Official GIGA P35-S3/DS3/DS3R/DS3P/DS4/DQ6 bios/info thread

[MaStA]

Well last night I decided to remove all the copper heatsinks off my DQ6 and was shocked to see that the contact points where .
The P35 chip had decent contact but the TIM was rock hard. There was only 1 impression in the heatsink pad for the MOSFET modules around the CPU.

I had applied Artic TIM and also used some heat resistant o-rings on the nuts for the north bridge to add additional pressure. I felt there was too much movement for the block on the northbridge.
Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.

The standard TIM pads that are shipped with the motherboard obviously do not do what they are meant to do.

Hmm...so they use thermal pads and not a paste? Also, if you remove the pad and use a paste like AS5 will there be a gap since the space before was larger than what a thin layer of paste will be? My heatsinks always are hot.

Thanks,
Jared

[Wallynut]

Hmm...so they use thermal pads and not a paste? Also, if you remove the pad and use a paste like AS5 will there be a gap since the space before was larger than what a thin layer of paste will be? My heatsinks always are hot.

Thanks,
Jared

The northbridge was paste but MOFETS and Southbridge is a thermal pad.
I also had the same concern about contact so I did a test first by installing the copper then removing to see if everything is touching as desired.

Looked fine as the copper had TIM residue on it.

[MaStA]

The northbridge was paste but MOFETS and Southbridge is a thermal pad.
I also had the same concern about contact so I did a test first by installing the copper then removing to see if everything is touching as desired.

Looked fine as the copper had TIM residue on it.

Hmm...well there has been some discussion about how Gigabyte says the warranty is void if the crazy cool plate is removed. I imagine the same would apply if you replace the stock TIM with better stuff. That is one heck of a crummy warranty if they void it for those reasons...I just wonder how true it is.

Jared

[mackintire]

I asked for someone to look into some of the stability issues regarding Mushkin memory and the P35 chipset.

Calls have been placed into Intel for verification and to Gigabyte to discuss the nature of the problem.

The most obvious part is that standby S3 does not work correctly, but this same issue also spills over into the general stability of the motherboard when running the 1/2 clock divider. Hopefully we will end up with F6 shortly.

[superray]

Well last night I decided to remove all the copper heatsinks off my DQ6 and was shocked to see that the contact points where .
The P35 chip had decent contact but the TIM was rock hard. There was only 1 impression in the heatsink pad for the MOSFET modules around the CPU.

I had applied Artic TIM and also used some heat resistant o-rings on the nuts for the north bridge to add additional pressure. I felt there was too much movement for the block on the northbridge.
Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.

The standard TIM pads that are shipped with the motherboard obviously do not do what they are meant to do.

Same thing here with my DS3P
The thermal pad was rock solid
I took it off and used AS5 on NB, but it seems no big improvement on the temp, and the heatsink is not very tight now.

[Richard Dower]

LOL! Richard Dower - you may want to recheck your link. I got a video clip of a hot naked British chick. I liked it, but it didn't update my drivers. Lemme guess - you copied the link from the wrong browser tab?

How embarrasing!

That wasn't deliberate at all, here is the correct link:

http://www.realtek.com.tw/downloads/...&GetDown=false

I've trned into a bit of a "fan" of Big Brother's Amy

[zsamz_]

Same thing here with my DS3P
The thermal pad was rock solid
I took it off and used AS5 on NB, but it seems no big improvement on the temp, and the heatsink is not very tight now.

i replaced the springs heatsinks are verry tight

[Rilla927]

Hmm...well there has been some discussion about how Gigabyte says the warranty is void if the crazy cool plate is removed. I imagine the same would apply if you replace the stock TIM with better stuff. That is one heck of a crummy warranty if they void it for those reasons...I just wonder how true it is.

Jared

Are you kidding me! I just read this and went whaaa...

I just bought the GA-P35C DS3R and even though it doesn't have all that crazy copper that's like saying if I choose to replace the TIM on the NB it will void the warranty. Do other manufacturers do this? I finally get the last of my parts tomorrow so I can put this thing together.

[kimandsally]

Well last night I decided to remove all the copper heatsinks off my DQ6 and was shocked to see that the contact points where .
The P35 chip had decent contact but the TIM was rock hard. There was only 1 impression in the heatsink pad for the MOSFET modules around the CPU.

