Official GIGABYTE X38-DQ6/X38T-DQ6 Info/BIOS Thread

[Buggalugs]

Well i got 5-6 us idle, with some spikes to around 50 ....max was 360 when i just started it. So wouldn't be too sure about the x38 causing it.

OK thanks, thats really good Kojo...how do you set up your PC? Do u use IDE/AHCI or RAID bios?
Its not the X38 per se but running AHCI is more prone to bios corruption .

Asus have completely removed AHCI function from P5k because of issues and i suspect it is causing issues on our Gigabyte boards too.


@ Richard. From what i've been reading Gigabyte at first were blaming an Intel Southbridge spec as the cause of the issue and they cant do anything about it. More recently they have released a beta bios for some boards but they are definatley aware of it.

[Nuckin_Futs]

OK thanks, thats really good Kojo...how do you set up your PC? Do u use IDE/AHCI or RAID bios?
Its not the X38 per se but running AHCI is more prone to bios corruption .

Asus have completely removed AHCI function from P5k because of issues and i suspect it is causing issues on our Gigabyte boards too.


@ Richard. From what i've been reading Gigabyte at first were blaming an Intel Southbridge spec as the cause of the issue and they cant do anything about it. More recently they have released a beta bios for some boards but they are definatley aware of it.

Wow, this is new to me. I have only ever used AHCI and RAID wich is AHIC complient and have had very good results on my origional P5K Deluxe and current X38 based P5E Deluxe and MAXIMUM FORMULA SE, all on same ICH9R in both RAID 0 and back up on single disc RAID Ready AHCI mode with good results. I admit it is hard to get out of corruptin cause by crashing OS due to stressfull RAM overclocking, but on stable profile, it is fine. I used my P5K since last year on this day, and my Maximus since Oct. 10th.

Keep us posted, I am quite interested in learning more of this.

[Buggalugs]

Yep, i think we are going to hear alot more about this problem in the coming weeks.

[Kojo]

OK thanks, thats really good Kojo...how do you set up your PC? Do u use IDE/AHCI or RAID bios?
Its not the X38 per se but running AHCI is more prone to bios corruption .

Asus have completely removed AHCI function from P5k because of issues and i suspect it is causing issues on our Gigabyte boards too.


@ Richard. From what i've been reading Gigabyte at first were blaming an Intel Southbridge spec as the cause of the issue and they cant do anything about it. More recently they have released a beta bios for some boards but they are definatley aware of it.

I'm runnig a single 750 G F1 in AHCI mode (waiting for the new raptor ), I disabled all the stuff on the mobo I don't use, so one of the gigabit lan , gigabyte raid, audio chip, think that was about it.

[truehighroller]

I use the gigaraid and my bios is acting like it is getting corrupt. I can no longer hold my overclock at all.... I can't even pull 2.7Ghz... This processor has seen 4.15Ghz in it's six months for a bench contest and 3.6Ghz normal use 24/7. The same thing started happening with my GA-P35-DQ6,, I bet I can flash the bios and it will be fine again for a while.

[TyphooN]

I have been trying to figure out why my 3870x2 keeps stuttering(it is a brand new card, I just got it a couple of days ago). If I reboot and do nothing my latency stays under 500us.. After about 30 minutes of playing call of duty 4 I am getting spikes of about 3000-5000. What can I do about this? I am getting very frustrated with my stuttering.. It seems to be a problem with the DPC latency.

[Alcibiades]

Are you having getting your Q6600 to 500FSB? It is tricky but I got it there months ago on my P5K Deluxe and now testing on my ASUS x38 in sig.

Wow! 500fsb on Q6600 is Nuckin_Futs!!!
Getting close to 100% overclock from 266FSB to 500 (4x266=1066fsb quad)
Much easier to hit 500fsb on a 333FSB based quad (1333fsb quad)
The wall for most 266fsb quads is around 470, but still many have hit 500+ so maybe I'm missing something.

for mine at 1.480v I have to use PLL: 1.632~1.648v , FSBT: 1.488~1.504v, GTL_Ref of .63x on CPU and 63x on CPU.

Anyone know how to translate that into X38-DQ6 terms speak now.

As for the trfc you reported of 31, is that default of you set? Have you tried 45~50? For me it allowed me to run at much less vdimm of 1.72 to now 1.22v for same 1165MHz on a 5:6.

