How to set up GTL Ref Values for 45nm & 65nm

Also to make things more precise "cross" exchanging correction numbers to "-" values like example below does not lead to same GTL resolute. They are completely different values which give completely different GTLs.

GTL Ref 0 = +50mv = GTL = 0.9336025v
GTL Ref 1 = +10mv = GTL = 0.9381305v
GTL Ref 2 = +50mv
GTL Ref 3 = +10mv

not equal

GTL Ref 0 = +30mv = GTL = 0.9136025v
GTL Ref 1 = -10mv = GTL = 0.9181305v
GTL Ref 2 = +30mv
GTL Ref 3 = -10mv

Above calculations made for vtt = 1.365v

Edit made to the first post as requested, thanks for the input guys. I made some small changes but they are essentially exactly as you wrote them, and credit is given to both of you. Cheers, and happy new year.

You too mate! Kickin back at home isn't all that bad, it's something different than the usual!

Hey ! :)

Yup happy new year to all of ya ppl :)

I got little present for every one for new year ;).

It is simple Excel file which is counting GTLs and NB GTL as well as new Multiplier for you and few other minor values for given Vtt. It is nothing special but saves time and allows to tweak GTLs without asking for help and getting too much into it.

How to use it:

First type in desired / currently used vtt voltage as in your bios (or rather the way it looks loaded in Everest for example - but bios one is also working fine)

Secondly put some correction numbers and watch how GTL values are changing (Max and Min values of GTL for given VTT are also shown there).

The goal is to figure out which GTL works best for you for one VTT - in most cases the same GTLs voltages will work for different VTTs or very similar ones (or sometimes not ). So when we switch Vtt we would need to adjust GTLrefs to values which will give same or close to same GTLs as working previous ones for different vtt.

Anyway have fun using it, I created it to make my life easier and it does the job.

PS. There are 2 sets of values to compare between different VTT.
PS2. This calculator is made for MF2 board but will work with all boards that got correction number in equation (like +40 or - 20 and so on).

Quote:

---

Originally Posted by mikeyakame

You too mate! Kickin back at home isn't all that bad, it's something different than the usual!

---

Ahh nice mate, same here, just having a quiet one in for once too ;) Thanks too for the descriptions of the GTL's you provided, I couldn't have said it any better myself.

Quote:

---

Originally Posted by seban

Hey ! :)

Yup happy new year to all of ya ppl :)

***snip***

---

You too mate, hope you're having a good one.

Also thanks for taking the time to make up that spreadsheet, I have added it and your description to the first post.

Do different chipset have different CPU, NB GTL as default.

Do P45 and X48 have the same

Example:

GTL Ref 0 = ,635
GTL Ref 1 = ,667
GTL Ref 2 = ,635
GTL Ref 3 = ,667

Is´t the CPU who determent what CPU GTL ref there´s, 45nm, 65nm or is´t the chipset.

What determent the NB GTL ref, does X48 and P45 chipset have the same GTL ref and does´t change if I run 45nm or 65nm CPU.



I run a EP45-UD3P board with a E8400@FSB 500 and I found that the most stable CPU ref are ~,635 and NB ref ~,600, the higher FSB the lower NB ref.

With my Q6600@FSB 475 I found that the most stable CPU ref are ~,667 and NB ref ~,700 the higher FSB the higher NB ref.


I have read the whole thread but I still think all this is quite confusing. :shrug:

I understand that there´s a lot of trail and error also involved in this, I try different settings and I´m learn new this all the time.

Is there anyone how can answer me questions it will be much apprecieted.

Quote:

---

Originally Posted by Ghostleader

Do different chipset have different CPU, NB GTL as default.

Do P45 and X48 have the same

Example:

GTL Ref 0 = ,635
GTL Ref 1 = ,667
GTL Ref 2 = ,635
GTL Ref 3 = ,667

**No not every board either X48 or P45 has the same setup, it depends on how the company wants to configure the BIOS options.

Is´t the CPU who determent what CPU GTL ref there´s, 45nm, 65nm or is´t the chipset.

**It's a combination of both the CPU and the individual board, for example if you keep the same CPU and replace the board with the exact same board again, you may need to re-adjust the GTL Refs as each set of board components/circuits behave slightly differently and give different resultant voltages even though you have set the BIOS options the same.

What determent the NB GTL ref, does X48 and P45 chipset have the same GTL ref and does´t change if I run 45nm or 65nm CPU.

**No the default GTL Ref multi of the X48 is different to the P45. What you set it to changes depending on the individual board (what is good for one board may nt be good for another board of the same type) and also if you run a 45nm or 65nm CPU, if you run a quad or a dual core, if you run 2GB or 4GB of ram, and if you run Crossfire or just a single card. It must be tuned for your individual components.


I run a EP45-UD3P board with a E8400@FSB 500 and I found that the most stable CPU ref are ~,635 and NB ref ~,600, the higher FSB the lower NB ref.

With my Q6600@FSB 475 I found that the most stable CPU ref are ~,667 and NB ref ~,700 the higher FSB the higher NB ref.


I have read the whole thread but I still think all this is quite confusing. :shrug:

I understand that there´s a lot of trail and error also involved in this, I try different settings and I´m learn new this all the time.

Is there anyone how can answer me questions it will be much apprecieted.

---

I have answered in your quote as best I can. HTH.

Heres a stab in the dark as to why there appears to be a small difference between X48 and P45 chipsets for GTL, it's just a shared thought.

P45 chipsets have a feature called CPU Margin Enhancement at least on Asus bios', what Intels official name is I don't know. From the little I know and I can figure out, the Margin Enhancement is a feature to address the electrical and physical constraints the front side bus had reached for the X48 based chipsets. These are most likely preset values either by OE or Intel that set up the signal skewing for the FSB/CPU clock waves to minimize clock jitter and ringback on the receiver, slew rate compensation, etc. Why you ask? AGTL+ is designed with limit of 400MHz on BCLK and beyond that integrity of the system can't be guaranteed. So P45 being mainstream and last ditch effort on FSB chipset to go beyond theoretical limits, this was added.

Now the point of the previous paragraph is, these presets most probably compensate for a percentage of signal problems that occur with high FSB frequencies that on the X48 chipset we we're using varying GTL Reference Points to handle compensation for clock signal integrity loss.

This is just a shot in the dark at why. None of us will ever know for sure, just theories and speculation.

You'll probably find that Asus' option names, Compatible / Optimal / Performance actually are presets for FSB frequency ranges. Compatible might be < 400MHz, Optimal 400-500MHz, Performance 500MHz ->. Thats just an example as its not that simple, but there will be ranges of frequencies that can be probably be compensated for with a margin of accuracy before those preset values go out of range.

If there is one thing that sucks though, it's that Asus, DFI, etc don't document what these settings and values do because they want to keep their design edge on their competitors products. This unfortunately has the consequence of us the guy who buys the product and uses it having to figure out what "insert corny performance name" option actually does and how it effects every other setting. I love having such flexibility, but some day the price of it without documentation that at least outlines its use it'll be too much for anyone.

Thank you CryptiK and mikeyakame

But how do I know what the default value for CPU GTL ref are when I calculate, lets say Q6600 on a P45 board and I can set the GTL for all cores.

Are the default (auto) like this on a P45 board

GTL Ref 0 = ,635
GTL Ref 1 = ,667
GTL Ref 2 = ,635
GTL Ref 3 = ,667

And if I want ,667 on all cores I hade to ad (+) some (how much is depended on the VTT value) for 0 and 2 and let the 1 and 3 left at auto so they all are at ,667.

Problem I've found on P45s is you dont know what the default should be, because of internal adjustments made by the bios code on top of changes you make! You just need to figure it out through a little trial and error is all.

Quote:

---

Originally Posted by Ghostleader

Thank you CryptiK and mikeyakame

But how do I know what the default value for CPU GTL ref are when I calculate, lets say Q6600 on a P45 board and I can set the GTL for all cores.

Are the default (auto) like this on a P45 board

GTL Ref 0 = ,635
GTL Ref 1 = ,667
GTL Ref 2 = ,635
GTL Ref 3 = ,667

And if I want ,667 on all cores I hade to ad (+) some (how much is depended on the VTT value) for 0 and 2 and let the 1 and 3 left at auto so they all are at ,667.

