** OFFICIAL ** Intel DX38BT & DX48BT2 BoneTrail Thread

[down8]

Newest INF doesn't fix it. Temps seem to be good now - great actually (TjMax is 100*C, right?) - I'm seeing 20sC even.

Is anyone else having issues with BIOSv2000? I am thinking about going back to 1997. Too many strange lockups and such.

Still can't get to 1600 on the RAM. Ideas?

-bZj

[Sean-E-Boy]

Down,

Good to see that someone is still trying (or at least has time) to push this board.

Couple things:
the unknown device in the device manager, where is it located in the device tree? your processor should appear on the page with everything collapsed...has its own branch. if that is where the issue is, i would contact intel. you can also get an unknown device listed elsewhere if you have Away Mode enabled in the BIOS.

400X8.5 should be a breeze for the 9550, and should require only minor vCore bumps to be rock solid.
However, 1600Mhz on the RAM with 4 dimms @ 1:2 is going to be your issue. to get anywhere close you are going to need every ounce of vMCH that the BIOS will let you provide and your timings loose, especially the Command Rate...and even then i have serious doubts that you will hit 1600 stable. I would suggest that you drop out of the 1:2 mem ratio to the next lower one (3:5 400/1333) and boost the CPU host clock. There is plenty of room in the 9550 to take you to 470~480 x8.5 or 3.995~4.08 GHz especially considering that youre on H20, and 480x8.5 @3:5 is 1600MHz on the RAM, just allows for looser internal timings of the MCH, but shouldnt affect your ability to tighten the RAM timings very much.

try to not get caught up the in the numbers game with RAM speed. Run some tests (everest or encode a video or something) at 380 FSB or so with 1:2 memory and take note of the performance numbers or processing time and then drop to 3:5 memory and bump FSB to 420x8.5 and run the same tests and compare. I think you will see that in most real world applications (>90%), increasing the FSB (which increases the CPU Freq, the CPU L1&2 Cache response and transfer abilities, the read/write performance of your HDs, and your video card performance) will more than offset any loss realized by the reduced RAM speed. And on top of that, a good % of the code executed in most programs once they have loaded will fit in the L1 and/or L2 cache on the CPU and not even touch the RAM.

Hope this helps

Sean

[GumaLive]

Down8 I had similar issue.

I was running my system well,
CPU 2,6 -> 3,2 GHz; FSB 1333 -> 1600 MHz; RAM 1600MHz

Then I started experimenting with Intel Integrator Toolkit.
I wanted to flash SMBIOS with some additional information. I booted Windows98 boot disk on flash disk and used ITOOLKIT.EXE to add information:

; iToolkit INI file
<< cut >>
[System]
INIversion=2.9
[SMBIOS.Values]
<< cut >>
ChassisManufacturer=Thermaltake
ChassisVersion=VH6000SWA Armor+
ChassisSerialNum=123412341244
ChassisAssetTag=VH6000SWA 1234124314

But after that system became very unstable. I also got 1 random reboot. And many troubles. CPU sensor "disappeared" from Everest. Also in BIOS there was no CPU temperature info. And CPU fan was running at 1500 RPM (maximum) instead of 600 RPM (usual).

I deleted all changed values from BIOS and issue seems to be resolved, however the system is not that stable as before. And there is message "Intel Integrator Toolkit has modified this BIOS" in BIOS -> Additional information.
I'll do more experiments later.

[down8]

Down,

Good to see that someone is still trying (or at least has time) to push this board.

Couple things:
the unknown device in the device manager, where is it located in the device tree? your processor should appear on the page with everything collapsed...has its own branch. if that is where the issue is, i would contact intel. you can also get an unknown device listed elsewhere if you have Away Mode enabled in the BIOS.

