** OFFICIAL ** Intel DX38BT & DX48BT2 BoneTrail Thread

[JohnZS]

I think it is more of a bug than anything, not sure actually what the symptoms are , other than stuttering/hitching and 0.1 lower score on Windows 7 experience index.
You are right, as far as I have been able to ascertain, Intel resolved the issue in July 2009, so all BIOS' for the DX38BT and the DX48BT2 motherboards released after this date (2000, 2006) do not have this issue.
ASUS on the other hand.... are playing catch up.
Hopefully this issue will be resolved soon.
John

[Sean-E-Boy]

dont think I will be of any help here...pretty sure my card is only pcie 1.1

[JohnZS]

Yeah that is correct. Sean, your Radeon 2900XT is a PCI-E 1.1 Graphics card.
As far as I am aware ATi moved to PCI-E with the 3000 series cards and nVidia with the G92 (8800GT, 8800GTS 512) cards and newer.
I have still not had any updates on this yet and am currently awaiting a nVidia PCI-E 2.0 card to test tomorrow.
John

[Manel]

Sorry Sean, didn't realize.
Anyway, it seems maybe the issue was fixed. Funny how it wasn't known at the time.
90 pages alreay!

[Manel]

You guys are awfully quiet these days. Nothing new to tell?
nVidia released new drivers, 196.21, will install them later today. See if the PhysX Control Panel issue is gone.
Wish they dropped the SLI issue already, it's silly doing it still at this time. Also, they should have GPU OC and temp monitoring as standard in Control Panel, instead of having to install that optional package.
Do you think there will be a new (final) BIOS for us this year?
There is a XTU v2.0 for X58 SmackOver and IDCC to go with it. We still don't have a final build of IDCC for our board. Intel does not even answer anymore. Guess that's it.

[redphil]

By the way: since I am using windows 7 (64bit) I can no longer use the IDCC. Any way to make it work on this OS?

[Manel]

redphil, I believe we're out of luck. I'm still on Vista, but the guys at Intel won't answer regarding new versions.
Doesn't it work on 7 then?
My opinion is that Intel really shines on hardware but when it comes to software it really sucks. IDCC looks good but it takes ages to get it right. Their applets look terrible, just look at the LAN driver tabs!!
Not sure IDCC for X58 runs on 7. If it does maybe we get it next, although I wouldn't bet on it.

[Manel]

redphil, tried running it under XP mode?

[redphil]

redphil, tried running it under XP mode?

In XP-Mode I can at least install idcc, but it won't start after installation. Just shows the loading-screen forever... .

[Manel]

Have you tried emailing Intel (idccfeedback) about this? Maybe they'll anwswer you regarding a new version that will run on 7.

[Manel]

Latest versions of IDCC are 7 ready for P45, P55 and X58 only. Emailed them again, waiting for the answer, if any.

[Manel]

nVidia released (well, not exactly but it's available) 196.34 beta drivers to correct the OC bug in WHQL'ed 196.21 drivers recently released. I'll wait for the WHQL version to install.
AMD has also 10.1 drivers out, they almost missed the January launch.

[Manel]

Intel's answer is that they don't provide info on unannounced products and future technologies.
So, IDCC for 7 is something they won't talk about since it's either one of the mentioned above!!
Not only are these guys slow developers, they are also very informative. I can even understand this, what I fail to understand is this lack of commitment in development, from those who want to capture the enthusiast crowd's attention.

[paradigital]

I think I might have cracked it guys, not 100% on the method to install yet, but once I work out exactly what procedure to take I'll be sure to post it up (working on it now):



Q6600, DX38BT, Windows 7 x64 (Professional)

[paradigital]

Ok, I can confirm that I have IDCC installed stably on Windows 7 x64 Professional, on my DX38BT. It allows me to make all changes that I have tried to make (FSB, Voltages, Memory Timings, etc). The way to install is as follows:

1) Install IDCC v4.2.0.757 (change IDCCSetup.exe to run in "Vista Service Pack 2" compatibility and to "run as administrator")
2) Find "Intel(R) Desktop Control Center.exe" (typically in %programfiles(x86)%\Intel\Intel(R) Desktop Control Center), and change it's compatibility to "Vista Service Pack 2" and to "run as administrator"
3) Reboot
4) Run IDCC, it should load, but fail to supply any data or allow any changes.
5) Uninstall IDCC from Control Panel
6) Reboot
7) Download IDCC for DP45SG (v4.3.0.1222 beta for Windows 7 x64) from http://downloadcenter.intel.com/
8) Install IDCC v4.3.0.1222, installer will fail, and roll back changes
9) Reboot
10) Reinstall IDCC v4.2.0.757
11) Reboot
12) Enjoy working IDCC

[crazysn2ke]

I can confirm that this works. Installed A-ok, all information and tempreture/voltage/fan speed appears to be accurate, so way to go paradigital !

[redphil]

Wow. Thanks, thanks, thanks.

It's really working on my system. Afters months of desperate waiting, idcc on my win 7-System.

[paradigital]

Can anyone tell me what is a "safe" maximum voltage to put through the MCH and FSB?

Just having a little play with bus speeds and max clocks I can get out of this board. Currently running 425MHz FSB, and 3.82GHz out of my Q6600 (G0).

Cheers.

[Sean-E-Boy]

paradigital,

First, any voltage over stock is going to increase the rate at which the particular component degrades and ultimately fails.

That being said, the type of cooling you have on the particular components can help to counteract this degradation. So, what type of cooling do you have on your processor and MCH?

The MCH on the x38bt, and presumably the x48 can get VERY VERY hot, but seems to be rather "bulletproof," if you will, with regard to voltage...it can take the max, so long as you have some real good cooling (i.e. water) going on. However, increasing the voltage here only helps you to increase your memory frequency at a particular FSB and Multi or maintain your memory frequency configuration as you increase your FSB or Multi. Also is useful when decreasing memory timings (to an extent).

and I have said this before, but DO NOT get caught up in "I must have my memory running 2:1 (meaning that the resulting memory frequency is the same as the resulting FSB frequency) with my FSB” the fact is that under most situations, using a slightly lower memory setting, say 5:2, where your memory frequency is lower than your FSB can make overclocking the processor much easier and you only take a minor hit on performance, if any. while memory speed does help to improve memory performance, increases in FSB frequency not only make the processor faster, but also improves the bandwidth efficiency of your memory, making it faster too.

at 2:1 it is unlikely that you will be able to run your FSB any higher than about 380MHz, if using 2 dimms, maybe 395 without stability issues, but if you drop that to 5:3, the internal latency of the MCH increases and the memory frequency decreases, allowing for stability at lower MCH voltages.

Now for the FSB, the stability of EVERYTHING is tied into the voltage you set here. Think of the FSB as the main voice of the system, the clarity and volume of this voice get carried over into the stability of the cpu cores and the quality of conversation between the FSB and the MCH. Somewhere earlier in this thread I wrote a little about GTL ref tuning. If you really want to either push your system to the max or tweak it so that (potentially) less voltage is needed for stability, I strongly urge you to look it up and read it.

In the mean time, the gist of it is to start with the FSB. So at stock voltages and stock FSB, drop your multi to its lowest and your memory to the slowest. then start increasing the FSB (performance will be lousy, but in this process you are trying to make the FSB rock solid, increasing the multi and therefore the GHz will come later.

Continue increasing in say 20 MHz steps until your system fails to boot or crashes or displays some sort of instability. When you reach this point, increase the vFSB. Once stability has been restored, increase the FSB. Repeat this process until you reach either the FSB for which you are aiming, or the highest vFSB that you would feel comfortable running 24/7, etc.

I strongly believe that the FSB in most of these processors can, under proper cooling, handle the max voltage...but it is your hardware, so if you cannot afford a burned out CPU you may want to think twice. Now also consider that it is unlikely that the combination of vFSB and FSB speed that you end up with at the lowest multi will be where you are as you increase the multi. Now while you are testing things a few minutes or an hour at a time, it is unlikely that you will cause massive damage through voltage.
In the end, you will likely be at a lower FSB and at that point can reduce the vFSB accordingly.
The next step is to start increasing the multiplier of the processor. If you keep note of FSB and vFSB as you go through the previous step, you will know the maximum FSB at X vFSB that your processor can handle. As the multi goes up, you will never be able to get your FSB any faster since the strain on it increases.