---

I can say for sure the Maximus II Formula has the GTL Ref options as:

GTL Ref 0 = ,635
GTL Ref 1 = ,667
GTL Ref 2 = ,635
GTL Ref 3 = ,667

But with other boards it may be different, also some boards tie two GTL Refs together (ie: 0/2 & 1/3) so you only have two adjustments possible, not four. In this situation 0/2 usually = 0.635x and 1/3 usually = 0.667x but unless they state what it is in the BIOS or in the manual it's anyone's guess.

EDIT - as Mikey said though there is often adjustments going on behind what you are doing and as such are beyond your control, so don't focus on what they are originally but through trial and error establish what settings give you the best performance and stability.

Thank you again CryptiK and mikeyakame

I get the picture.


Run Prime95 small with a little lower vcore the stable and see which core that fails and then adjust that GTL for that core and run Prime again.

Is this a good way to tune GTL?

or how do you guys test which GTL are most stable?

If Vcore is too low you'll probably you get hard locks or BSODs with stop error 0x09.

Always run vcore a little more than you need. That way you know it wont be the cause of error when fine tuning. Little more Vcore doesn't hurt to keep as a safe buffer anyway. At least that way GTLs are easier to fine tune. You coarse tune GTLs with Vtt by setting them roughly as Cryptik said. Then when the Vtt is roughly where it needs to be, you can fine tune the gtl's up or down depending on what yours needs. Longer it takes for a rounding error to appear the closer you are to nominal! If you get it right none should appear at all.

For a Q6600 I setup on a P5Q Deluxe, For 470Mhz FSB I found 0.635 and 0.655 for CPU and 0.645 for NB at 1.26V Vtt, and 1.34V Vnb with Margin Enhancement set to Performance I think? It's been a while so don't take this for gospel.

Thank you again CryptiK and mikeyakame for your time, much apprecieted, keep up the good work. :up:

Quote:

---

Originally Posted by mikeyakame

You coarse tune GTLs with Vtt by setting them roughly as Cryptik said.

---

Sorry but I am not able to find in the thread the Vtt value from which starting to tune.

What are roughly the rules for a dual core, X48 chipset (Asus Rampage Formula) and 4 sticks of 1 Gb memory?

Thanks,

start with a Vtt 0.02-0.04v above Vcore or so, set the NB ref to 0.63x, CPU ref to 0.65x, and Vnb will vary depending on strap, performance level and mem divider generally. From here once you get a baseline for yourself and find out what the minimum Vnb and Vcore you can work with is, lock Vnb down a step or two above minimum and setup NB ref, once you've got that merry, then CPU ref should nearly fall into place once you lock down a solid Vcore. You'll probably need to go up or down a few steps on Vtt till you find where it's most comfortable thanks to Asus minimalist gtl adjustment on the RF :D

EDIT - sorry guys - CUT

Hey guys

In my case seems that this chip or mobo likes negative GTLS...

I achieved to pass IBT at:

448x9 = 4032
vcore set 1,325 vcore idle 1,316
vtt 1,29
GTLS CPU +10/-50/+10/-50
pll 1,5
mems 1:1 (but tested on 1200 PL7 also pass IBT and 2500%MEMTEST)

IBT stable whenever i run it...BUT look this

450x9 = 4050
vcore set 1,325 vcore idle 1,316
vtt 1,2875
GTLS CPU +10/-45/+10/-45
pll 1,5
mems 1:1

I get it to pass IBT but today failed BSOD...then after some hours passed...again

Everytime i want to run on more VTT seems to be more unstable, also on positive GTLS BSOD comes faster
I'm kinda stuck on this, i want to start testing all over gain but really i don't know where to start on GTLS

I can also run it at 470x9 with low volts and i can play games etc...but so hard to get it stable

Cheers
Sergio

Quote:

---

Originally Posted by andressergio

Hey guys

In my case seems that this chip or mobo likes negative GTLS...

I achieved to pass IBT at:

448x9 = 4032
vcore set 1,325 vcore idle 1,316
vtt 1,29
GTLS CPU +10/-50/+10/-50
pll 1,5
mems 1:1 (but tested on 1200 PL7 also pass IBT and 2500%MEMTEST)

IBT stable whenever i run it...BUT look this

450x9 = 4050
vcore set 1,325 vcore idle 1,316
vtt 1,2875
GTLS CPU +10/-45/+10/-45
pll 1,5
mems 1:1

I get it to pass IBT but today failed BSOD...then after some hours passed...again

Everytime i want to run on more VTT seems to be more unstable, also on positive GTLS BSOD comes faster
I'm kinda stuck on this, i want to start testing all over gain but really i don't know where to start on GTLS

I can also run it at 470x9 with low volts and i can play games etc...but so hard to get it stable

Cheers
Sergio

---

Try 12:10 divider and see if its any more stable. I never had much luck with 1:1 on a P45, didn't like it all that much.

Quote:

---

Originally Posted by mikeyakame

Try 12:10 divider and see if its any more stable. I never had much luck with 1:1 on a P45, didn't like it all that much.

---

Thanks mikey i also noticed that, but i'm wondering now and i talked this with Cryptik if it's PSU related...

i have a TT toughpower 850W, but...

Cheers
Sergio

Ok after getting tired of my 450x9 i decided to push it a little and for my surprise took me less to make it stable, this chip seems hunger for vcore, more VTT has no effect at all, more VTT always lead me to unstability, same if i use positive GTLS...

The Gskills are great...

http://img50.imageshack.us/img50/170...227pl8pzx0.jpg

I haven't used a P45 yet that wasn't extremely anal when it came to Vtt, it either works great or it throws errors even when its only one step higher or lower. You should be able to enable DRAM static read, and I forget which skew is the lower one, cpu or nb, but for the lower one try around 400ps, higher one try around 100-200ps. Set top one first with other set to normal. Then set top to normal, and setup bottom one, then if you aren't too far off you should be able to set both to the valeus you had working and it should post without hassles. Took me an hour of failed posts to figure out how to set them up fairly quickly.

I remember the values and their relationship to menu, just not the correct skew name I used them for!

Btw very nice memory speeds there! I'm jealous :D Couldn't find the 9600s here in Australia when i replaced my last set of 4x1GB sticks, so had to settle for the 8800PIs instead. They don't unfortunately have a big working range, most I can get is ~ 1175mhz out of them as pushing more than 2.1v into them just throws errors, but they are rated at 1.9v 1100 5-5-5-15 and the pair I got do rated at 1.82V! I run them at 1148 5-5-5-16 1.96v works good here.

Whats the voltage you need for the 9600s?

Quote:

---

Originally Posted by mikeyakame

I haven't used a P45 yet that wasn't extremely anal when it came to Vtt, it either works great or it throws errors even when its only one step higher or lower. You should be able to enable DRAM static read, and I forget which skew is the lower one, cpu or nb, but for the lower one try around 400ps, higher one try around 100-200ps. Set top one first with other set to normal. Then set top to normal, and setup bottom one, then if you aren't too far off you should be able to set both to the valeus you had working and it should post without hassles. Took me an hour of failed posts to figure out how to set them up fairly quickly.

I remember the values and their relationship to menu, just not the correct skew name I used them for!

Btw very nice memory speeds there! I'm jealous :D Couldn't find the 9600s here in Australia when i replaced my last set of 4x1GB sticks, so had to settle for the 8800PIs instead. They don't unfortunately have a big working range, most I can get is ~ 1175mhz out of them as pushing more than 2.1v into them just throws errors, but they are rated at 1.9v 1100 5-5-5-15 and the pair I got do rated at 1.82V! I run them at 1148 5-5-5-16 1.96v works good here.

Whats the voltage you need for the 9600s?

---

Thanks mikey :up:

You won't believe but in Argentina they're selling them at 249 U$S, i bough them a month ago but payed almost 300...

Thanks for your help i will keep testing but this chip likes vcore more than all
The vdimm is on the everest screen on the gadgets

Sergio

Quote:

---

Originally Posted by mikeyakame

start with a Vtt 0.02-0.04v above Vcore or so, set the NB ref to 0.63x, CPU ref to 0.65x,

---

Thanks for suggestions. :)
I will try to find a smart spot......

Thanks for some great GTL advice, regarding looking at Gflops.

My findings (specs below) - stable 20x Linpack and 8h OCCT RAM.
Used intelburntest 1.9 at max memory (4GB total - about 3300MB used) in Vista x64 - 2 cycles.

