigabyte.
 
cerebrex said:
you people have got to read the posts, lol. I did purchase an Ultra 600w 37a dual rail PSU - and it's installed - thats no longer an issue, but yes it was I believe.
Click to expand...

He He..I didn't get that far into the thread...had NO idea it was this long.
Glad you finally realized the inevitable ;)
 
cerebrex said:
You think so? thing is, at *0.4v I get 2.31vdimm at 0.3v I get 2.13 .... nice gigabyte.
Click to expand...

2.31v should be fine 24/7..... I never had mine lower than 2.3v and when I talked to crucial tech support they told me to use atleast 2.2v..... Theres no benefit in undervolting your rams, they will last well after you have no more use for them... Personally I think you start harming your rams at 2.6v 24/7 or short sessions at 2.8v..... The rams that has died from high voltage from what ive heard is in the 3.1+v.......
 
luihed said:
2.31v should be fine 24/7..... I never had mine lower than 2.3v and when I talked to crucial tech support they told me to use atleast 2.2v..... Theres no benefit in undervolting your rams, they will last well after you have no more use for them... Personally I think you start harming your rams at 2.6v 24/7 or short sessions at 2.8v..... The rams that has died from high voltage from what ive heard is in the 3.1+v.......
Click to expand...

ehh I'd really like to believe that and used to myself until I went through too many d9's already. I ran a pair of buffalo firestix (d9gmh) at 2.2v with airflow going across the dimm's, and they only lasted me about 3 months. I don't know if I should blame the ram of the ds3 motherboard though.

Then I got some ocz 8800 d9gmh that say they can be run 2.4v. Initially, I was able to run 550mhz 5-5-5 with 2.2v. In my quest for tighter timings, I ran at 2.4v with direct airflow down on the sticks. Then, after a little under a month, they started acting flaky. I tried to test 550 5-5-5 again and this time it required 2.4v. 2.2v threw errors on me. Hmm :(


That said, I've got some buffalo d9gkx sticks now, and am running them at 2.1v for now. Once I get a new motherboard, I'll try higher again, but I figure it's time to play it safe
 
ziddey said:
ehh I'd really like to believe that and used to myself until I went through too many d9's already. I ran a pair of buffalo firestix (d9gmh) at 2.2v with airflow going across the dimm's, and they only lasted me about 3 months. I don't know if I should blame the ram of the ds3 motherboard though.

Then I got some ocz 8800 d9gmh that say they can be run 2.4v. Initially, I was able to run 550mhz 5-5-5 with 2.2v. In my quest for tighter timings, I ran at 2.4v with direct airflow down on the sticks. Then, after a little under a month, they started acting flaky. I tried to test 550 5-5-5 again and this time it required 2.4v. 2.2v threw errors on me. Hmm :(


That said, I've got some buffalo d9gkx sticks now, and am running them at 2.1v for now. Once I get a new motherboard, I'll try higher again, but I figure it's time to play it safe
Click to expand...

But really, who cares? lifetime warranty from crucial... they call don't call them them ballistix for nothing, hehehe....
 
I have 4 sticks of PC5300 ballistix, the fastest 2 I use to bench with and the 2 slow ones I only for 24/7 use..... My bench ballistix have seen 2.9v a few times and the other 2 just 2.4v but the funny thing is Ive only lost one stick and its one of the slow ones..... maybe these guys dont like low voltage lol.....
 
luihed said:
I have 4 sticks of PC5300 ballistix, the fastest 2 I use to bench with and the 2 slow ones I only for 24/7 use..... My bench ballistix have seen 2.9v a few times and the other 2 just 2.4v but the funny thing is Ive only lost one stick and its one of the slow ones..... maybe these guys dont like low voltage lol.....
Click to expand...

haha hmm. guess it comes down to cooling and duration. maybe you've got an extra fan on top or something. or it's because of long term continuous exposure to 2.4v by your slow sticks. then again, 2.4v is already a lot in my book ;) I wasn't even planning on running that much myself, but since that's what ocz specs, I eventually grew into running there, with adequate cooling, only to see them die too.
 
I'm having some scary problems now... my system locked up overnight idling.. and when I came back it once again cleared the cmos, so I had to load the 3.0GHz settings, and then load 3.2GHz settings.. I forgot to del back into the bios after loading the 3.0 and it started windows.. and locked... so I restarted and loaded the 3.2GHz.. it locked loading windows.. I loaded the 3.0GHz that always works, and it locked loading windows. I loaded the failsafe defaults at 2.4GHz and it locked during the boot post... I again restarted and it locked on the gigabyte motherboard splash screen, restarted again and it just stayed black nothing at all....

i opened the case up and wiggled my mem, the northbridge, pushed on the vid card a little, and it booted on failsafe now.....

grrrr..
 
cerebrex said:
I'm having some scary problems now... my system locked up overnight idling.. and when I came back it once again cleared the cmos, so I had to load the 3.0GHz settings, and then load 3.2GHz settings.. I forgot to del back into the bios after loading the 3.0 and it started windows.. and locked... so I restarted and loaded the 3.2GHz.. it locked loading windows.. I loaded the 3.0GHz that always works, and it locked loading windows. I loaded the failsafe defaults at 2.4GHz and it locked during the boot post... I again restarted and it locked on the gigabyte motherboard splash screen, restarted again and it just stayed black nothing at all....

i opened the case up and wiggled my mem, the northbridge, pushed on the vid card a little, and it booted on failsafe now.....

grrrr..
Click to expand...

Okay, tried again... 3.2GHz OC no longer functions. I swear, this board never stops hitting new highs of crap.
 
cerebrex said:
im going to head to frys to get a floppy drive and disk to flash to F10
Click to expand...


Gigabyte has a great windows based flash utility. I have not flashed via floppy in years since I have not isntalled one in that long ;)
 
while at fry's, check out oem e6600, it might be week 30... that's where I got mine, works fine. L630B something, default vid is 1.225, looks like an excellent clocker, at 1.325vcore which was default for my week 42 chip, I could bring it up to 3.3ghz, ran 8 hours Orthos stable (week 42 would go up to 3.06 @1.325v)... It runs hot but with your watercooling, it won't be a problem ;)
 
You must log in or register to reply here.
Back
Top