Default tRFC is 36, dropping to 31 gives good performance gains in everest benchmarks.

One thing I haven't been able to do on this board is run my ram at 4-4-4 latencies, even at standard relaxed settings @ 800mhz! which should do 3-3-3's easy.

[Buggalugs]

I have been trying to figure out why my 3870x2 keeps stuttering(it is a brand new card, I just got it a couple of days ago). If I reboot and do nothing my latency stays under 500us.. After about 30 minutes of playing call of duty 4 I am getting spikes of about 3000-5000. What can I do about this? I am getting very frustrated with my stuttering.. It seems to be a problem with the DPC latency.

You need to wait for a bios fix that Gigabyte is working on. 3000-5000 is really bad and will be noticable.

I suggest you take a screenshot of the latency checker program and email Gigabyte tech support with a few words that decribes your experience. IE Slow system, stuttering, audio dropouts etc.

With your card you might have the "powerplay" issue where it can downclock to low 3D mode while playing a game. Its a problem with ATI3870 and X2. The card has 4 power modes Boot mode,2D mode,low power 3D mode and Full 3D mode. There has been a problem where it can go from Full 3D to low 3D while playing causing stuttering. Some vendors have a VGA bios upgrade, or you can edit your bios so there are only 2 modes. If you go to your cards website/forum or the ATI website you can find more info.


Also some of you might see some error messages in Vista Event viewer that this latency issue causes.



You need to look in:

Event Viewer/applications and services log/microsoft/windows/diagnostics-performance/operational


You will probably see a bunch of errors, warnings etc about boot degradation, system degradation, etc.


@ Kojo. You are one of the lucky ones. From what i can tell it is very widespread affecting all Gigabyte boards since P35 and including X48.

[TyphooN]

Buggalugs, how do you know they are working on a fix? I really hope that they are. I am aware of the powerplay issues, but I have flashed my card to the same clocks/voltages across the board with computerbase bios editor. This is definitely not the cause of the stuttering. The stuttering isn't so bad when I first start gaming. Upon bootup I get between 100-800 us DPC latency. After gaming for about 10-20 minutes it gets very bad at 3000-5000 and above. I notice the stuttering when the DPC latency gets bad, so this must be the cause of my problem. I thought that maybe it was my OS... that is not the case, I just got done installing Windows Server 2008. I was using Vista SP1 and Windows XP x64 before.

How quick is gigabyte usually in responding to e-mail? I am going to game for a little bit and get a screenshot.

[TyphooN]

I got my screenshots faster than expected. Here are the 2 I am going to send to gigabyte. The first is me just using the computer for 10 minutes browsing XS. The second is after I have played call of duty 4 for 5 minutes.

[Buggalugs]

Buggalugs, how do you know they are working on a fix? I really hope that they are. I am aware of the powerplay issues, but I have flashed my card to the same clocks/voltages across the board with computerbase bios editor. This is definitely not the cause of the stuttering. The stuttering isn't so bad when I first start gaming. Upon bootup I get between 100-800 us DPC latency. After gaming for about 10-20 minutes it gets very bad at 3000-5000 and above. I notice the stuttering when the DPC latency gets bad, so this must be the cause of my problem. I thought that maybe it was my OS... that is not the case, I just got done installing Windows Server 2008. I was using Vista SP1 and Windows XP x64 before.

How quick is gigabyte usually in responding to e-mail? I am going to game for a little bit and get a screenshot.

Ok must be the DPC issue then. They will reply to your email in a couple of days but i'm not sure what answer they will give you.

However i know for a fact they have released beta bioses for some boards and they are aware of it and are currently working on it.

Thats why im asking anyone with the issue to send off the email so they know its widespread and they need to fix it properly.

@ Typhoon yours is really bad and must be affecting normal computer use.
Send the screen shot to gigabyte online support and tell them your computer is slow and unresponsive etc.

http://www.gigabyte.com.tw/Support/ServiceCenter.aspx

[TyphooN]

I sent them an email with both of the above screenshots, I hope they respond soon. Hopefully they respond with some sort of beta bios they are holding secret .

[Buggalugs]

They have only released a beta bios for a couple of the P35 flavours at this stage, but if we X38 owners complain enough we should get one soon too.

This really pisses me off. I have spent a couple of months on this and spent a small fortune on new ram and other parts that i didnt need to spend. This is meant to be a top of the range board...not good enough Gigabyte.