GTL GFlops (x2)
+50/+10 47.0 46.6
+30/-10 47.8 47.8
+10/-30 47.0 47.1

Auto/Auto 47.6 47.8
+20/+20 46.3 46.5
-20/-20 Refused to boot

Now using +30/-10 although it was interesting to note how close auto/auto was.

Andy

Quote:

---

Originally Posted by andressergio

Hey guys

In my case seems that this chip or mobo likes negative GTLS...

I achieved to pass IBT at:

448x9 = 4032
vcore set 1,325 vcore idle 1,316
vtt 1,29
GTLS CPU +10/-50/+10/-50
pll 1,5
mems 1:1 (but tested on 1200 PL7 also pass IBT and 2500%MEMTEST)

IBT stable whenever i run it...BUT look this

450x9 = 4050
vcore set 1,325 vcore idle 1,316
vtt 1,2875
GTLS CPU +10/-45/+10/-45
pll 1,5
mems 1:1

I get it to pass IBT but today failed BSOD...then after some hours passed...again

Everytime i want to run on more VTT seems to be more unstable, also on positive GTLS BSOD comes faster
I'm kinda stuck on this, i want to start testing all over gain but really i don't know where to start on GTLS

I can also run it at 470x9 with low volts and i can play games etc...but so hard to get it stable

Cheers
Sergio

---

Hi m8,

TBH I was in kind of similar situation beofre at the beggining of OC.

I could do no problem Linx 10 passess but blend would fail in minutes simetimes seconds, sometimes hours and also I had different resoults from day to day for same settings - the reason was simple they were not stable and I need mode volts on most of compononets.

IMHO IBT or Linx are good as finish touch and 12h blend is more reliable - but also as my own example showed even 12h blend stable + 10 linx + Pi32 did not gourntee stability (correct skews does ;) :ROTF: in my case ;) ).

I think you should go back to begginning and work your way up with P95 instead if IBT - I could also run IBT at low voltages but not blend - I am pretty sure your volts are too low for 24/7h.

But I might be wrong of course. Also it would help if you would find point at which you start to be unstable at Auto GTLs. Becouse from ther we can find some nice GTLs for your setup.

Wish you luck and try with all around higher volts and Blend test - in fact forget blend test and do Large FTT.

And about your PSU what volts you got on lines under load?

Cheers

Seems more people is interested in tuning GTLs when using Quad core processors.
I have an Q9550 what can pass IBT 10 tests with only 1.2v and I'm wondering if it's GTL / skews related? I've tried GTL0 +50mv but unfortunately the chip seems don't like the juice and hard lock my screen:rolleyes:

You could be lucky and have one that likes whatever the GTL Refs are set to on auto on your RE, for the given Vtt. Each CPU/board is different though so its a matter of trial and error to find what it likes. Whenever you change the Vtt, you usually have to change the GTL Refs to suit. Also as you have a quad, you should be adjusting all 4 GTL Refs.

Really nice chip btw, not often you find one like that.

OK I need a little help here. I'm currently running E8500 @ 3,825 GHz @ 8,5x450 @ 1,304V. GTL ref values are on auto. Mine FSB VTT is 1,18V (Everest) - 1,20V (Bios), nForce SPP 1,4V and NF200 1,2V. Without adjusting GTL ref values mine E8500 need 1,376V for 4,037 GHz. In BIOS of my board I can only setup GTL ref values in mV, can't anywhere see GTL multiplier. Downloaded GTL calculator and if I put VTT 1,18 I get values 90/50/90/50. Since I don't know english to good I got little lost in this topic so can someone please tell me is 50/90/50/90 my starting point, because I also read that 2 core processors more like negative GTL ref values...

Quote:

---

Originally Posted by ci2kla

Seems more people is interested in tuning GTLs when using Quad core processors.
I have an Q9550 what can pass IBT 10 tests with only 1.2v and I'm wondering if it's GTL / skews related? I've tried GTL0 +50mv but unfortunately the chip seems don't like the juice and hard lock my screen:rolleyes:

---

Hard lock usually means FSB wall. Only way I know of to get around this is CPU/NB clock skews and CPU differential amplitude adjustments.

Quote:

---

Originally Posted by mikeyakame

Hard lock usually means FSB wall. Only way I know of to get around this is CPU/NB clock skews and CPU differential amplitude adjustments.

---

Not likely that a Q9550 will hit a wall @ 471...sorry I have to disagree with you:D
With all GTLs on auto IBT can pass 10 tests with 1.2v, 4003Mhz.
Apart from CPU GTLs, NB GTLs becomes more important as well.
For my example, E8400 need +20mv to get a certain FSB run smooth. But for Q9550 w/ RE, +70mv for vNB GTL is the sweet spot, +60 / +80 setting or add more vtt will cause IBT calculation errors that's my finding so far. Probably now the vtt is higher then vCore (1.206v VS 1.192~1.2v) maybe that explained the problem but I'm not sure...need more time and testing to find out. Will post more soon;)

Quote:

---

Originally Posted by ci2kla

Not likely that a Q9550 will hit a wall @ 471...sorry I have to disagree with you:D
With all GTLs on auto IBT can pass 10 tests with 1.2v, 4003Mhz.
Apart from CPU GTLs, NB GTLs becomes more important as well.
For my example, E8400 need +20mv to get a certain FSB run smooth. But for Q9550 w/ RE, +70mv for vNB GTL is the sweet spot, +60 / +80 setting or add more vtt will cause IBT calculation errors that's my finding so far. Probably now the vtt is higher then vCore (1.206v VS 1.192~1.2v) maybe that explained the problem but I'm not sure...need more time and testing to find out. Will post more soon;)

---

You may disagree, but a FSB wall can mean a multitude of things. It's usually a wall that's in place due to clock signals skewing too far apart. If you have the flexibility of bios options to work around then the wall will be higher. All depends on how good you are fine tuning the settings and how fine they can be tuned. RE has a lot of flexibility so if you are stuck at 471 then it's more the fault of the guy playing with the settings than the chip itself ;)

I hit a wall on my Q6600 at 482mhz on RF which isn't a static wall, but without skew adjustment to get around the wall it might as well be a hard wall ;)

Why there isn't NB/CPU clock skew in RF 0410 BIOS? Or it's MoBo function and my P5E X48@RF doesn't have it?

I need to ask something - is VTT tight just with CPU or this voltage have something to do with the NB? I'm asking this, because i have trouble with RAM, i try it at 1130 (it's 1100 by default) and i got errors sec after starting memtest. vNB and vDIMM doesn't help, so, there reason is somewhere else.

If hardlocks are cause by fsb wall, mine is 470. Even 1mhz up and it locks in windows. So far, the only think, that helps are VTT and vNB.

Quote:

---

Originally Posted by seban

Hi m8,

TBH I was in kind of similar situation beofre at the beggining of OC.

I could do no problem Linx 10 passess but blend would fail in minutes simetimes seconds, sometimes hours and also I had different resoults from day to day for same settings - the reason was simple they were not stable and I need mode volts on most of compononets.

IMHO IBT or Linx are good as finish touch and 12h blend is more reliable - but also as my own example showed even 12h blend stable + 10 linx + Pi32 did not gourntee stability (correct skews does ;) :ROTF: in my case ;) ).

I think you should go back to begginning and work your way up with P95 instead if IBT - I could also run IBT at low voltages but not blend - I am pretty sure your volts are too low for 24/7h.

But I might be wrong of course. Also it would help if you would find point at which you start to be unstable at Auto GTLs. Becouse from ther we can find some nice GTLs for your setup.

Wish you luck and try with all around higher volts and Blend test - in fact forget blend test and do Large FTT.

And about your PSU what volts you got on lines under load?

Cheers

---

Thank you Seban, on 8,5multi this situation is different i am much more stable and with more patterns than on 9 multi

will keep on testing

Cheers
Sergio :up:

Quote:

---

Originally Posted by spoof

Why there isn't NB/CPU clock skew in RF 0410 BIOS? Or it's MoBo function and my P5E X48@RF doesn't have it?

I need to ask something - is VTT tight just with CPU or this voltage have something to do with the NB? I'm asking this, because i have trouble with RAM, i try it at 1130 (it's 1100 by default) and i got errors sec after starting memtest. vNB and vDIMM doesn't help, so, there reason is somewhere else.

