**Official DFI LanParty UT X48-T2R(and LT) Discussion

[Praz]

:"Remember up to the 913 BIOS the actual GTL core functions are mislabeled. The actual core each function relates to is shown below:

CPU GTL 1/3 REF Volt = GTL ref 0/1 (Cores 0 and 1)
CPU GTL 0/2 REF Volt = GTL ref 2/3 (Cores 2 and 3)" http://www.anandtech.com/showdoc.aspx?i=3129&p=14

That was an assumption a lot of us made back then. As the data sheets are very non-descrip for current processors the info was obtained from 771 pin data sheets. I have mapped out the traces on the board and done physical electrical measurements. The BIOS is labeled correctly (as in the P35).

[drnick79]

I've been looking through the DFI LanParty X38 thread, and it is chock full of bios setting goodness, very useful when you're climbing the learning curve of the DFI bios.

To those whom have used both X38 and X48 boards, are the X38 bios settings that many have posted in that thread generally applicable to the X48 boards (aside from perhaps slight voltage differences)? Or is the X48 board a different animal altogether???

[deadman'sclick]

far closer than the rabbit & his girlfriend in ure sig pic lol more a rabbit & a hair

joking aside the setting will pritty much carry across the two mobos, u might get away with slightly tighter settings on the x48.

[PnoT]

It's working for me

Seems as thou the page is loading fine for me but the images are taking days to come through. **After a quick refresh it loaded fine... must have been something on my end**

I think I've figured this out but this example doesn't make sense to me and I'll illustrate with a screenshot I've taken.

So for example 1.100v VTT is 1.060v real. 1.060v x .67 = 0.7102 .

The First and second GTLs that are highlight are the closes match to .7102 but the NB is off a little bit. Wouldn't 0.707v be closer than the highlighted 0.716v? This is of course if my understanding is correct.

[Clunk]

I think you are right about that one, 0.7102 looks to be the correct value.

Also, the second table, the values don't match the motherboard settings - I'm on that as well

I will fix it later, thanks for letting me know

[PnoT]

I think you are right about that one, 0.7102 looks to be the correct value.

Also, the second table, the values don't match the motherboard settings - I'm on that as well

I will fix it later, thanks for letting me know

No problem

[Clunk]

Ok, fixed that error. If you spot any more, please let me know

[wilson]

Does any1 know where i can find the led error code info please. Ive had 26 show up one or two times and would like to know what it means.

[Clunk]

Does any1 know where i can find the led error code info please. Ive had 26 show up one or two times and would like to know what it means.

Here.

[wilson]

Cheers m8..

[CoolZone]

Clunk,you are great with your little hardware information gems!Good job!

[CERO]

Got the board on my way what is current best BIOS?

[PnoT]

Got the board on my way what is current best BIOS?

There are only 2 so far:

1. initial retail bios
2. bios update to fix temp monitoring for e8xxx cores.

They plan on releasing another bios from what Clunk says tomorrow to fix the wording for the GTLs.

[adya]

hey ive reached 450x8/1081mhz . 3.6ghz core speed. prime stable.
ive tried to push it to x9 =4ghz but it crashes every time.
my volts are -
cpu vid 1.425
dram 2.35
s.b 1.51
n.b 1.456
cpu vtt 1.250
gtl 1/2-110
gtl 0/3-105
gtl nb 80
MY QUESTION IS , WHAT CAN I DO TO REACH 4GHZ STABLE. FOR A 24/7 CLOCK ? OR AM I EXPECTING TO MUCH AND SHOULD BE SATISFIED WITH MY CURRENT CLOCK?

[drnick79]

Hey Clunk, any idea why my cpu voltage remains at 1.42v according to everest when the bios cpu voltage is set anywhere from 1.381 to 1.40? I saw your suggestion before about the special added voltage and changed that from auto to the lowest percentage available, and it makes no difference. I know that software voltage readings are not terribly accurate, but I don't think that would explain why it is "frozen" at 1.42v across all those bios settings.

Also, to anyone that has used 2x2gb ram (I'm using 2x2gb gskill 1066) on this or the dfi x38 board, what is the highest stable memory overclock you've achieved and at what voltages? I thought I was stable at 1128mhz, but after about a half hour of OCCT it craps out and I get errors.

