see post Jun 25, 2010 12:52

Sorry for not responding earlier.

My three setups:

1) Two LCD displays, 1600x1200 each - Now using an NVIDIA card with zero errors.  Previously AMD 4850 and tons of errors and occasional freezing.

2) Two LCD displays.  1920x1080 and 1920 x 1200 using AMD 4850 with errors regularly appearing but at much less of a frequency than #1; no obvious hangs.

3) One LCD display.  1280x1024 using AMD 4850 does error regularly but no hangs or freezes.

The people at AMD/ATI have a new BIOS in testing that solves this problem.

Very cool. We just got a new BIOS for the X58SO board last week. I'll check my event logs tonight and post back to see if the issue has been corrected on the Intel side.

Sorry, but to clarify: "The people at ATI/AMD" means the people whose hardware is in all those RADEON video cards. The problem is with the bios on the graphics card.

There's good news and bad news. The good news is AMD has found the problem:

> The root cause of the problem is related to a Replay Timer Timeout  

> Error (Bit 12), which means the GPU was not able to send the  

> acknowledgment (transmission) to the Root Port (PCI Express BUS)  

> quick enough.

The bad news: They coincidentally found that they don't get the log errors after updating to
bios 5390 on the DX58SO, and want to mark it 'solved.' Given that this problem occurs with
many models of ATI card and only one card (not model, but specific card) not made by AMD,

and happens on multiple MOBO's...

Can we confirm that the latest Intel BIOS slows down the refresh timing? Does it solve this
problem?

Negative. I update the BIOS immediately when it comes out. I am still seeing the same errors I previously reported with my DX58SO and a GTX480.

Of course, becuase of this, I've always felt that it had nothing to do with AMD in the first place.

I'm getting the same, sporadic, message in my Event Viewer ---

Warning 7/23/2010 2:08:16 PM WHEA-Logger 17 None

Description: A corrected hardware error has occurred.

Component: PCI Express Root Port
Error Source: Advanced Error Reporting (PCI Express)

Video Card:
Nvidia Quadro FX-1800
Driver Version: 258.96-Quadro-win7-winvista-64bit-english-whql

Monitor: HP LP2475w

MY RIG:

OS Name Microsoft Windows 7 Ultimate (64BIT)
Version 6.1.7600 Build 7600
OS Manufacturer Microsoft Corporation
System Name xxxx
System Manufacturer INTEL_
System Model DX58SO__AAE29331-503
System Type x64-based PC
Processor Intel(R) Core(TM) i7 CPU   920  @ 2.67GHz, 2661 Mhz, 4 Core(s), 8 Logical Processor(s)
BIOS Version/Date Intel Corp. SOX5810J.86A.5390.2010.0614.1053, 6/14/2010
SMBIOS Version 2.5

Hardware Abstraction Layer Version = "6.1.7600.16385"

Installed Physical Memory (RAM) 6.00 GB
Total Physical Memory 5.99 GB
Available Physical Memory 4.54 GB
Total Virtual Memory 12.0 GB
Available Virtual Memory 10.2 GB
Page File Space 5.99 GB

All latest (Urgent Only) updates from Microsoft

All latest DX58SO Chipset and other drivers from Intel - each 'Intel' Chipset driver manually installed via Device Manager.

Yesterday, I received my video card back from ATI. They updated the VBIOS. Ran for 24 hours without a single WHEA 17. It is -- FINALLY -- fixed.

Glad to hear it. I'm fixing my warnings too....by buying an ASUS Rampage III Formula board and shelving this DX58SO once and for all (the latest BIOs didn't fix it). That will fix a lot of the issues I'm having, lol.

Have same problem with asus barbone V2P4V900, ATI HD4300, PROC INTEL D820  2 GB RAM, 200 GB HDD with win 7 pro 32 bit.

After disabling DPS service Diagnostic Policy Service. Logging of errors stoped.

This is just temporary solution.

Has a permanent solution been found for this  - a fix rather than a workaround?

I have been trying to resolve the problem for several months without success. Through my research on various fora and google searches it seems the only constant is to be using a motherboard with the x58 chipset on it, 64 bit win7, and a high end graphics card in a PCIe slot. The problem has been reported with different brand motherboards, different brand graphics cards, running under the 64 bit win 7 OS. This tells me the problem is more subtle than meets the eye and if we are looking for bios problems or hardware problems  we are probably looking in the wrong place.

When I first read the post above regarding disabling the diagnostic service I decided to give that a shot since it was about the only thing I hadn't tried. I didn't disable the service, I just changed the start up mode from automatic to manual to eliminate a possible windows polling start-up  error between the video card, PCIe bus, and and x58 chipset.

THE WHEA ERRORS FINALLY STOPPED - the first time I have received no errors at boot up. I have no idea if all of the permutations of this error will be solved by this approach but it works for me. It seems to me there is a timing and/or logic glitch occurring when the x58 chipset, the win 7 64 bit diagnostic service functionality, and a high end graphics card  try to play together. I am happy to find this because it sorta indicates that my intel DX58SO MB and EGVA GTX480 graphics card are probably OK and I don't have a "real" hardware issue but I'm also kinda unhappy because I still don't know the real nature of the problem and I hate to fix things via a workaround which I feel this is.

Somebody please chime in with a better idea?

I try to reply to this question once a month...

There is a lot of noise out there. People blame the board, the chips, the card... but the problem is (if you have this particular problem) a REPLAY TIMER ERROR. This can get annoying by filling up your log, but it is usually just recoverable and barely noticable, if at all, on the display. It CAN occur on any hardware (rarely), but certain video boards (RADEON HD 4000 and 5000, nVidia GTX, et. al.) push the limits, especially at HD resolutions (1920 x 1080 is a common culprit). You *can* turn off error reporting, but the AER is there to tell you other things; perhaps important things; things that you might want to know. You can also use a different aspect ratio, which will likely make the problem go away, but then you wind up with the problems associated with driving your display at a non-native resolution -- which can range from not-a-problem to performance-killer. On the Radeons, the problem was so bad that I forced AMD to fix it (literally! I wrote the VP engineering, exported a DX58SO from Intel to AMD's engineers in Canada, harassed them weekly until they wrote a new VBIOS). I imagine they don't advertise the fact that you can get a new VBIOS with the correct timing, since it costs more to do that than it costs to just replace the card with one that contains the fix.

ENVIRONMENT

Windows 7 x64

HP Elitebook 8540p

Nvidia 5100M

RAM HP-provided (4GB) + Crucial (8GB)

Same general issue as reported elsewhere in this thread. One thing that may not have been mentioned elsewhere is that the self-correcting or perhaps the logging process itself causes a hesitation in the system which is noticeable. This laptop is used for audio processing, and the hesitation causes a stutter if it occurs when the system is supposed to be converting audio.

HP had me try many things, including reinstalling Windows, but today it's going back for repair. I worked with as many as four different technicians over several months (hey, I had other fish to fry). The last one thought this was related to the graphics card.  Note that this is Nvidia not AMD.

HP and Microsoft diagnostics for this laptop as requested by HP failed to turn up anything.

See also my post in forum.notebookreview.com (search for "knowlengr"). I believe this laptop uses the same MBO as an Asus machine, and users there were having the same issue.