[Buggalugs]

Woohoo, I just got an email from Gigabyte with a beta bios for X38 to resolve this.

I'm gonna flash it now and report back.

[REVHEAD]

Woohoo, I just got an email from Gigabyte with a beta bios for X38 to resolve this.

I'm gonna flash it now and report back.

Care to share it?

[Buggalugs]

Ya its the F9b. Well i just tried it and it definately helps but doesnt completely solve it. The system overall seems faster. I'm gonna do some more testing but if anyone wants it , i put it on here:

http://rapidshare.com/files/110977377/x38dq6_F9b.rar

Remember this isnt a normal beta, its to resolve this specific issue at this stage so if you dont have a problem, maybe you shouldnt try it.

Please dont be lazy though, if you have a problem with DPC latency let Gigabyte know here:

http://www.gigabyte.com.tw/Support/ServiceCenter.aspx

So they will take it seriously.

Oh and a final Warning. This bios may make overclocking fail as it doesnt allow for Voltage increases.

[REVHEAD]

Ya its the F9b. Well i just tried it and it definately helps but doesnt completely solve it. The system overall seems faster. I'm gonna do some more testing but if anyone wants it , i put it on here:

http://rapidshare.com/files/110977377/x38dq6_F9b.rar

Remember this isnt a normal beta, its to resolve this specific issue at this stage so if you dont have a problem, maybe you shouldnt try it.

Please dont be lazy though, if you have a problem with DPC latency let Gigabyte know here:

http://www.gigabyte.com.tw/Support/ServiceCenter.aspx

So they will take it seriously.

Oh and a final Warning. This bios may make overclocking fail as it doesnt allow for Voltage increases.

Thanks Buggalugs

[TyphooN]

I hope they have one for the EX-DQ6 that allows voltage control ^_^. How long did it take for a response for you buggalugs? I sent my e-mail about 8 hours ago.

[Buggalugs]

I hope they have one for the EX-DQ6 that allows voltage control ^_^. How long did it take for a response for you buggalugs? I sent my e-mail about 8 hours ago.

It takes about 1 or 2 days.

[TyphooN]

How is that bios doing for you?? Did it improve your DPC latency?

[Buggalugs]

How is that bios doing for you?? Did it improve your DPC latency?

Well it improves it but doesnt completely solve it. They are still testing though.

A guy on another forum says his X48 beta bios completely solved the problem so hopefully we will get a resolution soon.

[Richard Dower]

New Intel Matrix Storage drivers en route:

http://www.station-drivers.com/forum...hp?p=8495#8495

Includes support for upcoming ICH10.

[Richard Dower]

Buggs, that was kinda quick of GB...to issue a fix that quickly....when are gonna issue a final F9?

[Richard Dower]

Drivers now available for download!

http://www.station-drivers.com/telec...ivers.com).exe

[Nuckin_Futs]

Wow! 500fsb on Q6600 is Nuckin_Futs!!!
Getting close to 100% overclock from 266FSB to 500 (4x266=1066fsb quad)
Much easier to hit 500fsb on a 333FSB based quad (1333fsb quad)
The wall for most 266fsb quads is around 470, but still many have hit 500+ so maybe I'm missing something.


Anyone know how to translate that into X38-DQ6 terms speak now.


Default tRFC is 36, dropping to 31 gives good performance gains in everest benchmarks.

One thing I haven't been able to do on this board is run my ram at 4-4-4 latencies, even at standard relaxed settings @ 800mhz! which should do 3-3-3's easy.

It really also depends on the chipset and BIOS tweaks to get 500FSB stable. I found all my olde 266 based C2D's did 500+ with propper BIOS options, especially on ASUS. (E6600, Q6600, E6300, E6400, 6750). E6300 and E6400 did best, w/ E6300 to 533FSB.

As for translating, please just PM me with a full BIOS settings template so I can make best of it to fill in. Gigabyte for my P35 based is not too clear on overvolting as it just offers +0.?? but never says what the starting point is for given iten (ie; MCH or PCI-E). Also please include the decoding for Gigabytes A,B,C,D strap values and RAM divider choices so I can better reply. In a nut shell, for mine (on P35 to 500x8) it required +0.2v on all major items. Just set and read from Everest.

As for the trfc, lower is even less stable for higher RAM spd and amount. If you have option of 50 or 55, try that instead.