If hardlocks are cause by fsb wall, mine is 470. Even 1mhz up and it locks in windows. So far, the only think, that helps are VTT and vNB.

---

Because Asus bios engineers too lazy to add it. It was added to P5E3 Premium and P5E64 Evo as far as I know. Rampage Extreme came with the option already there, and all P5Q bios have this too. We drew the short end of the straw it seems my friend.

Vtt and Vnb can help to a point, but from there it's clock skews and fine independant gtl adjustment which we have neither of.

Calling it a wall just simplifies the actual explanation of the problem.

From what I can make out, the "wall" itself manifests mostly from excessive skewing between independant clock waves within the scope of the MCH and CPU, which on top of becoming dirtier from jitter, external interference, ringback, etc as bus frequency, cpu frequency, cpu multiplier rises. It can be as minute as one of the clock signals for either die becomimg too dirty and slightly skewed more than the other and you will experience a "wall" or hard system locks. This is probably just a breakdown of inter-die communications or cross-bus comms, or even phase cycle timing going out of scope.

This is the simplest I can explain what appears to me to be actually happening.

I can reproduce the "fsb wall" on a P45 chipset just with cpu/nb clock skews alone and go between completely stable at high FSB on a Quad Core to near instant cpu load onset hard locks. This is the concrete proof I base my conclusions on.

Quote:

---

Originally Posted by ci2kla

Not likely that a Q9550 will hit a wall @ 471...sorry I have to disagree with you:D
With all GTLs on auto IBT can pass 10 tests with 1.2v, 4003Mhz.
Apart from CPU GTLs, NB GTLs becomes more important as well.
For my example, E8400 need +20mv to get a certain FSB run smooth. But for Q9550 w/ RE, +70mv for vNB GTL is the sweet spot, +60 / +80 setting or add more vtt will cause IBT calculation errors that's my finding so far. Probably now the vtt is higher then vCore (1.206v VS 1.192~1.2v) maybe that explained the problem but I'm not sure...need more time and testing to find out. Will post more soon;)

---

And i have to disagree with you.
IBT stable means you have no problem with your CPU stability and that is basically it.
Try to run Prime95 Blend mode and stress you Graphics card and all other components on your system and see if you can do that without problems.
See my sig and you will know what i mean.

[QUOTE=mikeyakame;3574725]Because Asus bios engineers too lazy to add it. It was added to P5E3 Premium and P5E64 Evo as far as I know. Rampage Extreme came with the option already there, and all P5Q bios have this too. We drew the short end of the straw it seems my friend.

I always thought that we paid enough to have a good board. Maybe ASUS thinks we didn't pay enough to have a good BIOS on it.:D

All i see are some quick fixes in the BIOS to keep the crowd happy.:clap:

At least we don't have to worry how to setup the GTL's on our board, they are as basic as they can get. Not a problem with a good quad that can run with the CPU GTL's we have in our BIOS. Big problem if one core doesn't like the voltage.

I might be wrong, but i think that most people who can't run there Quad on high FSB on our board have the problem that one or more cores doesn't like the same CPU GTL Voltage like the others.

Quote:

---

Originally Posted by mikeyakame

...

---

I'm extremely thankful for this explanation on what is actually the problem. So, i'm moving soon to P45 and see what it's gonna help.

Quote:

---

Originally Posted by A-Grey

I might be wrong, but i think that most people who can't run there Quad on high FSB on our board have the problem that one or more cores doesn't like the same CPU GTL Voltage like the others.

---

Correct. At same FSB where i can run a little 65x and 67x is needed with higher on Prime95 Benchmark first to cores gives error after to sec. With 65x there is no problem, but further FSB rising becomes imposible. What's why DFIs are for so many people the best OC MB, because it has tons of settings and when you understand them OC i like singing :P

Quote:

---

Originally Posted by A-Grey

And i have to disagree with you.
IBT stable means you have no problem with your CPU stability and that is basically it.
Try to run Prime95 Blend mode and stress you Graphics card and all other components on your system and see if you can do that without problems.
See my sig and you will know what i mean.

---

Are you trying to say I have to stress everything at the same time to make sure the FSB is stable or not? I'm not sure what you mean...
Just done some extensive test. The reason why I choose IBT as the main testing tool is because it is the fastest tool to find out if the system is stable or not (of course not for 7/24). Calculation errors exists not only CPU but also can be other components related...for example...NB GTL vRef :up:

Last Progress so far: 1.2v vCore, 1.32v VTT CPU GTLs all on auto, NB GTLs +70mv, IBT Test = Hard Lock / Reboot
With CPU GTLs slight fine tune, IBT Test 80% correctness, ORTHOS Large FFTs HARD LOCK after 1 hr 47 mins

Will start to try getting 485FSB stable after I'm off and will post up more if any progress made. Cheers:D

Quote:

---

Originally Posted by ci2kla

Are you trying to say I have to stress everything at the same time to make sure the FSB is stable or not? I'm not sure what you mean...
Just done some extensive test. The reason why I choose IBT as the main testing tool is because it is the fastest tool to find out if the system is stable or not (of course not for 7/24). Calculation errors exists not only CPU but also can be other components related...for example...NB GTL vRef :up:

Last Progress so far: 1.2v vCore, 1.32v VTT CPU GTLs all on auto, NB GTLs +70mv, IBT Test = Hard Lock / Reboot
With CPU GTLs slight fine tune, IBT Test 80% correctness, ORTHOS Large FFTs HARD LOCK after 1 hr 47 mins

Will start to try getting 485FSB stable after I'm off and will post up more if any progress made. Cheers:D

---

Yes i do. IBT is good for your CPU stability and the first step to a stable overclock.

I use 1.22V for VTT as you can see in my sig for FSB 465MHz with CPU GTL 0.63X and NB GTL 0.67X. I can get it IBT stable with 1.16V. Prime95 Large FFT's will fail within minutes with the VTT at 1.16V.
You might need a little bit more voltage on the NB to pass Prime95 Large FFT's than you use for IBT.
To find the right Memory voltage i use MemTest86+ V2.11. If i can pass MemTest86+ V2.11 without errors i have no problems at all with my Memory stability.
The reason that i stress everything on my system is to find the right VTT to keep everything stable.

This may look insane. We are at Xtreme Systems. Why not do extreme stability tests.:up:

Try the latest BIOS for your RE. The latest one for my RF fixed the problem with the VTT. Earlier BIOSes didn't allow me to raise the VTT higher than 1.16V without losing stability in IBT.
The ASUS Tech Support fixed the problem for the RF with BIOS 0701, they released a new BIOS for the RE at the same time. The BIOSes where released a few days after they have tested a system similar like mine to find out why i couldn't raise the VTT without losing stability. I don't know what they did with that BIOS but now i can get it stable with higher VTT.

I have to wait now for my new memory to arrive to see if i can get it stable at higher FSB. The 2 X 2 Giga OCZ Reaper HPC PC2-8500 can't go over 1120MHz with 5-5-5-15 timings.

Quote:

---

Originally Posted by A-Grey

Yes i do. IBT is good for your CPU stability and the first step to a stable overclock.

I use 1.22V for VTT as you can see in my sig for FSB 465MHz with CPU GTL 0.63X and NB GTL 0.67X. I can get it IBT stable with 1.16V. Prime95 Large FFT's will fail within minutes with the VTT at 1.16V.
You might need a little bit more voltage on the NB to pass Prime95 Large FFT's than you use for IBT.
To find the right Memory voltage i use MemTest86+ V2.11. If i can pass MemTest86+ V2.11 without errors i have no problems at all with my Memory stability.
The reason that i stress everything on my system is to find the right VTT to keep everything stable.

This may look insane. We are at Xtreme Systems. Why not do extreme stability tests.:up:

Try the latest BIOS for your RE. The latest one for my RF fixed the problem with the VTT. Earlier BIOSes didn't allow me to raise the VTT higher than 1.16V without losing stability in IBT.
The ASUS Tech Support fixed the problem for the RF with BIOS 0701, they released a new BIOS for the RE at the same time. The BIOSes where released a few days after they have tested a system similar like mine to find out why i couldn't raise the VTT without losing stability. I don't know what they did with that BIOS but now i can get it stable with higher VTT.

I have to wait now for my new memory to arrive to see if i can get it stable at higher FSB. The 2 X 2 Giga OCZ Reaper HPC PC2-8500 can't go over 1120MHz with 5-5-5-15 timings.

