- problem persists after BIOS update 0165
- problem 100% reproducable for: cold start, warm reboot
- problem does *not* occur for: wake from (hybrid) sleep
Looks like there is a difference in video card initialization for reboot and wake. This board can obviously work correctly with a Radeon HD4670 and two monitors attached, so it must be a software (BIOS firmware) issue. It would be nice if the BIOS could be fixed so I would not need the awkward boot procedure of detaching the secondary monitor during POST anymore.
Many of the problems in these forums are power supply related. At least list your model. See BEHardware and AnandTech for monitor information.
Thank you for your reply.
Let me give some more details on the system and why I think it is a software / firmware issue rather than a hardware issue.
Since you asked: the power supply is a Be Quiet Straight Power 450 W. But I don't think it has anything to do with it. Once booted, my system is rock solid and only needs to reboot for certain OS updates. I have not yet (in 3 weeks) had an OS crash. I use hybrid sleep (which works 100% reliably), RAID 10 (also 100% reliably), and every now and then I use all 8 logical cores of my i7 to do a full recompile of a rather large Visual Studio project. Temperatures never run high because I have a case which provides sufficient air flow. Believe me, the system is stable.
It's just a *boot* problem only, or rather, a POST problem.
This may be due to the graphics card, a Gigabyte GV-R467ZL-1GI, it's BIOS, or the DP55WB BIOS.
But rather I think it is a fluke incompatibility between this card and the DP55WB.
If it was just the DP55WB, more people would have noticed, and the same reasoning goes for the GV-R467ZL-1GI.
More exact description of the symptoms: at full reboot (either cold or warm), with both the DVI and VGA outputs connected, the system does not get through POST and remains at "E7". Noteworthy: at this moment both monitors *are* showing the boot screen as they should, so both outputs are working. I can press Ctl+Alt+Del for another go and it will stop again at E7. When I detach the VGA connection the boot always succeeds. As soon as I see the OS loading screen I can attach the VGA monitor and it will work properly.
When the system wakes from S3, there are no problems. The system happily wakes with both monitors attached and they function correctly. Obviously there is a different code path in the EFI for boot and wake.
I am still trying to confirm a problem between an 8600GT silent and a G965WH motherboard. The 8600GT has an odd post on a DG45ID in that a second screen appears, possibly because I did not disable onboard graphics prior to installation. Your system lacks onboard graphics so you would not see this. Minimum power requirement for your video card is 400 but a comparable nVidia card would be 460 or greater. Radeon has apparently changed but it had typically higher power demands. As far as POST goes this would not be an issue but any delay could upset the order of loading. I doubt the BIOS employs any threading, however the OS does not necessarily load the same way every time. The power demand of your system (which I am currently considering) is higher due to the i7. I am still concerned that you may not have enough amperage although I haven't looked at your specs. I had looked at a Be Quiet supply some time ago and was not motivated to investigate further (possibly due to availability). Once I install a different LAN driver and retry my 8600GT install I may have more to say. It isn't surprising that some 55 series issues would still exist.
you can rule out a PSU-problem, 450 is definitly strong enough. the maximum tdp of the 8600 is below 50W. a pcie slot provides up to 75watt afaik.
try a more recent card, like the nvidia 9400/9500.
No more attempts with the G965WH will be made since a 7600GS will start but not the 8600GT. The slot must not be providing enough power after the system failure. I'm still disturbed at the thought of hot-plugging a monitor. I must be wrong.
After installing the newest ATI drivers (9.11) the problem no longer occurs for warm reboot. It appears there is something weird with the graphics card or its graphics BIOS that makes the card initialize wrongly if 2 monitors are attached. ATI may know about this and "fixed" it by resetting the card properly upon driver initialization, so it does not occur anymore on warm reboot. However, the graphics BIOS itself is of course not modified, and the problem persists for cold start (when the drivers did not yet have a chance to patch the initialization problem).
So I now think it really is *not* a DP55WB problem, but rather a problem with the Gigabyte Radeon HD4670 card (or its BIOS). Looks like I will have to wait for a BIOS update of the graphics card instead.
I have had a problem like this .. Same MB but gfx card was Nvidia GT240 512 DDR3 .. System worked fine but to fix an audio issue, where audio would crash after a random time the BIOS of the MB needed to be updated.
I updated to latest BIOS which was release 190 and after reboot there was no video signal .. Tested as follows
1) removed RAM, and got BIOS error beeps
2) removed gfx card and system booted normally (could hear windows starting)
3) put old Nvidia 6200 card ind and system booted normally
4) tried GT240 card in another system .. card worked fine
5) "updated" BIOS to version 122 and system worked fine again .. although i dont know if the audio issue persist yet
So it would seem that your problem could very well be MB Bios related
Someone here has reported the need to use HDMI/DVI and DVI/HDMI adapters to connect monitors correctly depending on the default connection. If a monitor (or keyboard) is connected with the power supply on, a hardware change screen should appear (at least on restart), necessitating a reboot. I now have seen an audio synchronization issue for online video that is difficult to reproduce. This is primarily of concern due to sounds suggesting hard drive problems at that time. See http://communities.intel.com/message/77516#77516 for audio memory issues that I am concerned could also relate to the page file.