400X8.5 should be a breeze for the 9550, and should require only minor vCore bumps to be rock solid.
However, 1600Mhz on the RAM with 4 dimms @ 1:2 is going to be your issue. to get anywhere close you are going to need every ounce of vMCH that the BIOS will let you provide and your timings loose, especially the Command Rate...and even then i have serious doubts that you will hit 1600 stable. I would suggest that you drop out of the 1:2 mem ratio to the next lower one (3:5 400/1333) and boost the CPU host clock. There is plenty of room in the 9550 to take you to 470~480 x8.5 or 3.995~4.08 GHz especially considering that youre on H20, and 480x8.5 @3:5 is 1600MHz on the RAM, just allows for looser internal timings of the MCH, but shouldnt affect your ability to tighten the RAM timings very much.

try to not get caught up the in the numbers game with RAM speed. Run some tests (everest or encode a video or something) at 380 FSB or so with 1:2 memory and take note of the performance numbers or processing time and then drop to 3:5 memory and bump FSB to 420x8.5 and run the same tests and compare. I think you will see that in most real world applications (>90%), increasing the FSB (which increases the CPU Freq, the CPU L1&2 Cache response and transfer abilities, the read/write performance of your HDs, and your video card performance) will more than offset any loss realized by the reduced RAM speed. And on top of that, a good % of the code executed in most programs once they have loaded will fit in the L1 and/or L2 cache on the CPU and not even touch the RAM.

Hope this helps

Sean

Sean-e-boy, you're a lifesaver!

I am now at 3.83GHz/1500MHz RAM. vCore @ 1.28, vFSB @ 1.3, vMCH @ 1.3, vRAM @ 1.66[?]. I may back down the vRAM, and start trying to tighten up the timings. Think there's any room to lower the other voltages?

450MHz works for now - I tried 455-470, and couldn't get into Windows. I did downgrade to BIOS 1997 - v2000 just seemed unstable.

Idling high-30*Cs, load low-50*Cs.

Thanks again!
-bZj

[Manel]

Major facelift at Intel download center. Still, nothing new.

[redphil]

I had time for a little bit of ocing. Installed my watercooling again and pushed the Q9300 close to 3,4 Ghz.



For now I am not able to get more then 450 mhz host clock stable (even with lower multiplicator). What do you think about more voltage?

More annoying: I am using 4 1 GB sticks of Crucial Ballistix Tracer DDR3-1333. But can't get them stable above 420 mhz host clock. So I kicked two of them out of the system for overcklocking. Any chance to use all four of them @450 mhz?

[Sean-E-Boy]

Redphil,

Most likely you are running into voltage issues with the FSB and possibly the vCore. With H2O cooling on the CPU you can pretty safely max out the vFSB (although you shouldnt need to). For the quads, most of the instability comes from the vFSB and early on small bumps in vFSB will give you 10-15MHz on the host clock, but once you get to around 465 or 470 the same voltage increase will only give you 2-5Mhz on the host clock. I would suggest 1.5 on the vFSB (do this one first) and 1.4-1.5 on the vCore. increasing the vCore will almost always mask any instability found with the FSB, so dont increase them both at the same time...FSB first, then once you hit a wall where you are no longer comfortable increasing the vFSB, turn up the vCore in small steps.

As for the ram, 450 on the host clock @ 1:2 memory divider results in 1800Mhz for the ram. Now since you are at a lower multi (7.5) compared with the quads used by others (~8-8.5), you may be able to run all 4 dimms @ 1:2 up to about 430-440, but i seriously doubt that you will get to 450 or higher at that ratio (400x1333 is a much easier ratio, 3:5, to work with than 1:2).

if it is true that the Q9xxx's are all really the same with the exception of the multiplyer, reaching a host clock of 485 shouldnt be a problem, now getting stable there @ 7.5x will be difficult, but possible.

Sean

[redphil]

@Sean-E-Boy: Thanks for the advise. I tried 400/1333 with 2GB and lowered Multi (7), VFSB@1,45v. Wasn't that lucky. Reached 465 host clock, 470 was no good.
I then tried 400/1333 with 4GB, and after some starting issues it works quiet well. I am now on 455 host clock, that's a nice 3,4 Ghz on the CPU and 1516 mhz for the RAM.
Now running stabilty tests, if this will work then I'll try to lower voltages and to enhance timings on the RAM (now 9-9-9-24).
I am really satisfied with the results. It's nice to use all of the RAM-modules on this high clocking settings.
Thanks again @Sean-E-Boy for your great support here.

[Sean-E-Boy]

no problem, and remember that whenever you are messing with the RAM...changing timings, adding or removing dimms, changing ratios...this board gets upset. you may need to completely power down or even unplug the PC and restart several times to determine that those settings will not work.

I recall times having to power cycle 5 times just to see POST messages, and then everything was fine. So just because it wont boot in the first 2-3 times doesnt mean that it wont, it may just not want to right then.

sean

[GumaLive]

I was able to induce the issue with Intel Integrator Toolkit.
I booted Windows 98 rescue disk on flash disk, typed command:
itoolkit install -smbios -ini=file.ini

Command completed sucessfully, but after reboot again huge system instability+random reboots.