---

Thanks for the information man. Actually I was using 1003 BIOS for testing (sorry for my sig not up to date:p:)
to pass 485 FSB is still a no-go...trial and error takes time and makes me easily lost the way:shrug:

Start from a stable overclock and work your way up in steps of 5 or 10MHz.

It took me months and a lot of BIOSes flashes to have the stable overclock that i have now. Before BIOS 0701 i couldn't pass the 460MHz and it took me a lot of time, tests and frustrations before i knew that the VTT was the problem.
The problem is every time your higher the FSB it gets harder and harder to find out where the problem is situated and what to do to get it stable.

Quote:

---

Originally Posted by ci2kla

Are you trying to say I have to stress everything at the same time to make sure the FSB is stable or not? I'm not sure what you mean...
Just done some extensive test. The reason why I choose IBT as the main testing tool is because it is the fastest tool to find out if the system is stable or not (of course not for 7/24). Calculation errors exists not only CPU but also can be other components related...for example...NB GTL vRef :up:

Last Progress so far: 1.2v vCore, 1.32v VTT CPU GTLs all on auto, NB GTLs +70mv, IBT Test = Hard Lock / Reboot
With CPU GTLs slight fine tune, IBT Test 80% correctness, ORTHOS Large FFTs HARD LOCK after 1 hr 47 mins

Will start to try getting 485FSB stable after I'm off and will post up more if any progress made. Cheers:D

---

Give this a go and see what happens:

GTL 0 = +10mv
GTL 1 = -35mv
GTL 2 = +10mv
GTL 3 = -35mv

Also to test CPU use Small FFT's & IBT, large FFT's more tests the ram & NB. However if your cpu is quite unstable, it can still cause large FFT's to fail.

My CPU have FSB Wall and VTT is the only thing that helps for pushing it back. So i have this question - have VTT voltage peaks? I mean, currently in BIOS i have 1.44v, but in BIOS Hardware Monitoring VTT is 1.36v. For me, 1.36v is still safe, but is this value same all the time or while booting, restarting and under load it rise to 1.44v?

Here for 471 it need 1.42v VTT, but for 472 it needs 1.44v VTT. This is with 65x GTL vRef.

Quote:

---

Originally Posted by spoof

My CPU have FSB Wall and VTT is the only thing that helps for pushing it back. So i have this question - have VTT voltage peaks? I mean, currently in BIOS i have 1.44v, but in BIOS Hardware Monitoring VTT is 1.36v. For me, 1.36v is still safe, but is this value same all the time or while booting, restarting and under load it rise to 1.44v?

Here for 471 it need 1.42v VTT, but for 472 it needs 1.44v VTT. This is with 65x GTL vRef.

---

My suggestion. Try CPU GTL 0.63X and NB GTL 0.67X and try to lower VTT.
This is what i'm going to use to get to 475MHz on my RF. This is Prime95 Small FFT's stable, large FFT's have to wait for the new memory.

OK After yesterday struggling I'm completely lost...BIOS issue??

I've tested the following settings with IBT @ 485FSB
1. Changing CPU 1 or 3 GTL vREF only makes things worse both -mvs and +mvs I wonder~
2. The best so far, 1.32v VTT, 1.2v vCore, CPU 0/2 GTL vREF -30mv. My chip don't like +MVs at all...

Tested Prime 95 blend Overnight (7 hrs) stable, but not IBT:rolleyes:
vNB @ 1.57 which is rather high...

CryptiK, thanks for your suggestion first I'll try that out later and report back :)

If you can't get Linpack stable enough to run at all you've probably hit a FSB wall on you board. Try adjusting CPU & NB clock skews and see if it helps any at this point. I don't know if you have CPU Differential Amplitude/Drive option? If you do adding 2-300mv should help a little.

Next issue here - after testing i found, that ot 471 the board is perfectly stable, no lockups. But since i tested the chipset and RAM, CPU Multiplier set to 6. After all was done i just push it to 8.5, give enought vCPU and boot. I was expecting, if something is wrong, that should be vCPU. But isntead of reboots, BSOD i get AGAIN hard lockups? Why? It's stable on FSB, but with rising of the CPU Clock it lockup?

I think you use the wrong CPU GTL's. If i was using the wrong GTL's i would have problems to get it stable around FSB 400MHz.

I tried Mikeyakame's suggestion to use CPU GTL 0.65X and NB GTL 0.63X and higher VTT with my Q9450 on my RF board. This makes it completely unstable. Perhaps if i would use multiplier 6 instead of 8 it might be stable.

Quote:

---

Originally Posted by A-Grey

I think you use the wrong CPU GTL's. If i was using the wrong GTL's i would have problems to get it stable around FSB 400MHz.

I tried Mikeyakame's suggestion to use CPU GTL 0.65X and NB GTL 0.63X and higher VTT with my Q9450 on my RF board. This makes it completely unstable. Perhaps if i would use multiplier 6 instead of 8 it might be stable.

---

I'm using 65x, tried 63x, doesn't help. I'll try 67x, but cores 0,1 don't like that high GTLs and they fail in P95s benchmark. For the NB, only 67x help, with lower i have mass RAM performance fail.

You can always step down a few notches on Vtt you know. Higher isn't always better, just depends on the FSB frequency. Some frequencies need a lot, others need less, then a few MHz beyond might need lot more or less. Just depends its a variable factor and all you can do is set GTLs to 1 set of multipliers, ie 0.67 and 0.63 for RF and then try raising and dropping Vtt to find a point where the failures occur after the longest period of time. Then from there try and leave Vtt and fiddle with GTLs and see if you can stabilize it, if you can't drop Vtt by 0.02, and try again, if not raise by 0.02 above your set value and try again. and so on. It's lots of trial and error mostly, but you'll get the hang of feeling what it wants after a while.

Quote:

---

Originally Posted by spoof

I'm using 65x, tried 63x, doesn't help. I'll try 67x, but cores 0,1 don't like that high GTLs and they fail in P95s benchmark. For the NB, only 67x help, with lower i have mass RAM performance fail.

---

Try the 0.63X for CPU GTL lower the VTT to 1.24V and see what happens. You can try higher or lower CPU Voltage to see if this makes it more stable.
If that doesn't work raise the VTT one notch and try again. Each time you raise the VTT one notch you should try with higher or lower CPU Voltage.

Once you can pass 10 or 15 minutes Prime95 Small FFT's you can try to get it stable with IBT to find the the right CPU Voltage.

One question. Do you use Loadline Calibration or is it disabled in the BIOS?

I suggest not to use it with 45Nm Quads and to disable it in the BIOS. I do a lot better with it disabled and some guys at ASUS Rampage Formula find that to.

So, lets put everything together:
1. vCPU 1.3*v, 470FSB x 6, CPU GTL vRef 65x, 1.61 vNB GTL vRef 67x PL8, 1.40 VTT - lockup after no more than 10 minutes.
2. vCPU 1.3*v, 470FSB x 6, CPU GTL vRef 65x, 1.61 vNB GTL vRef 67x PL8, 1.42 VTT - rock stable.
3. vCPU 1.46(measured with multimeter), 470FSB x 8.5, CPU GTL vRef 65x, 1.61 vNB GTL vRef 67x PL8, 1.42 VTT - lockup, same time. => Rising vCPU, rising vNB, rising CPU GTL vRef - doesn't help.

So far, as i said before, it seams, that VTT is the critical value here. 1.42v or 1.36v as Hardware Monitoring measures is near to unsafe values. If i'm 100% sure, that 1.36v measuret in BIOS is the real value all the time, maybe i'll push it more to 1.42v real. If this isn't the case i think it's worthless even trying to find with what VTT case 3 would work. Lowering VTT under 1.36v take me to highest stable FSB at 450. With 470FSB and 1.24v VTT it won't even start to load OS. What is the most stange thing for me here is why rising just the multiplier and applying enough vCPU (1.48 vCPU doesn't help too and i ain't gonna use higher for 24/) it push me back to lockup. I thought that max CPU FSB ot max CPU freq are two separate things.

