Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
I'm on at 5k! LOL!! Screen shots take 10 mins to upload!! I will have my 8 meg connection soon, I hope!! I just wanted to get the CPU vcore requirement out of the way. You want no less that 1.45 While its loaded! It will be more like 1.465 to be 8 hour+ stable is my guess!
Now we know the CPU can do it, barely!
I dont know what yer board can do.. ya see? I can only tell you to do what everyone else does. A Basic 400 FSB x 9 multi, NB to 1.40 or 1.50, CPU PLL 1 point higher than VCore. 1.5000=1.6, no spread spectrum, no CPU options... etc. You could have an obscure field that sets something to On, that shouldnt be for OCing. I dunno about that part!
You have to feel out yer board!
--Lupi!
Now we know the CPU can do it, barely!
I dont know what yer board can do.. ya see? I can only tell you to do what everyone else does. A Basic 400 FSB x 9 multi, NB to 1.40 or 1.50, CPU PLL 1 point higher than VCore. 1.5000=1.6, no spread spectrum, no CPU options... etc. You could have an obscure field that sets something to On, that shouldnt be for OCing. I dunno about that part!
You have to feel out yer board!
--Lupi!
Hey,
Can't post pics tonight but guess what? I booted into windows at 3.6!!! Lupi, I don't know how you figured this out, but it booted up the first time with your settings. Only problem is that it failed prime95 within like 2 mins! haha. I disabled c1e, Max CPUID value limit , Vanderpool, CPU TM function, but left Execute Disable Bit, PECI enabled (as per the OC guide).
Here's the screenshot:
There's no way I'm even trying for 5k, lol, thats cool though man hard to believe! Don't worry bout the pics, I didn't realize you're on a slow connection (how do you have the patience? lol). I set my bios vcore at 1.5, win idle/load is 1.47. The NB is 1.55 (no 1.45) and the PLL is 1.6 as you suggested (this is the one setting I never played with, what does it do?), I also set the fsb voltage to 1.4.
Basically I got into windows with exactly what you suggested (except the two cpu options enabled, but this didn't affect stability, both ways it was unstable). So what's next? I'm at 3.6 but not stable and my temps are under 62!!! Thats incredible.
Thanks for all your help man, I was totally ready to give up. Just need a little more tweaking and I should be stable! Woot!
Can't post pics tonight but guess what? I booted into windows at 3.6!!! Lupi, I don't know how you figured this out, but it booted up the first time with your settings. Only problem is that it failed prime95 within like 2 mins! haha. I disabled c1e, Max CPUID value limit , Vanderpool, CPU TM function, but left Execute Disable Bit, PECI enabled (as per the OC guide).
Here's the screenshot:
There's no way I'm even trying for 5k, lol, thats cool though man hard to believe! Don't worry bout the pics, I didn't realize you're on a slow connection (how do you have the patience? lol). I set my bios vcore at 1.5, win idle/load is 1.47. The NB is 1.55 (no 1.45) and the PLL is 1.6 as you suggested (this is the one setting I never played with, what does it do?), I also set the fsb voltage to 1.4.
Basically I got into windows with exactly what you suggested (except the two cpu options enabled, but this didn't affect stability, both ways it was unstable). So what's next? I'm at 3.6 but not stable and my temps are under 62!!! Thats incredible.
Thanks for all your help man, I was totally ready to give up. Just need a little more tweaking and I should be stable! Woot!
Update: I lowered the fsb 389x9=3.5 and it prime 95 is stable so far (only 10 mins). I did however raise the cpu/nb glt voltage to .67 and .63 (not sure what this does but others mentioned it). I'll test without them at 3.5 and see if it's still stable.
3.5 stable isn't bad at all, I bet I can get 3.6 with a few more tweaks. I wish there was better explanations of all of these settings (a bunch are, but many are not in the guide).
3.5 stable isn't bad at all, I bet I can get 3.6 with a few more tweaks. I wish there was better explanations of all of these settings (a bunch are, but many are not in the guide).
PLL is a setting used to stablize (offset) the frequency comming from the FSB. That's it in a nutshell but it's acctually very complex and supposedly a CPU is very sensative to high PLL voltage. Do a google search and you can find some good articles on PLL.
I acctually uspet that my bios doesn't offer PLL adjustments.
Here's another thread I like and gives a decent explanation of most voltage adjustments that can be made. http://www.blazingpc.com/forum/showthread.php/explained_nb_fsb_sb_voltage-8907/index.html
I acctually uspet that my bios doesn't offer PLL adjustments.