I had applied Artic TIM and also used some heat resistant o-rings on the nuts for the north bridge to add additional pressure. I felt there was too much movement for the block on the northbridge.
Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.

The standard TIM pads that are shipped with the motherboard obviously do not do what they are meant to do.

Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.


This is not good, if the heatsink is now only warm, it means that it is getting less heat into it than before, if it were hotter then it would be doing a better job of getting the heat from the bits on the motherboard.

Check your temps via software to see what they were before and now.

[MaStA]

Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.


This is not good, if the heatsink is now only warm, it means that it is getting less heat into it than before, if it were hotter then it would be doing a better job of getting the heat from the bits on the motherboard.

Check your temps via software to see what they were before and now.

Say what? If the heatsink is cool then that means it is not absorbing heat. If it is warm/hot that means it is absorbing heat. Your airflow will determine how cool it stays, but if all you did was replace the tim and your airflow remains the same then the heatsink getting warmer in my eyes would be a good thing. If there was say no contact between the chips and heatsink before and now there is contact obviously the heatsink will be much hotter than before.

Jared

[aspms*]

F6 final for P35-DS4- http://www.station-drivers.com/page/...ga-p35-ds4.htm

only F5 and up (including all betas) bios is showing right freq for me! With earlier versions if I set FSB 417 at post and in win it shows only 416 and if I set 418 it shows right- 418 no matter what multiplier I set! So I couldnt get exactly 417 and one more lower value (I cant remember but it was under 400mhz)!

[K404]

DQ6 owners- how is your FSB clocking? Im wondering is my DS4 is slightly handicapped to make the DQ6 look better.

I really should have checked my NB/SB contact as well- ill be pretty mad if it turns out the contact is to blame

thanks for the link aspms

[Rebel_DL]

I was wondering if anyone using the new 1333fsb C2D's E6750/6850 still has the double post/boot issue when overclocking using the P35 motherboards.

Whilst I recognise that the double post/boot issue with Intel Chipset boards is not considered a fault, I am concerned about any potential damage to ther components such as hard drive etc.

I currently have an E4400 with a P35 DS3 and the only way around the double post/boot issue I have found is to use either the 3.33 or 4 X ram multiplier. The only problem with using these multipliers is that when I start raising the FSB, I start to run into problems overclocking my RAM

Current settings are FSB = 285 X 10 (2850 mhz) RAM @ 3.33 = 475 mhz @ 5,5,5,18 / 2.0 volts.

The chip is capable of more than 3.2 ghz, but my RAM is unstable over 500 mhz - which is not surprising since they are only Corsair PC6400 DDR2 - using ProMos chips.

Obviously I would not be able to overclock a 1333 FSB CPU very far using such a high RAM multiplyer, if I run into double post/boot problems.

Any feedback would be appreciated.

Thanks

[Wallynut]

Normally I could put my finger on the heatsink and not feel much heat, now when I do this the heatsink is warm.


This is not good, if the heatsink is now only warm, it means that it is getting less heat into it than before, if it were hotter then it would be doing a better job of getting the heat from the bits on the motherboard.

Check your temps via software to see what they were before and now.

All my temps are fine and my memory is more stable now. In the past I have had issues running 4 x 1GB modules. I had tried adding volts to the fsb and MCH but no difference. Now I have had no issues at all and have actually been able to increase the modules Mhz.

After using a better quality TIM, the motherboard is more stable which means I am very happy with the board now.
Before I was getting pissed with it and was going to RMA, of course a RMA is now out of the question.

As for the warranty when you use a better TIM it will be void, Does this also mean Gigabyte does not support liquid cooling on the Northbridge ??

[K404]

Does this also mean Gigabyte does not support liquid cooling on the Northbridge ??

I really doubt it!!

-> Just flashed to BIOS F6- I seem to have Spi32M stability to 505 FSB instead of 497 ALTHOUGH I cannot directly put this down to BIOS revision- I was dice benching last week, so its possible the FSB ceiling is linked to that.

F6 doesnt seem much different to F5G, but no worse either!