Running command
itoolkit remove -smbios
Resolved all issues and system became stable. I think I'll contact Intel tech support about that.
Manel do you have some idea how to contact Intel tech support? Is it better to call or email?

I will also test WindwsPE and Windows GUI utility. I want to flash different boot picture.
Also I'd like to get some info how to completelly reset the bios. Reflash with original BIOS+reset setting.

[Manel]

Gumai replied to this
in the past i used the support@intel.com support email, and they replied with the crmotherboard@mailbox.cps.intel.com address and I replied to this one and they answer as well.
Try it, it might work.

[Manel]

Guma, sorry man, that was some answer, he?! I was in a hurry and something went bad.
Guess you understood though, right?
Did you try any of those? Haven't used them for a while now.

[Manel]

Those guys at Intel are way sloooow, IDCC should have been final already for a long time but it seems something went wrong, I don't think they have the best coders anyway. Can't remember now but I believe it was June or July the release schedule. Guess just forget it.

[Manel]

Intel Extreme board DP55KG looks nice, really nice indeed. The lit skull is a nice touch, although I don't really fancy those kinda things. funny it's a DP board, not DX and still Extreme.
IDCC v5 for it has the same skull.
Dolby control Centre 2.2.1 looks good too, how come we don't get a new version? Still Pre-production though.

[Patyabc]

A new Inf file was posted but I didn't have success cause the dl page goes to a "return page".

Hope this could be fixed in any moment.

Sean e-boy, have you had a good news from Intel?

Intel DX48BT2 BoneTrail2
Intel Core 2 Quad
Mushkin DDR3 1600 Dual Channel 2x2Gb
Sapphire 4870 GDDR5 512MB
Be Quiet 650W

[Manel]

Paty, this is common for Intel, at first new dls never work. Still down right now.
Also, LAN drivers v14.5 out, but with the same problem.
Same old Intel.

[GumaLive]

Hey guys I can confirm that Chipset and LAN drivers are relased, but not possible to download.
Do you think there is some flaw in drivers and they removed them? Or they just not uploaded them?

[Tranzmit]

use this site for the VERY LATEST drivers. Its a french site but its got english links and its readable or use google translate if you need it

http://station-drivers.com/

[Manel]

Guma, this is typical for Intel, the very first days you just can't dl at all, go figure. This has happened for all the latest releases.
Still, I was able to dl in the morning so it should be fixed now.

[Patyabc]

Dear Friends:

If you got Intel new updates, you will be dissapointed cause these updates apparently were made for x58 MB's, and you will get a window warning you about no possibilities for installation in your PC , what is happenning with our friends from Intel?

P.S.: This issue is the same with the last update for my Network Desktop Adapter GT Pro 1000 .I got no Windows 7 Driver for my Intel GT card, well, hope in the near future I will.

[Manel]

Paty
What do you mean new updates? Maybe the INF? LAN? LAN drivers do not depend on the chipset really, only the network chip (PHY) used.
Haven't tried any of the latest but not sure why they shouldn't work though.
If the new INF won't work with X38/X48 why is it available in our dl page?! Well, Intel does screw up sometimes.

[Manel]

INF v9.1.1.1020 is intended also for our boards, so is LAN v14.5. Will try to install in the weekend or so.
The way I see it the SmackOver differs (apart from the obvious chipset) from our board, hardware wise, in the audio codec used, not much else. That way, BIOS and audio drivers will not be the same but the rest should work for all boards.

[fzfq3m]

hi there guys

can somebody explain me a bit what would be the order in wich i should encrease voltages to gain stability on this mobo while overclocking? Should I encrease CPU Voltage then FSB Voltage and MCH/ICH Voltage at last?? I´m a bit confused about FSB Volatge and MCH/ICH Voltage... I mean, MCH/ICH Voltage is the voltage of Northbridge and Southbridge Chipsets (I read somewhere that they are linked), but I´m not sure what FSB voltage is applied for. Found this http://www.xtremesystems.org/forums/...d.php?t=147164

BTW I did some research about Max voltages recommended from manufacturer (Intel) for safety/daily operation, but as far I know in order to get a high overclock I need to encrease the voltages beyond those values, but how high can I go without damaging my mobo??