Here's another thread I like and gives a decent explanation of most voltage adjustments that can be made. http://www.blazingpc.com/forum/showthread.php/explained_nb_fsb_sb_voltage-8907/index.html
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Sure, if it took minutes to fail, ad 2 notches of vcore, That should do it.
Now... because the voltage starts so high, you may encounter the 3.6 to 3.8 syndrome earlier. It may fail again, then again!
To get prime stable from 3.6 @ 1.3000 volts loaded, I needed enough VCore to hit 1.45 Loaded. Thats 22 notches in my Bios! That was the ONLY way it passes primes 8k test. I believe at the 3.6ish area depending on VID, the processor has simply reached its technological limitations, and to go beyond that barrier you need lots of VCore to alleviate the failing logic gates!
The technology simply isnt made for the speeds we want! Some will get it. (Low VID.) Some will not get near it! (high VIDs.) And even a High VID q6700 starting 266 Mhz faster is BARELY stable at 3.6! LOL!!
VID is CPU Gawd!
--Lupi
Now... because the voltage starts so high, you may encounter the 3.6 to 3.8 syndrome earlier. It may fail again, then again!
To get prime stable from 3.6 @ 1.3000 volts loaded, I needed enough VCore to hit 1.45 Loaded. Thats 22 notches in my Bios! That was the ONLY way it passes primes 8k test. I believe at the 3.6ish area depending on VID, the processor has simply reached its technological limitations, and to go beyond that barrier you need lots of VCore to alleviate the failing logic gates!
The technology simply isnt made for the speeds we want! Some will get it. (Low VID.) Some will not get near it! (high VIDs.) And even a High VID q6700 starting 266 Mhz faster is BARELY stable at 3.6! LOL!!
VID is CPU Gawd!
--Lupi
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Working wonders, lol! To be honest I was just tired of people always saying that each component OC's differently. I wanted a better explanation, and look what I found!
I am way sure there are other small factors involved in a q6600 performance, but VID seems to be the most critical factor!
As in, since it costs the same for them to make all qxxxx processors in a given series, it doesnt cost anything if someone like dell ordered 100,000 q6600s and intel only had 75,000 on stock, they have been known to re label their q6700s to fill the rest of the order, only after re flashing the chip!
I learned that from a call with intel, BTW! And why not? Think intel sets out to make a "lesser" chip? Hell no! They all start as the highest value, then once tested, intel adjusts their "speed" accordingly.
--Lupi
I am way sure there are other small factors involved in a q6600 performance, but VID seems to be the most critical factor!
As in, since it costs the same for them to make all qxxxx processors in a given series, it doesnt cost anything if someone like dell ordered 100,000 q6600s and intel only had 75,000 on stock, they have been known to re label their q6700s to fill the rest of the order, only after re flashing the chip!
I learned that from a call with intel, BTW! And why not? Think intel sets out to make a "lesser" chip? Hell no! They all start as the highest value, then once tested, intel adjusts their "speed" accordingly.
--Lupi
Hey,
Thanks for the link, I couldn't find any info anywhere like this! Too bad you can't adjust pll, believe me , I didn't even know bout the relevance of this until a few days ago...lol.
Hey,
Ok i'll try raising the vcore two more notches as you suggested just for testing purposes
.The vid argument makes sense to me, ( don't know what intel says) a high baseline vid, gives you less room before you reach that max vcore since you're already high to begin with. Logically and the laws of physics support this argument.
Yeah I don't know where you get the chips from (maybe you work in a store lol) but its a good argument. However, I would be not all chips are created equal in processing. Some have lower vids (and are overall more stable) than others, its the normal distribution. So if I'm intel and I have this broad variation in chips, if I can rebrand them, why not? I doubt they are going to make every single chip from scratch... Not the ideal scenario but it seems to be working for them so far.
Just something to remember, not all chips will be able to reach 3.6GHz.
I have a Q6600 1.3v VID here that just cannot reach 3.6. I need 1.52v at 3.4GHz (currently) to stay stable for an hour running Orthos, this also needs water cooling or it frys (core temps read -20C till thermal or 75C in coretemp). I have tried 3.6GHz but I had to push things well over 1.6v and the temps just overheated the waterblock and cooling setup.
I currently run at 2.975GHz (425x7) at 1.325v or 3.4GHz (425x8) at 1.525v (1.505v after droop). I may try and fiddle again as I now know what the x67, x61 options do (they adjust the logic level readings of a 1 and a 0, its todo with line ringing and noise rejection)
I have a Q6600 1.3v VID here that just cannot reach 3.6. I need 1.52v at 3.4GHz (currently) to stay stable for an hour running Orthos, this also needs water cooling or it frys (core temps read -20C till thermal or 75C in coretemp). I have tried 3.6GHz but I had to push things well over 1.6v and the temps just overheated the waterblock and cooling setup.