In my personal experience Vtt needs to be higher than Vcc but only by 0.02v or so. If you raise Vtt the best thing to do is drop CPU GTLREF to around 0.62 and go from there. Raise Vtt lower GTL, drop VTT raise GTL. Sometimes you need to do both, But 470 FSB on X48 doesnt need anywhere near that Vnb on PL8. Which Divider are you using? Helps if you give that because not all dividers are stable at high multiplier. I have same problem low multi no issues, raise it and breaks most of the dividers. You have same sticks as I do, try use 12:10 divider I find its best personally if you arent already. Half Multis might not be all that stable at higher FSB too. You need to take that into account if possible. Try a full multiplier and see if same problem occurs.

With Vnb I use 1.49v LLC enabled PL7 on 12:10 divider at 481mhz. 0.63x NB GTL.

Maybe i'll try to lower vNB, honestly didn't try this, but it will help just with the heat, thas chipset is generation, won't help with the lockups. PL is set to 8, but i think i can lower it direct or through Pull-In's, but this is when everything is stable. For the NB GTL i have interesting problem - it can not boot while using lower GTLs than 63x and higher than 57x (i'm no sure, but between this two i have 61x 60x and one more). No just this, but using different than 67x cause my RAM to fail (or it's NB fail, but i get mass error detection it Memtest). As for the memory divider i use 12:10 too, the only way to achieve my rated speed of 1100Mhz. The most strangest thing is that when i try 1:1 it refuses to boot. Dunno why, this two boot problems, connected with the chipset are very strange for me. Apparently the problem here is in CPU, but what exactly and why, no idea.

Now when i think a bit - i had 1.42v VTT and 1.44 vCPU and it locked up, than i lowered vCPU to 1.3* and now i have no problem for 30 minutes (mostly, when it's going to lockup it take couple of minutes). I realy have to test this, thank you!

What i don't understand is why someone can use low VTT and others don't? Is there an explanation for this?

Depends on what Vcc and Vnb you are using! If using low voltages you don't need high Vtt! Only need high Vtt for high Vcc/Vcore.

I think that's what I'm working with. Yesterday I was completely lost as original setting won't stable anymore~
Hence I reset and start from where I begin. What I'm playing right now is clock skew. By using different clock skews in the same FSB, vNB GTL sweet spot will change.
But due to lack of time these testing still don't have a strong clue. Will post more about that soon.

Current Progress: VTT 1.37v, vCore 1.2v, FSB 471Mhz, CPU Skew: 400ps, NB Skew: 100ps, vNB GTL: +60mv, other CPU GTLs temp. on auto

Yes NB GTL Ref sweetspot changes when other parameters are changed, and it can be a very influential setting. Also, I found that I was Linpack/Orthos/Memtest stable at a certain vNB, but in games I would sometimes crash to desktop as if I was very unstable. Crysis was the best at doing this, sometimes it would only last 30 seconds before it crashed. Even increasing vNB from 1.30v to 1.45v did nothing to help this situation, nor did changing clock skews ram skews, or NB DDR Ref's or CHA/CHB Refs. I finally dropped my NB GTL Ref to -100mv, and although it only passed 2/6 tests in linpack, it was absolutely fine in crysis, no more crashes. I then bumped it up to -80mv, and that made the system totally linpack etc stable again, and was still totally stable in crysis.

I have found on my M2F, I really only need to use cpu/NB clock skew when I am over 540 - 550 FSB. Behaviour is different with quads though, some guys need to start using the skews around 450 FSB.

Hey cryp!

If you get bored would you care to try setting CPU Margin Enhancement to Performance and see if you need any less CPU/NB clock skew with it set to that?

Quote:

---

Originally Posted by mikeyakame

Depends on what Vcc and Vnb you are using! If using low voltages you don't need high Vtt! Only need high Vtt for high Vcc/Vcore.

---

What if i don't have to use high Vcc and need high Vnb for that FSB?

It seems that i can't change the NB GTL from 0.67X to a lower multiplier.

Is this the point that you talk about where you have to use a CPU/NB Clock Skew to keep stability?

Can be yes! Yeah with such a high Vnb it makes it hard to not use high Vtt. It's a double edged blade unfortunately. This is really where we near clock skew adjustments. High Vtt is just simply compensation for GTL/Skew related issues. I need a lot less Vtt on a P45 board than X48 just because I can set up skews and GTLs very finely. 475FSB on a P45 with 65nm chip I can get away with 1.26v Vtt with NB clock skew around 100-200ps, CPU clock skew around 400-500ps, same FSB on a X48 I need 1.47V Vtt real for same stability.

Quote:

---

Originally Posted by mikeyakame

Hey cryp!

If you get bored would you care to try setting CPU Margin Enhancement to Performance and see if you need any less CPU/NB clock skew with it set to that?

---

I would but I am using my M2F which doesn't have that setting. I was doing some testing with the P5Q-D last night (which does have that setting) but it's back in the box now.

With older bioses I found it didnt really have any effect on anything, but it may have been poorly implemented in earlier versions. It really seems to affect quads more than duals, but more in the sense of simply providing stability, as it was causing quad users some major hassles if set incorrectly before BIOS 1406 IIRC (I only tried it with BIOS 1201).

Quote:

---

Originally Posted by CryptiK

I would but I am using my M2F which doesn't have that setting. I was doing some testing with the P5Q-D last night (which does have that setting) but it's back in the box now.

With older bioses I found it didnt really have any effect on anything, but it may have been poorly implemented in earlier versions. It really seems to affect quads more than duals, but more in the sense of simply providing stability, as it was causing quad users some major hassles if set incorrectly before BIOS 1406 IIRC (I only tried it with BIOS 1201).

---

That makes sense indeed. I found same problem so I can only assume it affects clock skews in a major way. I'm surprised M2F doesn't have that option! Another thing is Mem OC Charger, I think that forcefully sets DRAM Clock Skews as I remember with it enabled the DRAM Clock skews for all slots were all locked when checked with MCHBAR dump or Everest.

I think that the options was designed to aid overclocking of low end cpu's (think e4300 etc). However being Asus' top of the line P45 board, the M2F was probably not expected to be paired up with low end chips, hence the option is not present (that's my take on the situation anyway).

Interesting what you said about Mem OC Charger, I always assumed it was asus-ese for ram drive strength. With my M2F, I cannot overclock the ram over 605 MHz with it set to disabled, but with it enabled, I can bench at around 650 MHz. Also with it enabled my D9's are stable at 600 MHz 5-5-5-15 @ 1.98v, whereas they need around 2.05v with it disabled. So you are saying that when it's enabled, ram clock skews when altered have no effect in reality as it presets & locks them? Any more info you can give regarding that option would be appreciated.

Quote:

---

Originally Posted by CryptiK

I think that the options was designed to aid overclocking of low end cpu's (think e4300 etc). However being Asus' top of the line P45 board, the M2F was probably not expected to be paired up with low end chips, hence the option is not present (that's my take on the situation anyway).

Interesting what you said about Mem OC Charger, I always assumed it was asus-ese for ram drive strength. With my M2F, I cannot overclock the ram over 605 MHz with it set to disabled, but with it enabled, I can bench at around 650 MHz. Also with it enabled my D9's are stable at 600 MHz 5-5-5-15 @ 1.98v, whereas they need around 2.05v with it disabled. So you are saying that when it's enabled, ram clock skews when altered have no effect in reality as it presets & locks them? Any more info you can give regarding that option would be appreciated.

---

I haven't really checked with it disabled. But on P5Q deluxe with it enabled I see Dimm1/2/3/4 Fine Clock Delay at 6/2/2/4 or so. CHanging DRAM Clock Skews for any dimm slot doesn't effect those values what so ever. Those values correspond to MCHBAR registers for DRAM Fine Clock Delay which DRAM Clock skew is supposed to set as well as DRAM Address Fine Clock Delay. It's odd to say the least. Don't have one at hand to check so can't say for certain.

I wish Asus instead of labelling options with stupid names would actually call them the right thing and explain what they change. Would make life of an overclocker so much easier.

With it disabled and with ram skews on 'normal' I have Dimm 1 - 4 fine clock delay of 6-2-2-4 respectively. IIRC changing ram skews had no effect on them even when Mem OC Charger was disabled, but I may be mistaken and will re-examine this later. Changing Ai Clock Twister settings also did not effect them. The boost in ram overclocking speed and reduction in required vdimm at any given speed with it enabled led me to conclude it was drive strength related, and as such I run it disabled 24/7, but I'd really like to know exactly what it does. I also really dislike the way asus name BIOS options, perhaps they think it sounds more appealing, but come on, we have already bought the board, they shouldn't still be using marketing strategies in the BIOS of all places.

