** OFFICIAL ** Intel DX38BT & DX48BT2 BoneTrail Thread

[Manel]

Hey red, you're online.
Have you tried it yet? Guess you didn't have time yet.
Intel makes me laugh sometimes, I was reading the readme inside the zip and it's dated 27.May.2009, they haven't updated that file yet. Careless. Still mentions XP, which is not supported anymore in this new version. All the bugs are still the same. All documentation is outdated, even for the newer version of IDCC for P45 which was released by the same time.
Saw somewhere mentioning XTU v2.0, don't know where now. Not sure if it's a new version of the service or if it's already used in our version of IDCC.
Anyone tried the PassMark BurninTest and Performance test? Downloaded but never installed.

[redphil]

Hey red, you're online.
Have you tried it yet? Guess you didn't have time yet.

I gave it a short try. It's working, but a bit sluggish. Maybe I should have de-installed the older version before the installation.

[GumaLive]

Amazing that Intel released final version of Desktop Control Center. Love you guys :-)

I have tested my DX48BT2 with 8 GB of memory:
2 x 2GB OCZ DDR3 PC3-12800 Intel Extreme Edition (OCZ3X16004GK)
2 x 2GB OCZ DDR3 PC3-12800 Platinum Enhanced Bandwidth Edition (OCZ3P1600EB4GK).
And it is definitivelly not working. Again random reboots. For sure it is causing OCZ3P1600EB4GK.
- I had also tryied to decrease timings and everything. It POST on 1333 FSB, but random reboots. On 1600 not POST.
Anyway now I'm running with 4GB OCZ3X16004GK and it is running perfect on 1600 FSB.

I'm interested if new DDR3 memory (lover voltage) would work on this mobo with 8GB.
- Did someone of you sucesfully tested 8GB on 1600 FSB?
(As I undesteand it... My modules needs 1,9 V. New modules(OCZ3X1600LV4GK) needs only 1,65V. Does it mean less voltage -> less buss interference -> running 8 GB?)

[Manel]

Hey Guma
Did you loosen timings with all 4 sockets populated? Or tried lower mem freq, just to test?
Since they're different mem sets, didn't check, are voltages the same for both kits? 1.9v seems high, mine are 1.8v, but I guess there are higher voltage sets. Lower timings to 9-9-9.
I also wanted 8GB with low voltage, newer 1.65v kits, or the ones to come with only 1.5v or even 1.35v. DDR3 should be rated 1.5v anyway.
8GB would be sweet, don't know if anyone could have it at 1600 though.

[Sean-E-Boy]

- Did someone of you sucesfully tested 8GB on 1600 FSB?
(As I undesteand it... My modules needs 1,9 V. New modules(OCZ3X1600LV4GK) needs only 1,65V. Does it mean less voltage -> less buss interference -> running 8 GB?)

Guma,

now take this with a grain of salt as i do not have 8 gigs to check, nor an x48, but have tried 4 dimms before...I believe the real issue is with the MCH and its ability to handle the throughput of all 4 dimms at 1600mhz, and by 1600mhz i am talking about the resulting memory frequency, not the 1600 in 1600/400. I was never able to get all 4 dimms, whether paired 1gigs or 2x2 + 2x1 to run stable any faster than about 1530 or 1560, and it takes every bit of the vMCH to get there.

as far as the memory voltage goes, and i am not certain, but i think it simply deals with how much voltage they require for operation, nothing more. so 1.5 dimms compared to 1.9 may just have more efficient circuitry--kind of like going from 65 to 45nm with the CPU. I seriously doubt that there is any less interference or signal distortion because of the required voltage.

sean

[Sean-E-Boy]

UPDATE

Ok, all this talk of memory stuff got my gears cranking again so i decided to mess around some today to test a few things out.