[T_M]

Just for info, i bought a retail E6550 and P35-DQ6 yesterday.
Today i flashed from F4 to F5 and with exactly the same BIOS settings and gained at least 5fsb stability
This horrible setup is now about to be fully vmodded with a hope to gain way more fsb

[Lestat]

I'M BAAAAAAACCCKKKKKK
after many months away from this ever growing thread im back...

Got my DS3R yesterday.

running my X3210 @ 3.6ghz now.

i am dumbfounded and peeved that gigabyte still uses the retarde +2 +3. +4 etc etc voltage settings.

also even more appauling is the fact that you have to overvolt in the bios for vcore to get the proper voltage.

and not a little, alot.
1.56 = 1.48v retarded.

i also had a helluva time installing windows, infact probably the hardest time i have had in years.

lastly,,,, and again VERY shocked and pissed off that gigabyte did this sh|te again.

they are shoving all the key hardware items on the same fvcking IRQ's.
NIC's Sound cards, Video cards all need their own isolated IRQ god damn it!
gigabyte WAKE THE FRICK UP!!!


Richard -
long time no talk bud how ya been?

I'm on the F4 bios now
is there any overclocking or stability reason to go to the F6a ?????

[Synicalcop]

I'M BAAAAAAACCCKKKKKK
after many months away from this ever growing thread im back...

Got my DS3R yesterday.

running my X3210 @ 3.6ghz now.

i am dumbfounded and peeved that gigabyte still uses the retarde +2 +3. +4 etc etc voltage settings.

also even more appauling is the fact that you have to overvolt in the bios for vcore to get the proper voltage.

and not a little, alot.
1.56 = 1.48v retarded.

i also had a helluva time installing windows, infact probably the hardest time i have had in years.

lastly,,,, and again VERY shocked and pissed off that gigabyte did this sh|te again.

they are shoving all the key hardware items on the same fvcking IRQ's.
NIC's Sound cards, Video cards all need their own isolated IRQ god damn it!
gigabyte WAKE THE FRICK UP!!!


Richard -
long time no talk bud how ya been?

I'm on the F4 bios now
is there any overclocking or stability reason to go to the F6a ?????

What did you expect when buying a 130$ mobo
My DQ6 Run's my q6600 @ 3.6Ghz and my balistix @ 1200 5-4-5-7 2.2v without the slightess problem. 0.02 vdroop (4 x prime95)
I switched my soundcard so it wouldnt be on a shared IRQ just in case.

I built a q6600 system with a DSR3 and the vdroop and undervoltage is horrible

[irev210]

There is nothing wrong with having "rock hard" TIM. It is actually usually pretty good to use thermal pads for chipsets, etc since the surface of those heatsinks tend not be as flat. The thicker pad fills the void. Probably the most useful thing you can do is add some tension to the heatsink screws by adding paper washers or what have you.

When the pad heats up, it melts and fills in the empty space. It stays in a liquid state when above 40-50C and when it cools back down it hardens again. Perfectly normal function.

[LifeStealer]

I'm wanting to get closer to a 500fsb but I'm having trouble getting over 400 stable. I know 400 stable isn't anything to be depressed over but I would like to see it get a little higher.

Currently:
e6400: 3.2 @1.45v [Everest reports 1.41] (Don't want to go higher than this because I'm on air [This was purchased around Aug 06])
FSB: 400 @ +.1v
PCI: 100 @ +.1v (Not sure if I need this just upped it .1 for the hell of it)
MCH Overvolt @ +.1v
Team Elite 667 3-3-3-8: 800 @ 2.15v (3-3-3-8)
Temps: 44c idle, 52c load on a Zalman 9500 (This could be limiting? But the temps aren't that bad)

All told prime/orthos stable @ 2hours (Wanted to mess around and lower voltages before I go all night)

Thats pretty much it, regardless of how I change the multiplier or what voltages I put on FSB (I've gone up +.2v still no change) it hangs on Windows loading (It boots @420 w/x7 multi). The ram will go up to 550 @ stock voltages (Tested) so I'm not sure where the limiting factor here lies. I'm assuming that my e6400 just doesn't want to go over a 400mhz FSB. But I've fed it and been good to it

Also, I know 1:1 gives the best performance but if the ram will run 1200 should I change divider and let it run 1200?