the max voltages I found on Intel site was

CPU Voltage: 1.5v for Q6600
FSB Voltage: 1.3v
MCH/ICH Voltage 1.375v

Currently I´m trying to get my rig stable at 3.8Ghz (422*9) using the following voltages:
CPU Voltage: 1.52v
FSB Voltage: 1.375v (2 or 3 steps beyond max)
MCH/ICH Voltage 1.425v (2 or 3 steps beyond max)

with this settings my systems fails on Prime95 small FFT test on about 10 seconds, and I must say that is not a matter of temperature because I'm got my rig water cooled... and memory isn´t the problem neither because I'm using 400 of Mem ref and 800 as mem speed...and I also checked with memtest86+ (2 full test) and didn´t find any memory errors..

I really think that my rig needs a little more juice on voltages but after getting voltages this high i'm affraid to get further without ask...

Advice Please

[Sean-E-Boy]

fzfq3m Welcome!

Now i know that this thread has some 2000 posts in it, but much can be learned (and therefore many hours of frustration avoided) by starting at the beginning and just reading through it.

but more directly to your question, read this post:
http://www.xtremesystems.org/forums/...postcount=1754

then this post:
http://www.xtremesystems.org/forums/...postcount=1765

and dont get owerwhelmed with the GTL talk, it is just another setting in the BIOS that you can change to attain greater performance...but these posts should give you an idea of what steps to take and when.

The biggest thing to remember is that successful overclocking takes a fair amount of time, and the more you try to rush it the less successful you will be. now that doesnt mean that you wont get any performance gains, but it is just like about everything else: there is a quick way and then there is the right way.

If you dont have 2 weeks of time to figure everything out, let us know and we will try our best to get you where you want to be as soon as possible.

Which is where, exactly...any ideas on FSB MHz or processor GHz?

Also, dont believe what intel says about voltages, for the most part their chips are built like Sherman tanks.

another note, I used to have a Q6700, no H2O though, but it seemed to top out stable around 375 FSB (also prior to understanding and tuning GTLs), so not so sure how feasable that 3.8 (422x9) is.

Sean

[fzfq3m]

Thanks Sean-E-Boy

I printed the posts for further reference...
I still reading through this thread but as you say it will take me some time.

Anyway I just can´t get my hands off my rig... last night i was able to gain a little bit of stability at 422*9 by bumping vCore up to 1.55v, vFSB at 1.4v and vMCH/ICH up to 1.45v... I'm still getting BOD a while after firing Prime´s small FFT test but with this settings Prime was operative for longer time than before... b

I found that enabling Enhanced Power Slope(EPS) reduces the amount vDroop of the board, without EPS my mobo show almost 0.09v of difference between idle and load vCore but with EPS the difference is just 0.05v

Once again thanks for the links, I will give GTL a try but following are my specs in case you got more advices

Hardware:
Cpu: Intel Q6600 G0
Mobo: Intel DX48BT2
Mem: 4x1GB Corsair XMS3 PC3-12800 DHX (CM3X1G1600C9DHX) 9-9-9-24 2T 1.8v
PSU: Corsair TX-750w Single 12v Rail
Vga: MSI Geforce 9800GTX 512MB OC edition
THE COOLING SYSTEM:
CPU: Swiftech Apogee GTZ
Chipset: Swiftech MCW30
VGA: None
SB Chipset: stock passive HS
Pump: Swiftech MCP655
Radiator: Black Ice X-Flow GTS Stealth 360
Reservoir: Swiftech MCRES-MICRO v1
Rad Fans: 1x 2200rpm (cfm not know), 2 x NIDEC TA450DC 110cfm
Thermal Compound: Artic Cooling MX-2 on both cpu and chipset
Tubing: Tygon 3606 1/2" ID (Don't remenber OD)
Coolant: 2 x 2oz HydraX mixed with 1l Distilled Water

Current settings:
Fsb: 422
Cpu Multi: 9
vCore: 1.55v (on Windows everest reports 1.46v idle - 1.31v at load)
CPU voltage Offset: Disabled
Enhanced Power Slope: Enabled
vFSB: 1.4v
vMCH/ICH: 1.45v

Mem Ref Clock: 400
Mem Speed: 800
Mem Timming: 10-11-11-28-2T
vMem: 1.82v

regards