First of all, these tests were conducted with the vMCH and vFSB at the max and it should also be noted that my system has had GTL Ref tuning done. Additionally, my vCore was set at my max safe, which can take my Q9550 to almost 4.2Ghz at 8.5x. And just to be thorough, i know that my FSB at 6x can run to 505 before things get wacky, and the memory used (2 different pairs, 2x2g corsair and 2x1g corsair) can make it well into the 1880Mhz range, the 1 gigs over 2050Mhz (reportedly). I am telling you this so we are able to feel confident that any instability discussed below is purely from the MCH and nothing else.

So, here is what I did:
while leaving voltages as mentioned above, i dropped the host clock to 200Mhz and the multi to its lowest (for me this is 6) so CPU frequency is about 1.2Ghz. I next loaded into this with my memory settings as listed in the sig + 2 1 gig dimms in the other slots to make sure that it would boot to such a low host clock.

next, i created a list of all the possible settings for memory frequency (ie 1600, 1333, 1066, 800) and Reference frequency (200, 266, 333, 400) and ordered them from fastest resulting memory frequency to slowest. working my way down the list i tried to boot into each setting, and if i could i would run a LINPACK test, which to my knowledge is one of if not the most demanding stability test out there. I also loosened up ram timings all the way out for the first 4 plus forced command rate to 2T.

So, and in line with what we have found or been told before, resulting memory frequency faster than FSB frequency will not boot or POST. making 200/1600, 200/1333, 266/1600, 200/1066, 266/1333, and 333/1600 completely worthless settings for all of us and eliminating 3 of the 4 1600 memory frequency combinations.

Now the fun starts...
next in the lineup is 200/800 with 200 ref frequency being the most restrictive (and presumably the best performing) setting in the category. also this is the last 200 configuration available, as the other 3 result in memory frequencies faster than the FSB.

at 200Mhz host clock by 6x, 200/800 booted and passed some short tests. (this setting has a ratio of 2:1, resulting in the memory frequency being the same as the FSB.) then i increased the host clock by 50Mhz until it wouldnt boot. wouldnt boot at 350x6, so i backed it off by 25 and it booted at 325x6 and passed some tests. while not trying to get exactly to the limit, i moved on to the next fastest combination.

starting with the host clock at 300Mhz, 266/1066 (another 2:1) was selected in the BIOS. Booted and tested just fine. bumped by 25, tested, and repeated until problems occurred. wouldnt boot into 400x6, but did POST. as my vMCH is maxed and my timings all the way out, nothing more to do but move on.

next, 375x6 with 333/1333. boots and tests fine up to 450 host clock (thats 1800Mhz resulting FSB and Mem freq). 475 (1900 fsb and mem) didnt want to play...would POST and start to boot sometimes, and others it would just randomly crash during the POST if it loaded at all.) It is possible that the 2x2g set i have was causing this as they start to have issues around 1880 i think

Somewhat hopeful from these results and the continual 50Mhz host clock jump gained in the previous tests, i set it up for 400/1600 @ 400Mhz. Nothing. dropped to 375, nothing, 350, nothing. while this sucks, i remember not being able to load into this before. I dont know if it i were using an x48 with a 1600Mhz FSB processor if things would be different, but i doubt it.

now i know from previous experience that settings at or below 400/1333 with the 2x2 and 2x1g sticks installed will work at 8.5x up to 479Mhz for me with some minor instability, so i set my focus on testing out where 4 dimms at 333/1333 would fail.

and since many people get hooked on the 1600Mhz ram number, i set the host clock back to 400 and bumped the multi to its next step, which is 7x for me. POST and boot went fine, and so did testing. so i jumped again, now to 400x8. worked like a dream. now to 8.5x400 and to my surprise, no problems (although my antivirus seems to have difficulty loading properly now. not sure if this is related to memory frequency or the timings, but i do know that using IDCC to change into settings that you are unsure of is a REAL BAD IDEA, as it lead me down the road of ChkDisk land, which could be the real problem behind the antivirus.)

anyway, now at 400x8.5 (my max multi), thats 1600 FSB and Memory with the CPU at 3.4Ghz, and everything working well, i begin to bump the host clock again. felt adventurous, so i jumped to 425 (1700Mhz), everything still working well...add another 10 to 435 (1740Mhz) still things are pretty good.

