** OFFICIAL ** Intel DX38BT & DX48BT2 BoneTrail Thread

[Manel]

Someone at MS is paying attention after all, the nVidia display driver update is gone.
Sean, I kinda like IE so I'm sticking with it, even if it's slower and all.
Funny that you mention Apple, not sure if there is another browser for OS X, maybe IE also? They also have MS Office so why not?! Guess they'll have it coming as well, if all is fair.

[Manel]

Seems there is a 196.78 beta nVidia driver out there with support for Fermi already, at least GTX470. Hopefully, the fan issue was fixed.

[Manel]

nVidia today made available in their site drivers 197.13 beta, or WHQL-candidate as they call it. Fan issue fixed from 196.75 finally. PhysiX driver newer version, as already available in 196.75 and HD audio driver update.
Numerous bug fixes as usual and performance improvements.
Wait and see if it's ok this time.

[Manel]

DivX Plus 8 is finally out, for those who care. HD all the way.

[Manel]

I found it odd that the site was down but they were testing the new page, still beta.
Also, new RAID drivers out v9.6.0.1014.

[petr0id]

Hello there! Long time no see 8)
Intel® Rapid Storage Technology Driver is available for download - TRIM function for SSD drives is enabled now.

All what i can say now, after upgrading to latest BIOS and Win7, i'm still missing a lot of useful features in BIOS. Like ability to save more than one custom settings for example. And cooler fan speed control would be great... And more detailed info on some undocumented features. And more, and more, and more... IDCC is sucks.
It's looks like we will recieve new BIOS only in case of EXTREME bug in chipset 8(
Anyway Bonetrail is solid as rock at defaults, so this is enough for me.

[Sean-E-Boy]

RE RST 9.6.0.1014:

Can anyone find a Readme about this update? maybe i am just missing it, but i dont see it and would like to know if they addressed the drive failures when windows powers down drives to save energy, or at least what they fixed.

sean

[Manel]

New Intel Chipset Drivers Bring TRIM Support for RAID Setups
The TRIM feature introduced by the newest generation of operating systems makes solid state drives more efficient with write performance. However, it does not work when RAID arrays are built with SSDs, until now. The latest version (9.6.0.1014) of Intel Rapid Storage Technology driver enables TRIM for each of the SSDs that are part of a RAID volume, of all RAID types, with the exception for RAID 5. The software can be downloaded from here.

Introduced with Microsoft's Windows 7, Windows Server 2008 R2, and with Linux 2.6.33, TRIM is a feature which uses system idle time to physically erase deleted data, since unlike with magnetic hard drives, data can't be simply overwritten on areas with deleted data. Deleted data must be erased before a NAND flash device can store fresh data in its place, and this causes additional write-cycles on setups without TRIM support. TRIM erases deleted data whenever feasible so that lesser number of write cycles are spent when there's data to be actually written in its place, increasing performance.

[petr0id]

@Sean-E-Boy

There is a readme file in installation dir, but it is really poor on info: just instructions on how to install.
There is no bugfix list at all. Intel guys are acting as usual...

May be You'll be able to find more info @
http://communities.intel.com/search?...=rapid+storage

[JohnZS]

Thanks for the info Manel
I must admit this new Rapid Storage Technology software really does love my X-25M drives.... even if they are Gen 1 drives!
ASUS have not responded to me, so my guess is they are brushing the PCI-E Gen 2.0 bugs under the carpet
The bug is confirmed by ASUS forum users on the following boards.

ASUS P5E3
ASUS Maximus Forumla
ASUS P5E
ASUS P5E Workstation Pro
ASUS P5E3 Premium

However ASUS have not as of yet provided a fix for something Intel fixed back in July 2009!

Ah well.... I wonder when the next Intel DX38BT/DX48BT2 BIOS update is going to be released and whether it will have a new RAID Option ROM!?!
John

[Sean-E-Boy]

petr0id, yeah, i saw that crappy readme...hate having to look for the changes.