I personally like how DFI name them. The right names and do exactly what they are named to do! Weird what you just said about them never changing. That has me thinking what the hell. Mem OC charger might do just what you think it does, with DQ drive strength stabilization. I think we'll never know what their options do exactly till we can figure out the nature of AMI 1B bios module and how to locate instructions for options.

i have a q9650 on asus striker 2 extreme

if my vcore is 1.38 vcore idle 1.38 load 1.33
and vtt 1.22

i would work this out at

1.22 x 0.635 = 0.775
1.22 x 0.667 = 0.814

0.775 - 0.814 = 0.039

as its a 45nm cpu i would look for closest setting i think are -40 and -36

-40
-36
-40
-36

then nb gtl i would set to -40

is this correct

as i change vtt the calculation would change

i need to under stand this to get better clocks but makes my head hurt lol

Yes as you change either Vtt or any voltage which uses Vtt as a termination for logical high you change GTLs. heh It's a tough one.

I'm personally not at all familiar with Nvidia's implementation of Vtt on their chipsets so I can't be of much help. It should be how Intel says to implement it for Vcc but I am not certain since NV doesn't use FSB rather it uses A hyper-transport like bus similar to AMD for Intel chips.

Still nothing here - i tried almost everything: lower VTT, higher VTT, vCPU, vNB, GTLs, PLL, combination, tried vCPU and VTT with 0.2 difference - still suffeting from hard lockups. Somethink interestin is, that i managed to lower my vCPU, but still this damn lockups. I'm out of any ideas. 471FSB rockstable with 6x multiplier, just rise it to 8.5x and everything goes to hell.

Quote:

---

Originally Posted by CryptiK

Yes NB GTL Ref sweetspot changes when other parameters are changed, and it can be a very influential setting. Also, I found that I was Linpack/Orthos/Memtest stable at a certain vNB, but in games I would sometimes crash to desktop as if I was very unstable. Crysis was the best at doing this, sometimes it would only last 30 seconds before it crashed. Even increasing vNB from 1.30v to 1.45v did nothing to help this situation, nor did changing clock skews ram skews, or NB DDR Ref's or CHA/CHB Refs. I finally dropped my NB GTL Ref to -100mv, and although it only passed 2/6 tests in linpack, it was absolutely fine in crysis, no more crashes. I then bumped it up to -80mv, and that made the system totally linpack etc stable again, and was still totally stable in crysis.

I have found on my M2F, I really only need to use cpu/NB clock skew when I am over 540 - 550 FSB. Behaviour is different with quads though, some guys need to start using the skews around 450 FSB.

---

That's the reason why i stress the whole system. Before i used UT2004 with Prime95 i had something similar like you've been experiencing with Crysis.
My CPU is at 100% load all the time with BOINC, and before i used UT2004 together with Prime95 a 3D application crashed within minutes.
The idea to use UT2004 together with Prime95 is not my idea. Some guy posted it on a forum, don't know what forum it was, to do it like this to find full system stability.

Like i said before, i do this to find the right VTT. With the wrong VTT UT2004 will crash sooner or later while Prime95 is still running without errors.

Cryptik!

I have a mission for you if you choose to accept it :D How would you feel about doing some MCHBAR dumps for me :D :D :D

http://www.tweakers.fr/baredit.html

You want this software to save the dumps.

All you need to do if you feel like it is this.

Change NB and CPU clock skews individual of each other without changing any other options in bios (that is the critical part), and boot into Windows load up Baredit, set MCHBAR to 32bit and click bar.txt.

CHange the file name to represent what setting you used. Do this 2 or 3 boots for each setting, and save dump each time.

Try a few different values if possible.

Ie
NB Normal, CPU 200PS
NB 100ps, CPU 200ps
NB Normal, CPU Normal

etc

Critical part is you remember which is which ! hehe. The 2 or 3 boots and dumps for same settings is to just compare against each to figure out which registers I don't recognize that are incrementing or not relevant. More dumps I have to analyze and compare the easier it will be to find what I'm looking for.

THis is just a hunch I have that NB & CPU clock skews are values stored in MCHBAR registers and not CPU MSRs or PLL Clock Gen registers.

Don't need to rush with it, and don't worry if a setting wont post just try another. Don't kill your system doing this either not even for sake of results! heh.

If its too much to ask just let me know, take all the time you need after all it's only time.

What I would like to achieve is being able to adjust these values dynamically inside Windows to aid with tuning provided they are within MCHBAR and they arent RWO (read write once) registers which require PWROK to clear.

By the way mate if you are interested in meddling with DRAM clock skews inside Windows let me know and I'll explain to you which MCHBAR register offsets to read and write to change them ;) Baredit allows that of course too thanks to Felix's awesome handywork.

Yeah mate for sure I'll do that, sounds very interesting I have read a bit about it but I have not got into it. That program if used as you directed is not going to cause any problems will it? I assume it just reads and dumps into a text file. So I open it, select 32bit (under configuration space & mchbar drop down boxes) then click bar.txt button. Where does the txt file get saved to?

So far nothing - lockup or reboot, all day long. Now i decided to do this - i set FSB to 470 and multi to 6x and start to lower all the voltages. Of course, it lockup while booting windows, what isn't new. The first repeatable sing, that i found is that i lockups just before entering LogOn screen, when 3 green light on the keyboard are flashing all 3 at a time. Lowering voltages doesn't make it lockup earlyer or later. So far i'm at 1.28 vCPU, 1.41 vNB, 1.5 CPU PLL, 1.3 VTT and going down. GTLs are 65x for CPU and 67x for NB.

So, i didn't do anything with finding max FSB and then testing the cpu. I'll set all voltages to minimum, 333x8.5 and i'll rising 10mhz at i time and 1 step on required voltages at a time. That's all i can do ...
Question - how to set GTLs just to be sure, that they aren't too high? From testing so far 65x on the CPU and 67x on the NB give the most stability, but since now i'll be going from down to up, may be i need to set them to min available values - 62x CPU GTL and 52x NB GTL? Any help here?

Quote:

---

Originally Posted by CryptiK

Yeah mate for sure I'll do that, sounds very interesting I have read a bit about it but I have not got into it. That program if used as you directed is not going to cause any problems will it? I assume it just reads and dumps into a text file. So I open it, select 32bit (under configuration space & mchbar drop down boxes) then click bar.txt button. Where does the txt file get saved to?

---

wherever you save it, default is baredit install folder!

Nah it wont cause any probs for reading. It just reads pci register space is all!

Looks like I don't have much luck with my E8500. Since I have EVGA 750i motherboard I only have positive values to set up. Tried with 40/0/40/0 and 50/10/50/10 with 1.35V in BIOS, actually 1.336V in load at 9x445 with FSB VTT at 1.2V (1.18V Everest says) and nForce MCP 1.40V, but Prime crashed with first core after 2 minutes...

With VTT 1.2V I get -0.038V that's way i tried combinations 40/0/40/0 and 50/50/50/10 for GTL ref values...

Quote:

---

Originally Posted by mikeyakame

wherever you save it

---

:rofl:

Quote:

---

Originally Posted by CryptiK

:rofl:

---

Indeed :D

For real though it should bring up a file browser window to select save location and file name!

@ Cryptik

On my Board (P5Q Deluxe) i can set for example 0.635 as GTL Value. So if i want to find my right gtl setting how must i start to set this value ?

For example if i select a higher value 0.640 the GTL increases,right?

So if i am running 1.20V FSBTer i can try what?

I have an E8600 CPU.

Yes the higher you set the multiplier the higher the Vref you will get. To start with, set them both to 0.635x and work from there. I have found with both my dual cores they like the address bus GTL Ref (the second GTL Ref on your board) to be a little higher than the data bus GTL Ref, so you could set 0.635x and 0.655x for example. Test for general stability gains at several setting, and find the most stable setup for the given FSB & vFSB. I'm now running on my M2F 1.25 vFSB and the equivalent of about 0.690x and 0.70x on your board, it helped to stabilize my current clock.