So keeping voltages where they are drop your FSB by 50MHz or so and increase the multiplier by 1 step. If stable, slowly increase the FSB again back towards the max found in step 1. If not stable, add some vCore. Repeat this process and keep note of resulting CPU GHz and FSB speeds as well as voltages. Once you hit a wall and don’t feel comfortable increasing vCore, drop the FSB some again and increase the multi by a step.

At some point, you will probably need to add some vMCH to keep up, but remember to change 1 thing at a time and keep track of what it was.

So, once you get up to your max multi (note that 1 step down may result in faster CPU speeds with a higher FSB) you can play with memory.

A few notes: since all of these things are interrelated or at least dependent upon some other component, changes in one area can mask faults in another. For example, increasing vCore can mask instability with the FSB. In other cases, increasing the FSB can cause stability issues not correctable with vFSB, and this is because increasing the FSB also increases your memory frequency adding to the load on the MCH, in which case you would either need to lower the frequency of the memory, reduce timings, increase the vMCH, or all of these things.

This should give you a good start. It is a long road filled with frustration...as has been said before, "this board is just like a woman" most of the time it wont do what you want it to because you havent caressed it properly, and other times it just doesnt want to do what you want it to, but her attitude may be different tomorrow and you just might get lucky...

I know that I am long winded, and for that I apologize, however there are many many hours of experience crammed into this post and those of others, so try to absorb all you can (and in some instances just take our word for it!).

Sean

[Manel]

Sean, as usual you're right on it. Do you ever find the time to sleep? You're wife hasn't kicked you out yet? Just joking.
Nice detailed procedure as always.
By the way, which were the devices in Device Manager that need to be manually updated with newer INF drivers? Which are not automatically updated once you run the installer?

[Sean-E-Boy]

last time i dealt with the inf stuff, nothing updated either automatically or manually...i assume that this is because the inf package supports the x58 boards too and the update was not for us.

as a caveat, the directions below are somewhat from memory, as i am not on the x38 pc right now.

but, in the case where the inf will update our system, it seems that under 32 bit windows, the installer updates everything fine and no manual update is needed. On 64 bit windows nothing, or at least very few things, update automatically. you will need winrar or 7-zip to extract the inf package to a known place (desktop is easy), then go to device manager and you need change the view through the menu at the top to view by connection. then expand acpi x64 based pc, then expand microsoft acpi compliant, then expand pci bus.

you will see things such as the pci-pci bus or bridge, x38-pci (or something like this) bus or controller as well as the sata controller, etc.

as for what to update manually? it depends on what they updated in the inf (usually the usb stuff updates, but periodically others will too). i usually just try to update each major item under pci bus and do not expand further. then just right click on the item, choose update, then browse for driver software, then choose the location of the folder you extracted and make sure to check include subfolders.

sean

repeat for each item.

[BoneTrail]

Hello everyone.
Intel has released The latest version of IDCC for W7 and Vista: IDCC x38/x48 W7/Vista

I am awaiting The Audio IDT driver with Dolby Digital for Windows 7.

Regards.

[Manel]

Beat me to it BT.
Yes, Intel has posted IDCC v4.2.0.769 final (finally), after so much pressure from us all - coincidence maybe or do they actually read this thread or even hear/read our emails?!
redphil/paradigital, no more work-arounds needed, this time it's 7 compatible. You need to install to confirm of course but should be a sure thing.
Great Intel, good work.
Readme refers a Iron City (which has been in other boards docs as well), which I believe is for mobile. Documentation seems to be general, not X38 specific. Very informative now!!!
Dated already 4.Jan.2010 or 5.Jan.2010 depending on where you read it. Typical Intel, taking this long to release it. I understand that they take their time to fully test before release, we do it too in the company, but I guess we get itchy with new stuff.

Thanks Sean, I had that recollection as you described it, just checking.

[Manel]

BT, IDT drivers are also a pain to come out. IDT seems to provide updates quite often, Intel doesn't always release them to the public though. It seems at least by the documentation.
I'm still on Vista but there are a few things I'd like fixed in the drivers, latest 20001.
Dolby is still locked for 7, maybe soon they release new drivers.

[redphil]

Finally... .