I currently run at 2.975GHz (425x7) at 1.325v or 3.4GHz (425x8) at 1.525v (1.505v after droop). I may try and fiddle again as I now know what the x67, x61 options do (they adjust the logic level readings of a 1 and a 0, its todo with line ringing and noise rejection)
Demonic Goldfish
Distinguished
- Mar 31, 2008
- 45
- 0
- 18,530
Have you tried playing with the RAM voltage? That was the problem when my E8500 wouldn't boot past 3.4GHz...
Hey Qwak:
I couldn't find a stable 3.6 without taking my vcore way too high for my liking (over 1.53). I'm stable at 3.2 though, with most options on auto. I'm playing around with 3.5 but this is kinda silly now. 3.2 is 14 hrs stable, so i might just stick with that, I want to get around to actually PLAYING the games as well!
I couldn't find a stable 3.6 without taking my vcore way too high for my liking (over 1.53). I'm stable at 3.2 though, with most options on auto. I'm playing around with 3.5 but this is kinda silly now. 3.2 is 14 hrs stable, so i might just stick with that, I want to get around to actually PLAYING the games as well!
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Hmmm, thats interesting. But useful. I think I may need to hunt down info on that... the GTL Vref junk. I think I have x67 x65 x63 in there...
If that is a way to lower the level of "power" required for it to accept it as a 1 and not a 0, that would be awesome. Lowering that threshold SHOULD in crease your OC and technically lower your current VCore needs. Hmmm.... That would be very handy to "stretch" a few more Mhz out of your OC.
Damn 5 k connection and com casts slow speeds for instal of a new connection! waaaahhhh! Surfing the net blows at 5 k!
--Lupi
If that is a way to lower the level of "power" required for it to accept it as a 1 and not a 0, that would be awesome. Lowering that threshold SHOULD in crease your OC and technically lower your current VCore needs. Hmmm.... That would be very handy to "stretch" a few more Mhz out of your OC.
Damn 5 k connection and com casts slow speeds for instal of a new connection! waaaahhhh! Surfing the net blows at 5 k!
--Lupi
That's strange, I run a VID 1.3125 Q6600 at 3.6ghz with 1.45vcore, prime95 stable for 5+ hours, and stablized full load at 58 degrees C on air cooling. Pushed it up to 4ghz at 1.65vcore, prime95 stable for 1 hour, stablized full load at 70 degrees C.
Maybe something else is at work here.
Evilonigiri
Splendid
- Jun 8, 2007
- 4,381
- 0
- 22,780
Evilonigiri
Splendid
- Jun 8, 2007
- 4,381
- 0
- 22,780
dagger :
Weak part? Like a bum core? Yeah, I got that one. 
Sucks for you! 😛
Still, you managed an impressive OC for such a VID.
Dagger:
I tried my q6600 (vid 1.325) bios vcore all the way up to 1.525, and I still couldn't get it prime stable. I had nb at 1.55, sb 1.2, fsb 1.5, pll 1.6 and no go. My temps were up to 67 as well.
I don't know what it is, I'm not comfortable with vcore voltage up over 1.5, and 3.2Ghz is stable with most things on auto (14 hrs) so I might just leave it there with a very minor voltage inc.
I tried my q6600 (vid 1.325) bios vcore all the way up to 1.525, and I still couldn't get it prime stable. I had nb at 1.55, sb 1.2, fsb 1.5, pll 1.6 and no go. My temps were up to 67 as well.
I don't know what it is, I'm not comfortable with vcore voltage up over 1.5, and 3.2Ghz is stable with most things on auto (14 hrs) so I might just leave it there with a very minor voltage inc.
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Sounds about right, mine passes primes first test with 1.440, so it may need a lil more, but not much to be stable.
Just so you know, while I may be testing only through primes 8k test, once its stable enough to pass that, and within acceptable VCore limits, it takes maybe one or two more notches in my bios for an 8 hour stable, due to zero droop.
And I just have to much stuff to 8-12 hour test it yet! Even have a new OEM q6600! LOL!
I dunno why he has so much trouble...
Underclock everything but the CPU!
--Lupi
Just so you know, while I may be testing only through primes 8k test, once its stable enough to pass that, and within acceptable VCore limits, it takes maybe one or two more notches in my bios for an 8 hour stable, due to zero droop.
And I just have to much stuff to 8-12 hour test it yet! Even have a new OEM q6600! LOL!