At 435 i noticed that the vMCH started to droop under load...something i never noticed before, but wasnt looking for either. vMCH was set to 1.7, displayed 1.65 in IDCC at idle, and 1.63 under load. while not much of a drop, we can use all we can get.

feeling a bit more ambitious, i bumped the host clock some more to 444 (dont remember why i chose this number, but i had serious doubts that 450 was going to happen.). 444x8.5 comes out to 1776Mhz FSB and Mem with the processor at 3.774Ghz. still booted and made it through the LINPACK test, but the results indicated instability. and since i know what all the other components are capable of, the MCH must be the culprit. vMCH droop was a bit worse and IDCC displayed 1.62 under load.

From here, i stopped testing. 440Mhz might be solid, but i think that the exact limit is not as important and determining the general limit as each MCH is slightly different and some may be able to handle 450 and others not able to handle 425Mhz with these memory settings.

Hope that this solves some things for everyone.

Sean

[redphil]

btw: I complaint at the IDDC-Developers a few days ago. Yesterday they responded and told me about the availability of the new Version. They told me that I shouldn't worry about the Beta-Status but should "focus on the features the software provides rather than the label (beta or production) that we assign to it".

[Manel]

Way to go Sean, very thorough.
red, that was exactly the answer I got some time ago. I'm still waiting for the answer to my last email, but I'm sure it will be the same, mentioning also the new version (which wasn't available at the time of writing).
I will install the new version shortly, maybe tomorrow or even today. Anyone installed it yet? Any differences or improvements?

[Sean-E-Boy]

it works pretty well, as before it can slow down system loading a bit and takes some time to load the interface. Some nice things about it is that it allows you to store 3 different custom configurations that can easily be loaded (may require restart) although i have yet to discover how to know what settings a particular configuration has assigned to it.

other than that I suggest that you use caution when adjusting the host clock through IDCC...crashed once on me even though settings i changed to are stable. upon restart the system wanted to run chkdisk and found a few little errors.

Other than that, you can change vCore, vMCH, and vFSB without restarting and i havent had any issues with that yet.

Note for last post:
Those of you with dual core processors, you should be able to attain at least slightly higher memory speeds than those of us with quads, as the quads tend to saturate the MCH.

Sean

[Sean-E-Boy]

fyi:
the memory speed lost between 2 and 4 dimms slams encryption, compression, and string sorting. However, the increase in RAM improves prime number finding and gigaflop rating using LINPACK testing, as well as some memory allocation functions.

for realistic things, seems encoding (or transcoding...DVD) is quite a bit faster with 2 dimms at a higher speed than more total memory at a slower speed.

what we need is for someone to get their hands on 2x4g DDR3 dimms and see if the x38/48 board can use them.

[JohnZS]

Not meaning to thread jack, but can anyone who has a PCI-E Gen 2.0 nVidia graphics card (Geforce 8800GTS 512MB) or newer, PM me a GPU-Z and CPU-Z screenshot showing your card running @ PCI-E 2.0 and the CPU-Z showing your BIOS revision. Preferably with forceware 196.xx
Thanks
With regards to 4 DIMM sticks of RAM (8GB) I have found that 1.33V is required on the MCH for stable operation @ 1333Mhz anything less and you can get freezes on ASUS X38
John

[Manel]

Hi John
I was waiting on the 196.34 beta to go WHQL before installing, since 196.21 seem to have issues (OC mostly).
Anyway, mine seems to be working fine @2.0 x16. Is this still related to the X38 issue you mentioned before?

[Manel]

John, you're using 4 sticks, 8GB, in your current setup now? I guess more than 1333 is not possible, right? Too much load on the MCH? Remember, this is on an Asus board, their engineers tend to tweak better their boards than Intel, so I'm not sure about the BT getting there.

[JohnZS]

John, you're using 4 sticks, 8GB, in your current setup now? I guess more than 1333 is not possible, right? Too much load on the MCH? Remember, this is on an Asus board, their engineers tend to tweak better their boards than Intel, so I'm not sure about the BT getting there.