As an aside I'm not sure I believe the temps that coretemp is giving me, bios shows 36c while coretemp shows 44-46c idle. Bios also shows NB at 26c which is close to ambient but speedfan reports 28-30c idle. I turned the warning on at 60c in the bios but nothing happens when I reach that. Although to be honest I'm not sure what the bios would do when it reaches that temp (Does it throttle?) but I see no change.

[zsamz_]

I'm wanting to get closer to a 500fsb but I'm having trouble getting over 400 stable. I know 400 stable isn't anything to be depressed over but I would like to see it get a little higher.

Currently:
e6400: 3.2 @1.45v [Everest reports 1.41] (Don't want to go higher than this because I'm on air [This was purchased around Aug 06])
FSB: 400 @ +.1v
PCI: 100 @ +.1v (Not sure if I need this just upped it .1 for the hell of it)
MCH Overvolt @ +.1v
Team Elite 667 3-3-3-8: 800 @ 2.15v (3-3-3-8)
Temps: 44c idle, 52c load on a Zalman 9500 (This could be limiting? But the temps aren't that bad)

All told prime/orthos stable @ 2hours (Wanted to mess around and lower voltages before I go all night)









Thats pretty much it, regardless of how I change the multiplier or what voltages I put on FSB (I've gone up +.2v still no change) it hangs on Windows loading (It boots @420 w/x7 multi). The ram will go up to 550 @ stock voltages (Tested) so I'm not sure where the limiting factor here lies. I'm assuming that my e6400 just doesn't want to go over a 400mhz FSB. But I've fed it and been good to it

Also, I know 1:1 gives the best performance but if the ram will run 1200 should I change divider and let it run 1200?

As an aside I'm not sure I believe the temps that coretemp is giving me, bios shows 36c while coretemp shows 44-46c idle. Bios also shows NB at 26c which is close to ambient but speedfan reports 28-30c idle. I turned the warning on at 60c in the bios but nothing happens when I reach that. Although to be honest I'm not sure what the bios would do when it reaches that temp (Does it throttle?) but I see no change.

try 510fsb with 6 divider
i did 510 x7 with my e6400
nb 2-3+

chek your cpu if its flat or else lap it
i never went above 50c with my old 9500

[LifeStealer]

Yeah on my old board (POS Asrock but w/e) core temps were reported to be around 35c~ but meh I dunno. I'll take your advice though and go ahead and lap it and the 9500. Its an OK cooler I guess, would install a tuniq but the damn thing is too big

[diablos]

Hi guys,
I've a DS3p, are there any mod (particullary the vdrop)?
thanks

[dnottis]

I'M BAAAAAAACCCKKKKKK
after many months away from this ever growing thread im back...

Got my DS3R yesterday.

running my X3210 @ 3.6ghz now.

i am dumbfounded and peeved that gigabyte still uses the retarde +2 +3. +4 etc etc voltage settings.

also even more appauling is the fact that you have to overvolt in the bios for vcore to get the proper voltage.

and not a little, alot.
1.56 = 1.48v retarded.

i also had a helluva time installing windows, infact probably the hardest time i have had in years.

lastly,,,, and again VERY shocked and pissed off that gigabyte did this sh|te again.

they are shoving all the key hardware items on the same fvcking IRQ's.
NIC's Sound cards, Video cards all need their own isolated IRQ god damn it!
gigabyte WAKE THE FRICK UP!!!


Richard -
long time no talk bud how ya been?

I'm on the F4 bios now
is there any overclocking or stability reason to go to the F6a ?????

Im using F5 official. Fixed performance issues with 2900XT for me. Not much else seems different. I hate the lack of a pencil vdroop fix. Missing my p5b-d...

[road-runner]

Not sure if its my board or what but machine crashed crunching in the middle of the night and now even at stock speeds I am getting so many errors I can not even get windows installed without errors at stock speed. I think linux said cpu soft error on the screen when I went in there yesterday morning and I have had no luck getting linux or windows installed, I tried all day yesterday. I tried a pci video card since the 2900 was hanging at boot, I tried a new raptor 150 hard drive so far. Must be the cpu ram or MB.

Edit: Update got a bad stick of G. Skill HZ....