This is a puzzle/problem that has now consumed hours and hours over several days of my debugging, yet i have not arrived at any solution yet. Hoping you guys can give some insight.
First, my setup:
Gigabyte G1.Guerrilla X58
Windows 7 x64
24GB GSKILL RAM
HD - irrelevant (tried this with both HDD and SSDs, doesn't really matter, but the point is i tried this under numerous W7 installs, so not believed to be SW/driver related).
1250W OCZ PSU
Core i7 920, no O/C during debug
The GPUs I am playing with are two EVGA SC 570s and a Galaxy 460: this is intended for 570's in SLI and 460 to support 2nd and 3rd monitors (as well as PhysX dedication). I also have old 8800GT i used for testing at one point.
Mobo PCIE slots are x16, x16(shared), and x8 (takes away from the shared 2nd slot x16, forcing that 2nd slot to become x8 and thus the 3 slots become x16, x8, and x8 respectively).
Next, the background:
I had been running with these 3 video cards just fine for a while. I wanted to finally plug in a USB3.0 header (for front case panel) into the motherboard. But when I did that, the 3rd card became undetected (the 460 at the time). Since I was also working on some new upgrades (new SSDs), I went ahead with fresh install on SSDs, but the same problem persisted. TO avoid wearing my SSDs, most of my debug after that was on another HDD with yet another fresh W7 install. In the process of this debug, I have collected the following clues, maybe you guys can give me some ideas as to what's going on.
-3rd PCIE slot's card is NOT detected by either BIOS (No display) nor W7 (Device Mgr, scan for new devices) whenever USB 3.0 header is connected, even if i disable onboard USB3.0 controller in BIOS. It is also the case if i use fewer than 3 PCIE slots (in other words, if i populate just 3rd slot, or 3rd slot + one other, whatever is in 3rd slot - be it 560, 460, or 8800GT - goes undetected). Although the USB connector is next to the 3rd PCIE slot, i do not believe it is a physical contact/restriction issue, as i moved the USB3 header to the side to make sure no contact with vid card. I also did an experiment where I pulled USB3 cable out without unplugging any of the inserted video cards.
- If I unplug the USB3 header, but do not move any of the vid cards around, then reboot: the 3rd slot STILL goes undetected. What it takes for it to become detected after USB3 is unplugged is for me to first pull out the card from the slot (I only tried the 3rd slot itself, not sure if this would work by pulling a different slot's card), rebooting into windows in this new state, then shutting down and re-inserting the card into 3rd slot, and rebooting again: now it will be detected.
- Numerous BIOS updates/downgrades/DMI cleanses did not seem to make a difference.
- I believe the PCIE bus is shared by resources like the USB3 itself, plus the SATA controllers. In case of some "overload" i tried disabling unused SATA controllers (Gigabyte ctrlr and Marvell ctrlr, as my debug HDD setup was on Intel ctrlr). As mentioned, I also tried disabling USB3 as well. None of that helped the video card issue, however.
Does "onboard USB 3.0 controller" refer to a single controller that controls both the "header" USB3.0 connection and the attached USB3.0 ports on the back panel of the motherboard? If so, then it is puzzling to me that disabling this controller has no effect on the issue, and the physical connection of the cable to the header still kills the 3rd PCIE slot somehow. I know this may sound like a physical restriction of card's contact with the slot, but I tried this many many times as I also thought this was the root cause, and I am confident that it is NOT the cause. I've tried 3 different cards of varying sizes, the 460 really fits quite comfortably and clicks into position. I've also tried physically pushed cards down etc to make sure full contact.
Any help would be appreciated. I am really close to contacting Gigabyte about this.
More about :detect 3rd pci card usb3 header plugged