Thanks for the help!

[drnick79]

hey ive reached 450x8/1081mhz . 3.6ghz core speed. prime stable.
ive tried to push it to x9 =4ghz but it crashes every time.
my volts are -
cpu vid 1.425
dram 2.35
s.b 1.51
n.b 1.456
cpu vtt 1.250
gtl 1/2-110
gtl 0/3-105
gtl nb 80
MY QUESTION IS , WHAT CAN I DO TO REACH 4GHZ STABLE. FOR A 24/7 CLOCK ? OR AM I EXPECTING TO MUCH AND SHOULD BE SATISFIED WITH MY CURRENT CLOCK?

Hi Adya

Is this with your x6800 cpu in your sig?
I tried pushing the x6800 cpu I had to over 4GHz on an nvidia 680i board, and if I recall correctly, it took quite a bit more than 1.425v to run it stable, so I never ran it that high for fear of sending it to an early grave. Higher CPU volts perhaps?

[adya]

Hi Adya

Is this with your x6800 cpu in your sig?
I tried pushing the x6800 cpu I had to over 4GHz on an nvidia 680i board, and if I recall correctly, it took quite a bit more than 1.425v to run it stable, so I never ran it that high for fear of sending it to an early grave. Higher CPU volts perhaps?

yep, its with a x6800. i can run it at full load and temp wont go past 40c (both cores) thats at 1.425v .
i thought 1.425v was way to high , but maybe i can increase it abit and see if i can hit 4ghz. i could go 3.8ghz but then id have to increase fsb , and dont want to do that.
ive tried to increase my 1066 modules to 1200mhz, they wont do it, they just crash every time. maybe i need to update to new bios. Jor3lBR i noticed has his 1066 modules running at 1200 mhz on this board, so it must be possible?

[Clunk]

Seems as thou the page is loading fine for me but the images are taking days to come through. **After a quick refresh it loaded fine... must have been something on my end**

I think I've figured this out but this example doesn't make sense to me and I'll illustrate with a screenshot I've taken.

So for example 1.100v VTT is 1.060v real. 1.060v x .67 = 0.7102 .

The First and second GTLs that are highlight are the closes match to .7102 but the NB is off a little bit. Wouldn't 0.707v be closer than the highlighted 0.716v? This is of course if my understanding is correct.

I've just finished talking to DFI.

The original value of 0.716V is correct. This is a real figure measured with a DMM, so that's why he said there is some variation from the 0.67% and also the fact that a half measure was used (75) instead of 70 or 80.

I think, because it is only one figure out of the entire set of tables, the best thing to do with that is try it and see. If 0.716V isn't working, try 0.7102V.

That brings me onto the BIOS that was promised to fix the wording for the GTL REF settings in the BIOS. They have double checked everything and the BIOS is correct after all (I think Praz said it was), so the tables have been updated again and as far as I can see, we are all fixed and good to go!

Sorry for any confusion

http://www.clunk.org.uk/forums/12436-post11.html

[Clunk]

Hey Clunk, any idea why my cpu voltage remains at 1.42v according to everest when the bios cpu voltage is set anywhere from 1.381 to 1.40? I saw your suggestion before about the special added voltage and changed that from auto to the lowest percentage available, and it makes no difference. I know that software voltage readings are not terribly accurate, but I don't think that would explain why it is "frozen" at 1.42v across all those bios settings.

Also, to anyone that has used 2x2gb ram (I'm using 2x2gb gskill 1066) on this or the dfi x38 board, what is the highest stable memory overclock you've achieved and at what voltages? I thought I was stable at 1128mhz, but after about a half hour of OCCT it craps out and I get errors.

Thanks for the help!

I'm not sure, I haven't noticed it sticking, but it does read higher.

When you set it to the lowest percentage, did you remember to add that figure onto the voltage?

[Jor3lBR]

I've just finished talking to DFI.
The original value of 0.716V is correct. This is a real figure measured with a DMM, so that's why he said there is some variation from the 0.67% and also the fact that a half measure was used (75) instead of 70 or 80.

I think, because it is only one figure out of the entire set of tables, the best thing to do with that is try it and see. If 0.716V isn't working, try 0.7102V.