*********************

Mikey, I have to agree with you about fine tuning with IBT, it's quite effective. The times & GFlops are very close and consistent when tuned correctly. I have noticed that the first iteration is always a little lower & takes longer though, this is not due to speedstep delay I have that disabled. This run isnt perfect but its pretty close:

http://i266.photobucket.com/albums/i...z564MHzPL8.jpg

Quote:

---

Originally Posted by CryptiK

Yes the higher you set the multiplier the higher the Vref you will get. To start with, set them both to 0.635x and work from there. I have found with both my dual cores they like the address bus GTL Ref (the second GTL Ref on your board) to be a little higher than the data bus GTL Ref, so you could set 0.635x and 0.655x for example. Test for general stability gains at several setting, and find the most stable setup for the given FSB & vFSB. I'm now running on my M2F 1.25 vFSB and the equivalent of about 0.690x and 0.70x on your board, it helped to stabilize my current clock.

---

Thank you Cryptik

At the moment i am running Both GTL's in Auto and NB GTL Auto but i want to get them manually to see if i can decrease a little Vcore and some other(if this was possible).

At the moment i am Running the CPU 420x10 RS 15 Passes of IBT.
I tried also 430x10 RS 15 Passes of IBT and 500x8.5 RS 15 Passes IBT.
Maybe i can play a little more to get higher FSB using a smaller multiplier.
The memories i have work very good the for the things i do, Crucial Ballistix Tracer Led Pc6400 @ 560Mhz with only 2.08V measured with Tester on the Board.

I will let you know when i make some tests these days.


Thank you.
Vincent

Agree with both of you guys, i tuned my GTLS (Thanks CryptiK :up:) and my runs are very close now

http://img168.imageshack.us/img168/2...1207pl8os0.jpg
http://img168.imageshack.us/img168/4...pg/1/w1280.png

Sigh...Another problem...IBT Pass 20 runs that fails ORTHOS Large FFTs P9 in minutes...really disappointed and lost :shrug:

Does increasing vNB, raising the PL or increasing vdimm help? Is your system at least 1000% windows memtest stable?

Quote:

---

Originally Posted by CryptiK

Does increasing vNB, raising the PL or increasing vdimm help? Is your system at least 1000% windows memtest stable?

---

Or NB GTL, this was what help by me, when i was testing for FSB. It's NB<->RAM issue and by me not even relaxed timing help, just NB GTL one (or two, if nessecery) steps up.

Well, using 2 days 1.8VTT on your Q9550 don't kill it, let's see if it affects OC ... *speachless*

Quote:

---

Originally Posted by ci2kla

Sigh...Another problem...IBT Pass 20 runs that fails ORTHOS Large FFTs P9 in minutes...really disappointed and lost :shrug:

---

Try MemTest86+ V2.11 to test your Memory. If you can pass test 5 without errors you won't have any problems with the Memory.

I'd actually recommend using HCI memtest over bootable memtest, I have had ram pass bootable memtest that didnt last till 3% HCI memtest, and failed blend almost instantly.

It is found here: http://hcidesign.com/memtest/download.html

I assume you're running 2GB of ram so open 2 instances of HCi and set 860MB of ram in each and let it run till around 1000%. If you get an error keep tweaking your ram/NB settings. I also recommend making a custom run in Orthos of 4096K and letting it run for at least an hour.

If you can pass IBT then your cpu is at least quite stable, but running small FFT is also recommended.

Also do Blend in Prime95 it will find most memory / gtl errors and report as rounding errors. The severity of the rounding difference from what it should be gives you a rough estimate on the severity of the problem.

Small FFT tests L1/L2 cache mostly.
Large FFT tests small amounts of memory and L2 cache.
Blend tests large amounts of memory and copying between memory.

I find the custom run at 4096K FFT length I suggested is more stressful on the ram and NB than blend, but use whatever you want really stability is a subjective & personally definable thing.

OK I'll try the following tests and report back. Thank you guys...

1. Prime95 Blend: overnight
2. HCI Memtest 3.8...Quad run each 400MB for 1000%
3. ORTHOS SP2004 Small FFTs Priority 9

This step seems very complicated is this necessary to have a stable overclock like 4GHz on a licquidcooled i7 920

No. I7 doesn't use GTLs. it's for FSB only. I7 has its own architectually complex features you can get your head around though if it interests you ;)

Quote:

---

Originally Posted by mikeyakame

No. I7 doesn't use GTLs. it's for FSB only. I7 has its own architectually complex features you can get your head around though if it interests you ;)

---

Good to know, it interests me, but til a sertain point, to complex and I leave it on auto. Thanks

Update: Small FFTs Failed 1 hr 4 mins, error in result, not rounding
CPU GTL: -30, auto, auto, auto

Quote:

---

Originally Posted by CryptiK

I'd actually recommend using HCI memtest over bootable memtest, I have had ram pass bootable memtest that didnt last till 3% HCI memtest, and failed blend almost instantly.

---

Funny, i was Prime95 Blend test stable but couldn't pass MemTest86+ V2.11 Test 5 without raising the DRAM Voltage 3 steps in the BIOS.

Okay I hate to say this but on three separate occasions I've decided I would try and get my head around it and set my GTL reg values to break the 4GHz on my Q9550 and get stable (Prime95 overnight blend stable but tend to get random freezes when gaming/browsing/idle) but I just get lost >.<.


I've understood the math behind it, but my problem is I have everything set to Auto except Vcore (1.4V in BIOS, 1.36V CPU-Z) and NB/SB (1.5V / 1.5V).
Current setup: Asus S2E w/ Q9550 @ 3.966GHz (1866FSB) - 1x2GB DDR3 1600MHz 7-7-7-20 (Need to RMA other module, doesn't pass memtest).
Other than the ram values set to spec and the voltage I've already mentionned, everything is pretty much on auto.

What VTT value would be a good one to start off with if I set it to manual? What's considered a maximum for a 45nm quad? Without this starting point I can't aim to optimise my GTL Ref values.
Any additional S2E/Q9550 info that one may know would also be appreciated :)

Considering I got this far with only touching Vcore and NB voltages I would assume I could break 4GHz, my goal being a FSB of 1920 :).

Quote:

---

Originally Posted by A-Grey

Funny, i was Prime95 Blend test stable but couldn't pass MemTest86+ V2.11 Test 5 without raising the DRAM Voltage 3 steps in the BIOS.

---

Stress tests can be weird. Some guys pass IBT and fail small FT, some pass small FFT and fail IBT etc. There's so many variables between systems I guess there's no specific rules. That's why I use every stress test I can.

Quote:

---

Originally Posted by Bobly

Okay I hate to say this but on three separate occasions I've decided I would try and get my head around it and set my GTL reg values to break the 4GHz on my Q9550 and get stable (Prime95 overnight blend stable but tend to get random freezes when gaming/browsing/idle) but I just get lost >.<.


I've understood the math behind it, but my problem is I have everything set to Auto except Vcore (1.4V in BIOS, 1.36V CPU-Z) and NB/SB (1.5V / 1.5V).
Current setup: Asus S2E w/ Q9550 @ 3.966GHz (1866FSB) - 1x2GB DDR3 1600MHz 7-7-7-20 (Need to RMA other module, doesn't pass memtest).
Other than the ram values set to spec and the voltage I've already mentionned, everything is pretty much on auto.

What VTT value would be a good one to start off with if I set it to manual? What's considered a maximum for a 45nm quad? Without this starting point I can't aim to optimise my GTL Ref values.
Any additional S2E/Q9550 info that one may know would also be appreciated :)

Considering I got this far with only touching Vcore and NB voltages I would assume I could break 4GHz, my goal being a FSB of 1920 :).

---

Freezing/BSOD at idle but stable under load usually happens to guys using load line calibration (no vdroop) where under load the vcore rises slightly but at idle it's just one step too low.

I'd try setting your Vtt to 1.25v first and experiment with your GTL's and see what gives you the best stability. You can always work up from 1.25v as you test. From what I've read, around 1.40 - 1.42v for Vtt is about as high as most people go on air/water.

Some tests update again:
1. Small FFTs Priority 9 fail in 1 hr 4 min
2. Prime95 Crashed during overnight run therefore I don't know the duration

But one thing for sure is that the system is not stable still :shrug:
for IBT testing, 1.25v VTT will fail 1/10 runs while 1.2v can pass 20 runs. I wonder if vNB GTL is wrongly set (+60mv). Currently running 4x Memtest 3.7 (410MB each). Will report back ASAP after the result is out. I guess it's CPU GTL related but I don't have a clue still...comon:shakes:

:rofl: don't stress too much about it mate :D I've spent accumulations all up of weeks playing around with minor settings. Everything takes time I suppose.!