You are right, More than 1333Mhz is not possible for me, although some have reported up to 1500Mhz (there or there abouts) on the P5E3 Premium.

With regards to nVidia drivers, 195.62WHQL should be fine too

Sadly ASUS are trying to pass the buck and say that the issue was never fixed on the DX38BT and DX48BT2 motherboards, however the release notes for your BIOS do indicate otherwise.

Updated support for high power PCIe adapters

Source is Intel BIOS release notes

The issue is we need people to provide proof with nVidia Gen PCI-E Gen 2.0 cards on these platforms. Then ASUS should hopefully get the fix sorted out before they discontinue support for these boards.

Please if you do have a 8800GT, 8800GTS 512MB, 9800GT, 9800GTX, 9800GTX+, GTS 250, GTX 280, GTX 285 and the DX38BT or DX48BT2 please provide a GPU-Z shot with Forceware newer than 191 series AND a CPU-Z shot showing your BIOS revision and motherboard model.

Please note 9800GX2 and GTX 295 boards "spoof" Gen 2.0 support because of the NF200 chip linking the GPU's.

John

[Manel]

John, I have posted the screenshots some time back, need it again? They should be 1 or 2 pages back.
Funny how Asus couldn't fix it yet, usually they are pretty quick fixing stuff.
Too bad the MCH can't handle the load though, sure would be sweet having all slots populated, loads of mem.

[JohnZS]

John, I have posted the screenshots some time back, need it again? They should be 1 or 2 pages back.
Funny how Asus couldn't fix it yet, usually they are pretty quick fixing stuff.
Too bad the MCH can't handle the load though, sure would be sweet having all slots populated, loads of mem.

The screenshots you posted Manel were perfect thanks
Just sadly ASUS are playing hard to prove this too, they need more proof, sadly one screenshot of one system is not enough. I think they need at least 3 different system screenshots to prove this is not an isolated incident?!

It is annoying as many people on the nVidia and ASUS forums have this issue, yet ASUS are not responding with the BIOS update!

John

[Manel]

Just making sure you didn't need any further info. Haven't installed newer nVidia drivers since I'm waiting for the next WHQL release, so nothing new in my setup anyway.
I used to be an Asus fan but at some point I got sick of it, now I don't buy Asus stuff anymore. I guess they got too big in the process. I know they are good, at least designing their boards and tweaking their BIOSes, but I'm not into it anymore.
Still, they are just trying to let this get forgotten.
Isn't there anyone else who qualifies?

[JohnZS]

Sadly the only people who can help on this one are DX38BT and DX48BT2 users with nVidia PCI-E Gen 2.0 cards as this issue also effects MSI, Gigabyte and DFI X38 and X48 motherboards.
It is a bit like the old RAID option rom issue, I had to press ASUS really hard to update their ICH9R RAID option ROM from 7.5 (a P35 revision) to 8.7 in order for my X-25M drives to work in RAID.
Intel X38/X48 based boards are now enjoying 8.9 RAID option ROM's yet ASUS refuse to update from 8.7 to 8.9
Gigabyte I think are still on 7.6!
Basically all third party board vendors are not applying the latest fixes from Intel, I am not sure why, it is almost like they do not see the point unless end users report issues and proof that Intel resolved the issue on their homebrew boards.
Thank you once again for your help. I hope other users help too
John

[Manel]

Glad I could help somehow, maybe others will too. I guess most people with this board use ATI cards.
Funny how Intel always seems to trail other manufacturers but in this case the're right in front of the pack. They better be anyway, it's their hardware. Still, can't help to think they're somehow lazy.
I'm not into taiwanese boards really (although Intel's are made by Foxconn if I'm not mistaken) so I don't follow BIOS updates of others, still I'm amazed how they don't keep up on those updates.
At least they listen to you and end up updating, some of the stuff anyway. That's good.
Good luck with Asus, hopefully they'll listen.