John, i fear that there will be no further BIOS updates for us, and if there are, i seriously doubt that they will add any functionality unless the x58 boards that use similar BIOSes get something that is backwards compatible to us.

sean

[Manel]

John, it wasn't much but it sheds some light on what was changed/new features.
It's odd that Asus can't or won't fix that issue, they are either not capable (which I find hard to believe), or they don't think it's worth it or yet it's not fixable on their boards, maybe power constrains. Still, they should at least say something about it.
Regarding BIOS, I'm with Sean, unfortunatelly. I would like to see new BIOSes coming up with at least these updates but I can't imagine Intel doing any further development on it. Still, I hope.

nVidia released drivers 197.13 on thursday, with GTX470/480 coming in on friday, which is funny. Still the drivers don't support officially the GTX400 series (although we know they do). How do you release a new tech without drivers support?! nVidia style as usual. Hope these new drivers don't come with an extra feature/bug!!!

[Manel]

Looks like nVidia is coming out with GTS450/440/430 based on GF104, cutdown version of GF100. GTS440 seems to be a good part, for me that is. At least a 256b mem interface board. 2Q release though, I'll wait for the reviews and see if the specs hold true.
GTS450 20% I believe. GTS430 only 192b interface.
Lower power consumption finally. GTX480 rated at 250W after all, it was said to be 296W.

[GumaLive]

Hey guys, I just bought one more Raptor today. And luckilly I see new storage drivers.
I'm updating it right now :-)

Edit: So I have updated to Intel Rapid Storage... Previously named Matrix Storage.
- Software looks comletelly new :-)
- I had to tweak "Intel Rapid Storage" in services.msc. By default it is set for Automatic (delayed) start. Better is to use "Automatic". It will start it faster and not show error message in buble.
- I have also found bug... You must have 2 RAID volumes. Click Manage, Open Help, Click back to Management console and keep Help opened. Now on right side click on Volume1, Volume2, Volume1... While switching volumes the "Data strip size" item is not updated :-( (Possibly only GUI bug)
- Migration from 3 to 4 drives was sucessful, but took few hours.

[petr0id]

Hiya GumaLive!
I dont know is it right decision - to set RST service into the plain Automatic start mode. Maybe it is starting Delayed for a good reason - to wait for other services to starts and confirms their disk activity? Dunno. And is Yours first (meant on which Windows is installed) disk is treated as a "System Disk"? My one is not... Scrn attached.

[GumaLive]

No, my "System" volume is also not marked as system :-/

[JohnZS]

ASUS require 2 more screenshots and then hopefully they will fix the PCI-E Gen 2.0 bug.
They require a screenshot of a DX38BT running BIOS 1997
and then a screenshot of a DX38BT running BIOS 2006 from the same system. Ideally the BIOS 1997's nVidia graphics card would be running @ PCI-E Gen 1.1, yet resolved by the BIOS upgrade.
I have provided ASUS with the screenshots you have all provided however they say that it is not conclusive enough to prove that the P5E3 X38chipset is a PCI-E Gen 2.0 board
They are of the opinion that X48 is PCI-E Gen 2.0! NOT X38!
I provided ASUS a screenshot of a 9800GT running on a DX38BT using BIOS 1997, the card was stuck @ PCI-E 1.1 mode, yet all new(er) BIOS show nVidia cards happily running @ PCI-E 2.0

John

[Sean-E-Boy]

They are of the opinion that X48 is PCI-E Gen 2.0! NOT X38!
John

Well, then clearly they are F-ing idiots...as Intel has admitted that the x48s are nothing more than x38s that passed more rigorous testing at higher FSB and mem frequency speeds...ie "Higher binned" as the techies call it. and EVERYTHING ELSE on the board is identical to the 38...

Now i suppose that it is possible that when the BIOS package detects that there is an "x48" rather than an "x38", then PCI-E Gen 2.0 is enabled, where it is not with the x38?

