Intel DX38BT/DX48BT2 (Bone Trail)

[Sandog]

FYI, Intel provides specification updates for all their processors/chips which contain errata references. The DX38BT has none for the mobo itself. The X38 MCH has a few but most serious is 1.1 PCIe (slot 1X I believe) causing some issues with 1X card vendors who have not adhered to the PCIe1x spec. The ICH9 has one or two I believe. Both chips (MCH/ICH9) have no fix plans so the whole board looks good to go. Other compatibility issues may arrise later though.

On a side note, my discontinued X6800 has quite a few bugs in her. I have the first B2 stepping. There has since been a B3 and G0 stepping. I have had fatal errors with manipulating large ammounts of data which I originally thought was a memory issue but it is not. Hopefully the new Q9450 Penryn will be pretty solid as I plan on getting it when it arrives.

Also, I believe Intels X48 chipset board will be called DX48BT2 and not DX48BT (if and when it comes out).

[ATINUT]

Originally Posted by Sandog

FYI, Intel provides specification updates for all their processors/chips which contain errata references. The DX38BT has none for the mobo itself. The X38 MCH has a few but most serious is 1.1 PCIe (slot 1X I believe) causing some issues with 1X card vendors who have not adhered to the PCIe1x spec. The ICH9 has one or two I believe. Both chips (MCH/ICH9) have no fix plans so the whole board looks good to go. Other compatibility issues may arrise later though.

On a side note, my discontinued X6800 has quite a few bugs in her. I have the first B2 stepping. There has since been a B3 and G0 stepping. I have had fatal errors with manipulating large ammounts of data which I originally thought was a memory issue but it is not. Hopefully the new Q9450 Penryn will be pretty solid as I plan on getting it when it arrives.

Also, I believe Intels X48 chipset board will be called DX48BT2 and not DX48BT (if and when it comes out).

Don't feel bad, the Core 2 Duo cpu's have some errata bugs in them also. As such Microsoft released a patch in June for XP and now for Vista 32 and 64.

A microcode reliability update is available that improves the reliability of systems that use Intel processors

Here's some of the bugs that exist in the Core 2 Duo line of dual core cpu's.

Write-protect or non-execute bit for a page table entry is ignored.
Floating point instruction non-coherencies.
Allowed memory corruptions outside of the range of permitted writing for a process by running common instruction sequences

My Core 2 Duo E6600 cpu was discontinued within the past 8 weeks.

[Sandog]

...and this goes to show what happens when a company is in a rush to come to market like the Conroe was. All CPU's have errata but this is the first CPU that it has effected me.

[SpeedDMN]

Q: What issues are fixed by the microcode reliability update?

A: The microcode reliability update fixes the following issues:• A possible Intel processor marginality
• A potential source of unpredictable system behavior
• A "stop 0x7E" error that may occur during startup on some systems that are running an older Pentium 5 CPU

Humm I wasn't aware of the Pentium 5 series Must have been a short release. I see no reason to apply this patch since I don't have any of the symptoms that I know about. I guess if if aint broke don't fix it..

[ATINUT]

I haven't run in to any major problems, that I can tell yet with my Core 2 Duo cpu.

[Sandog]

I can make mine crash. Just scan a high rez photo and when it tries to transfer the data it locks up. I will post a pic of the error later. It is not my memory malfunctioning. I have good tested memory now and even running at SPD of 5-5-5-18 it still does it. It's the CPU. There are at least 125 erratum for the early C2D's including mine. Several of which cause unpredictable system behavior or hang ups. Intel added erratum just this month and have had 22 updates since these CPU's arrived.
http://download.intel.com/design/pro...t/31327922.pdf

By the way, Corsair replaced my dual kit with another one. I sent them a kit of black Dominators I paid over two hundred dollars for and they sent me a silver set valued at $60 on NE after $40 instant rebate. They are the same spec though (6400/4-4-4-12/2.1V).

[ATINUT]

Ouch, the Core 2 Duo E6600 is one of the early C2D's. I get a problem, where when I download something, or copy something to my second harddrive, it will freeze up and ask me to try again, not sure if it is a windows problem or a cpu bug popping up. This happens occasionally, I might add.

[SpeedDMN]

Could there be any harm done installing this patch? I have a T7200 in my lappy. It was not a launch time chip, but it was a fairly early C2D. The mobile chips were mentioned as possibly having this errata.

[ATINUT]

Checked, and found out I have a B2 stepping Core 2 Duo E6600, which means I could possibly have one Core 2 Duo's with the bugs.

[Sandog]

FYI, here is the types of memory configurations for latest Intel boards. (this applies to the new X38 chipset board DX38BT).

Desktop Boards - Single / Dual / Flex memory modes

[traveler]

Originally Posted by ATINUT

Checked, and found out I have a B2 stepping Core 2 Duo E6600, which means I could possibly have one Core 2 Duo's with the bugs.

Send it back for another.

[ATINUT]

Originally Posted by traveler

Send it back for another.

Gotta have a good reason to RMA it to Intel otherwise Intel won't have anything to do with it.

[GWillakers]

Originally Posted by Sandog

I can make mine crash. Just scan a high rez photo and when it tries to transfer the data it locks up. I will post a pic of the error later. It is not my memory malfunctioning. I have good tested memory now and even running at SPD of 5-5-5-18 it still does it. It's the CPU. There are at least 125 erratum for the early C2D's including mine. Several of which cause unpredictable system behavior or hang ups. Intel added erratum just this month and have had 22 updates since these CPU's arrived.
http://download.intel.com/design/pro...t/31327922.pdf

By the way, Corsair replaced my dual kit with another one. I sent them a kit of black Dominators I paid over two hundred dollars for and they sent me a silver set valued at $60 on NE after $40 instant rebate. They are the same spec though (6400/4-4-4-12/2.1V).

Hey Sandog,

Howd you come to the conclusion that it's your CPU?

I realize that processors often have errata, and multiprocessors are going to have exponentially more problems then uni-processors... just the nature of the beast... sharing data and all that...

I would think that if you were really experiencing hardware level problems...
your overall experience with your system would be MUCH more erratic...

A problem where you have it narrowed down to transferring lots of data, like scanning a large photo sounds like it is emanating from the upper levels of the application programming... or maybe in the operating system itself.

I would tend to look for buffer overruns or bad drivers...

A LOT of old software was NEVER intended to run on multiple processing systems.

Programmers need to take extra precautions when writing code that could possibly be grabbed by more than one engine.

It will work fine, if only one cpu executes it.
And it will fail sometimes if more than one gets to it

Try testing your scan by either disabling a CPU for the the scanning process, or more drastically disable one of the engines in bios.

Operating systems like Vista, XP.. experience many errors with concurrency. Even though they were designed to handle multiple processors from the start.

Application programs for the most part, were only designed and tested with one processor in mind.

[Sandog]

Intel might be soon releasing the P45 chipset. This is a mainstream chipset similar to P35 but it will have CrossFire support, 65nm chipset, and supposedly 1600 FSB support. I may have to take a look at this one since it would use considerably less power than either the X38/X48 chipset boards.

Intel P45 might arrive when X48 does :: TweakTown

[Sandog]

Originally Posted by GWillakers

Hey Sandog,

Howd you come to the conclusion that it's your CPU?

I realize that processors often have errata, and multiprocessors are going to have exponentially more problems then uni-processors... just the nature of the beast... sharing data and all that...

I would think that if you were really experiencing hardware level problems...
your overall experience with your system would be MUCH more erratic...

The program works fine at scans <= 1200dpi so the program works. I get a memory error message from the scanner program (writing to high addresses I believe). I am almost positive my memory is functioning correctly.