[crazysn2ke]

Hey JohnZS, here is my screenshot for you, hope it helps and you get some love from ASUS

[JohnZS]

Thank you very much crazysn2ke
I have relayed this information onto ASUS.
Hey do you guys know that Intel have officially released Rapid Storage Technology 9.5.0 for the DX38BT and DX48BT2 motherboards. You can get this download from your appropriate download section on Intel site.
It has boosted my SSD performance slightly.
FYI Rapid Storage Technology is the replacement to Intel Matrix Storage Technology
Once again Manuel thanks for your support
John

[Manel]

John, no sweat. Maybe with my shots and now with crazy's you can get Asus to deliver, they can't really delay this forever.
Yes, Intel RST 9.5.0 has been out for a while but it's funny that, as far as I can tell, it has been released only maybe a couple of weeks ago and it's dated October 14th. As usual, Intel takes it's time.
It was time they release audio and Dolby updates. The audio driver UI sucks, I have now some question marks on the IOs, don't know why, didn't use to be there before as far as I can tell.

[JohnZS]

The delay in releasing Rapid Storage Technology software might be both a good thing or a bad thing. I would hope it meant Intel put it through rigorous testing before releasing it to the public, in which case that would be a good thing

However, I, like you think that Intel are so slow at updating their pages and their software it was probably sitting there since November, yet not linked to, because department A that does the web design had not created the page yet

Does Everest say that your graphics card is connected into Slot #1 or Slot #2?

My card is in the primary port near the CPU yet Everest claims it is in PCI-E X16 slot 2?!?



John

[Sean-E-Boy]

I tell you what...

RST 9.5 and windows power option to turn off hard disks dont play well together. since installig 9.5 my system has locked up several times and i get SMART events on 2 of 3 drives that are in 1 of my RAID arrays. and it seems to coincide with the drives spinning down as i hear them turn off and then bam, system freezes.

Not positive that RST 9.5 is to blame...maybe 9.5 is better at picking up SMART errors, or my drives by coincidence started to fail just after installing 9.5?

However, this RAID array has no OS data, page file, program files, or anything of the sort on it required for the system to run, so not sure why a SMART event in this array would cause the system to freeze, but that seems to be the case.

Disabled power option in windows to turn off drives and backing up important stuff from the array in question now. will keep you posted.


sean

Update

So far i have transferred 400+ GBs out to other drives without incident. Drive power down option still disabled. Plan on deconstructing the array and rebuilding, then running scandisk on it, then putting everything back. will see what happens.

Well, have been error and freeze free since i disabled the turn off drives option in the power center. could be that 2 of the 3 are starting to have motor issues, but still doesnt explain the freezing part.

[JohnZS]

That is really bad news Sean.
Glad to hear you have backed everything up, as a precaution I have also turned off the drive powering down feature within Windows 7.

Out of curiosity when your system freezes do the lights go out on your mouse and keyboard?!

I myself have just been plagued by this problem too. (within the past few days). However I did not have the SMART issues you had.

I am LinX maxstress stable (10 runs) as well as loops of 3dmark Vantage and 3 hours of Memtest86 4.00 stable.

Yet get random freezing, either on idle, or playing a game, or just surfing the web.

Nothing is overclocked and the RAM is actually running slacker than spec (CL8 as opposed to CL7). I do have to run my NB @ 1.31V as using 8GB @ 1333Mhz is impossible with less than this.

I am an unhappy person in the ASUS camp, but like I said have this freezing issue too. If your symptoms are the same then perhaps this new Rapid Storage Technology needs a RAID OPTION ROM update? or was released by mistake for our motherboards?!?

John

PS can anyone who runs an nVidia PCI-E Gen 2.0 card do a screen print of Everest and a CPU-Z showing their motherboard revision, the screen print I would like is from the "Chipset" page which shows you what slot the graphics card is plugged in. ASUS boards seem to get confused where by Slot 0 is seen as Slot 2 and Slot 1 (or 2) is seen as Slot 3?!?
Thanks
John