And the other thing is, some people might have noticed that their VTT voltage values in the BIOS don't correspond to the second table. Please update your BIOS to 0307 or later and it should match.

That brings me onto the BIOS that was promised to fix the wording for the GTL REF settings in the BIOS. They have double checked everything and the BIOS is correct after all (I think Praz said it was), so the tables have been updated again and as far as I can see, we are all fixed and good to go!

Sorry for any confusion

http://www.clunk.org.uk/forums/12436-post11.html

Hey Clunk,

Why do we have 3 shaded values on the 1st row of the table instead of 2 (like in this post you have shaded green 75, 90, and 95)?

90 is to CPU GTL 1/3 Ref; 95 is for 0/2 and whats the 75 for?

[wilson]

Ok now ive had ASUS boards in the past so i dont know how to flash the bios on this DFI so any help would be great.
I got this of the DFI site but could someone fill the blanks pls
A:\>[executable] [file name].bin /F
OR A:\> [Executable] [file name].bin /F/QI
OR, A:\>[executable] /?

Cheers all

[drnick79]

Thanks for the reply Clunk. Even adding on the percentage from the special volts add doesn't explain the readings, unless I am missing something (happens quite often !).

Basically, these are the readings:

Bios CPU volts: 1.381 then everest reads 1.42v
Bios CPU volts: 1.3875 then everest reads 1.42v
Bios CPU volts: 1.393 then everest reads 1.42v
Bios CPU volts: 1.40 then everest reads 1.42v

It doesn't matter if the special add volts is set to auto or the lowest percentage...these values stay the same. If I am trying to set the cpu volts to 1.4 for example, I can do a workaround by setting the cpu bios volts to 1.35ish and then upping the special add volts, but I'm curious if you or anyone else is experiencing this "stuck voltage" phenomenon. My main concern is if this is some indicator of a faulty board, or if this will lead to instability when overclocking.

[drnick79]

Adya-

Just for a frame of reference, my gaming rig is a evga 680i mobo with a e6850 cpu. Not the x6800 as you have, but they are both 65nm architecture if I'm not mistaken. To run 4ghz with this cpu stable, I have to set the bios cpu voltage to almost 1.6v (yielding an everest cpu volt reading of about 1.55v). I'm watercooling, and temps are acceptable. But I don't know what most people feel is a "safe" cpu voltage ceiling for your cpu (if there is such a thing). I think that the 65nm cpus require more voltage (and are more robust) than the 45nm cpus. On another motherboard, my e8400 can run 4.05ghz stable at a bios voltage setting of only 1.3875v in comparison.

Perhaps others can chime in if I'm off base here.

[Clunk]

Hey Clunk,

Why do we have 3 shaded values on the 1st row of the table instead of 2 (like in this post you have shaded green 75, 90, and 95)?

90 is to CPU GTL 1/3 Ref; 95 is for 0/2 and whats the 75 for?

If you look at the first table, it gives you a value (a BIOS setting & real voltage) for CPU GTL 1/2, CPU GTL 0/3 and NB GTL REF. - Notice that in the first table, each of them has a different value. Scroll down to the corresponding VTT table for the voltage you wish to use and you will see the 3 set of values for each different thing 1/3, 0/2 & NB GTL.


That make sense?

[Clunk]

Thanks for the reply Clunk. Even adding on the percentage from the special volts add doesn't explain the readings, unless I am missing something (happens quite often !).

Basically, these are the readings:

Bios CPU volts: 1.381 then everest reads 1.42v
Bios CPU volts: 1.3875 then everest reads 1.42v
Bios CPU volts: 1.393 then everest reads 1.42v
Bios CPU volts: 1.40 then everest reads 1.42v

It doesn't matter if the special add volts is set to auto or the lowest percentage...these values stay the same. If I am trying to set the cpu volts to 1.4 for example, I can do a workaround by setting the cpu bios volts to 1.35ish and then upping the special add volts, but I'm curious if you or anyone else is experiencing this "stuck voltage" phenomenon. My main concern is if this is some indicator of a faulty board, or if this will lead to instability when overclocking.

In that case, I'm not sure why it's doing it then.

I'll see if I can make mine do it and if I can I will ask DFI later tonight.

Are you on the latest BIOS 0307 I think it is?