I dunno why he has so much trouble...
Underclock everything but the CPU!
--Lupi
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Oh, and when I mention a VCore for stability, I am usually meaning under load. So when I say 1.440 thats loaded. If you have a huge VDroop, good luck!
As in, if I was on a 780i with the .06 VDroop... I'd need like 1.5250 in the bios to reach that LOADED droop.
--Lupi!
As in, if I was on a 780i with the .06 VDroop... I'd need like 1.5250 in the bios to reach that LOADED droop.
--Lupi!
Lupi:
Everything is underclocked. In order to reach 1.48 vcore under load, my bios vcore (with LLC enabled) has to be 1.525 + (1.44 loaded isn't even close to get stability for me) . I'm sure I can raise it more and get it stable at 3.6, but I'm not to keen on hurting the chip. My temps are very good at 3.2 (under 55) and I can play everything all my games at max so I'm happy there. If I could get 3.6 with a lower bios vcore that would have been ideal....
The ram is running 1:1 (ddr2-800, cpu 400x9), and the fsb has to be overclocked to 400fsb (its 333 naive), other than that, nothing else is OC'd when I was trying this.
Everything is underclocked. In order to reach 1.48 vcore under load, my bios vcore (with LLC enabled) has to be 1.525 + (1.44 loaded isn't even close to get stability for me) . I'm sure I can raise it more and get it stable at 3.6, but I'm not to keen on hurting the chip. My temps are very good at 3.2 (under 55) and I can play everything all my games at max so I'm happy there. If I could get 3.6 with a lower bios vcore that would have been ideal....
The ram is running 1:1 (ddr2-800, cpu 400x9), and the fsb has to be overclocked to 400fsb (its 333 naive), other than that, nothing else is OC'd when I was trying this.
Lupiron
Distinguished
- Feb 9, 2008
- 1,711
- 0
- 19,780
Thats true. LLC still has V Drop. Dont worry about it. It looks like the drop on my board is EXACTLY .2500, but then there is no droop.
1.42500 is exactly 1.4000 in windows. Or so it works on my board. Anyone know if intels Drop is just .2500, before testing and vdroop is applied!
Other than that, I have asked a few people and their 1.32500 VID chips OC with the same voltage as mine! 1.44/1.45 ish under load. There must me something more going on.
I have yet to buy the Lemon chip excuse. I have 7 q6600's now, and a pair of q6700's and none of them have a problem reaching 3.6 just near where I thought it would be!
Now here is a piece of info. There is something that I'll just call 3.6 to 3.8 syndrome. Its where the chip cant really OC without huge amounts of voltage.
Prime stable from 3.6 to 3.8 takes me 26 bios notches! It takes me 8 notches to go from 2.4 to 3.6 with my LOW VID. Then 26 more to be prime stable at 3.8!!
Now the higher VID's start to show 3.6 to 3.8 syndrom early. As a matter of fact, my 1.2625 does not show signs of that syndrome at 3.6. BUT my 1.2750 sure does!! So that means that the VIDs that start at 1.27000 and up will show the syndrome earlier because their VID and off the top voltage requirements are so high!
Hope that helps! It will all be in my review if I ever write it!
--Lupi!
1.42500 is exactly 1.4000 in windows. Or so it works on my board. Anyone know if intels Drop is just .2500, before testing and vdroop is applied!
Other than that, I have asked a few people and their 1.32500 VID chips OC with the same voltage as mine! 1.44/1.45 ish under load. There must me something more going on.
I have yet to buy the Lemon chip excuse. I have 7 q6600's now, and a pair of q6700's and none of them have a problem reaching 3.6 just near where I thought it would be!
Now here is a piece of info. There is something that I'll just call 3.6 to 3.8 syndrome. Its where the chip cant really OC without huge amounts of voltage.
Prime stable from 3.6 to 3.8 takes me 26 bios notches! It takes me 8 notches to go from 2.4 to 3.6 with my LOW VID. Then 26 more to be prime stable at 3.8!!
Now the higher VID's start to show 3.6 to 3.8 syndrom early. As a matter of fact, my 1.2625 does not show signs of that syndrome at 3.6. BUT my 1.2750 sure does!! So that means that the VIDs that start at 1.27000 and up will show the syndrome earlier because their VID and off the top voltage requirements are so high!
Hope that helps! It will all be in my review if I ever write it!
--Lupi!
TRENDING THREADS
-
-
Discussion What's your favourite video game you've been playing?- Started by amdfangirl
- Replies: 4K
-
-
-
Facebook
Twitter
Instagram