Sean

[Manel]

This is why I ditched Asus, they're a bunch of idiots, full of it.
It's not that they're in this business only for a couple of days, if they don't know the chipset (X38), which they should of course, they could as well research instead of saying stupid things like that.
X48 is indeed a higher binned version of X38, you can even see it by the same BIOS used for both, right?! X48 is rated for 1600 mem/FSB while the X38 only supported (at first) 1333. Now, with the newer BIOSes both can reach 1600, though X38 not officially.
I guess they'll have to come to their senses. Isn't there anyone else complaining about this issue? If there are quite a few of you they'll have to do something I guess.

[JohnZS]

Thanks guys, I thought that was pretty much the case, I thought the X48 was higher binned and worked with lower voltages? Than the X38. Aside from those differences the feature set was identical.

ASUS, Gigabyte and Foxconn boards are effected.

ASUS and Foxconn symptoms are the same, however Gigabyte boards are half fixed, they report a slot capability of PCI-E Gen 2.0, just a card cabability of PCI-E 1.1.

If you are interested a lot of people are discussing this on the nVidia forums here

I am trying to help people by coordinating testing etc. I have found that the ASUS P5E and an ASUS P5E Flashed to a Maximus Formula is unaffected by this bug, however my P5E3 and a Geniune Maximus Formula IS effected by the bug?!? Weird huh!

On the whole most X48 boards are unaffected and only 2 X58 boards (A Biostar board, and 1 ASUS board) are effected so far.

Thanks for your support guys.

John

[Manel]

Didn't have the notion that this issue even existed before you mentioned it, good thing I went with Intel after all.
Funny how the top dogs are affected, not others. Even Foxconn, which I believe makes the boards for Intel, although BIOS development is different.
X58 having the same issue is weird.

Seems nVidia will have a new GF100 SKU out, the GTX460, this one looks good to me, 256b wide mem bus, 1GB mem, normal card. Just wonder about power draw. Just a reference design though, no actual card.
Also, EVGA seems to have a 4 way SLI, not Quad SLI, running on their Classified boards, wow.

[Manel]

John

It seems there were some issues regarding PCIe 2.0 with the SmackOver (Intel DX58SO) as well, fixed almost a year ago though. I imagine it would be the same problem, not sure though. Power issue.

[JohnZS]

Thank you very much for the info Manuel.
It appears that Intel were having a link modulation issue on the DX58SO, apparently they fixed it in BIOS 4196 and addressed another PCI-E issue in BIOS 5020.

With regards to the current situation facing nVidia cards and non-Intel branded Intel chipset boards. P45 chipset boards do NOT have the issue. X38 and some X48 based boards have the issue, with a handful of X58 and P55 boards having the issue.

Gigabyte are playing the "Gigabyte branded cards work fine, so therefore if you use eVGA, Zotac, BFG or another branded card then it is your fault" card to Gigabyte owners.
ASUS are stalling and making out that the X38 chipset is flawed, yet the X48 is perfect.
Foxconn MIGHT try and fix the issue, IF they can.

I had a report today of someone experiencing this issue under Linux using a GTX 480 on an ASUS P5E. For them the issue was causing a huge performance degradation as they are doing CUDA computing under Linux. The performance was less than half on the ASUS P5E compared to the nForce 790i with exactly the same system components. ASUS really need to sort this out.

Have you guys had a new BIOS yet?
Intel seem to have slowed down somewhat on X48 and X38 BIOS, they also seem to be holding out on the new Option ROM's.

John

[Manel]

John, thanx for the update.
You are right, Intel seems to have dropped BIOS development for these boards nothing new since BIOS 2006, and this was October last year. Even new drivers aren't being released. Audio for one is something to forget about, they are not only terrible but Intel doesn't provide any updates from IDT.
Option ROM seems to be forgotten, yes.
Latest update out was RAID, a month or so ago.

[Manel]

The nVidia issue is odd, since it reveals itself only on certain configurations as I understand it. Is it a mix of hardware and software problems? Is it a flaw in nVidia design or Intel's? Since not all chipsets have the issue, even hardware that is essencially the same (X38 and X48) might or might not have the issue, could it be an OEM implementation/design problem? Even different drivers have different behaviours. So, where is the problem really? I guess everyone is trying to pass the issue to someone else.