Over the last month I have been encountering sporadic boot issues with my homebuilt PC. As mentioned I have an ASUS Extreme Rampage MoBo and Radeon 4870x2 card. When I turn on the PC, the drive/power lights come on, as does the mouse. But the keyboard doesn't light up, and no monitor output.
This may not happen several times in a row, or I may have to restart the machine many times before it finally posts and the screen lights.
I'm wondering if this could be a mobo, GPU or power supply issue. What is most likely and how do I go about troubleshooting?
More about :recent boot issues asus extreme rampage radeon 4870x2
Since your running on crossfire, try using one and boot the pc and see if there's a display right away from a fresh boot up. I had experience that sort of thing which after a few months my gpu failed, most probably due to heat issue and kinda shorted the gpu life.
I've had this exact same issue with my Gigabyte MA790X-UD4P and an XFX HD4890 (489A-ZDEC) for a year-and-a-half, and recently it's gotten a bit worse. (Phenom II X3 720 CPU installed.) I've never found a solution to the problem, despite having spoken to both Gigabyte and XFX about it. (Gigabyte tech support is pretty much worthless, IMO.) I've tried 4 different mobo BIOS versions, two PSUs, and different memory - no effect.
@ apache_lives - I manually set my memory speed, timings, and voltage during the initial build, as well as when I tried the different RAM, yet the problem still arose. Both sets require only 1.8V to operate at 1066Mhz (it's DDR2), but I ran also them at 1.9V. Basically, I had already tried what you suggested and it didn't help. While the OP can give it a shot, I don't believe it's a memory-related issue at all. I tend to think it's something inherent to some 4800-series cards. I'll explain my experience with the issue.
There are the 3 different start-up scenarios I've experienced:
1. Starts up "OK." (I guess it's OK...) There's an initial power burst where all the fans spin up, followed shortly (~2 seconds) by a strong burst from the GPU fan. The GPU fan will settle quickly (~2 seconds), and about 5 seconds later, my KB and mouse light up, which is immediately followed by the display of the POST screen. This is the fastest it's ever started - about 10 secs from the press of the Power button to POST screen display. This might happen about 30% of all times the system is started, and will occur every time the system is shutdown and restarted from within Windows.
2. I call this a "stalled" POST. Sometimes my system takes a minute or so to POST, perhaps attempting to initialize video over and over. I can hear the GPU fan spin up and down repeatedly, and see the HDD activity LED blink a few times while the GPU fan is going through it's spin-up/spin-down cycling process. During this scenario, pressing the reset button after a GPU fan cycle will often cause it to POST, but can require 2, sometimes 3 attempts. It should be noted, the first GPU fan cycle is significantly more pronounced than those that follow. The fan initially seems to jump to 75-100% speed, but rarely seems to spin beyond 40-50% speed during any following cycles. This happens about 70% of all system starts.
3. The PC shuts down without starting. This used to happen very, very rarely, but is the scenario that's occurring more often recently. Initially, there's the burst from system fans, but there's no burst from the GPU fan at all. Eventually, perhaps 20-30 seconds later, I'll hear the GPU fan kick in very strongly. The GPU will then repeatedly cycle as it does in scenario 2. Pressing reset after any of these cycles fails to cause the system to POST as it does in scenario 2. When reset fails, eventually the system just dies and shuts down. At this point, I have to turn off the PSU via it's switch, press the Power button to dump electricity stored within the CAPS, flip the PSU back on and start over. This situation has never repeated itself twice in a row, thankfully. Until a few days ago, I can only remember this happening 2 or 3 times over the past year-and-a-half. Since then, it's happened 3 times.
The key seems to be whether or not I hear the fan of my 4890 kick in and slow down. In scenario 1, I hear the GPU fan burst right after the initial system power-up. It bursts to what I'd guess approaches 75-100% speed, then slows down. After a brief pause, the system will POST. In scenario 2, this fan bursting process repeats itself two, three, and even four times. After one of these 2-4 cycles, it will POST on it's own. Or, as I said, if I hit the Reset button following a fan burst and slow-down cycle, the system will seemingly roll into scenario 1 and POST. In scenario 3, the burst takes time to even begin, and then seems to repeat almost endlessly. Resetting after a fan cycle has no effect, and eventually the system just plain fails, shutting itself down. Without cycling the PSU and dumping the CAPS, the power button is totally unresponsive. In all scenarios, I never get any beep codes.
As I said in scenario 1, the system will start properly any time I shutdown and restart within Windows. Similarly, I've never run into this problem if I shut my system down for 5, 10, maybe 15 minutes or so. But, any time it's off for an hour or overnight, it will usually occur. I tried a couple different cards (4850, 8800GT, 4670) in the system to see if this still happened, and it didn't. But, I only left the system sit 15 minutes or so while I had those cards available. I was never able to leave any of those cards in the system overnight for a thorough test.
I've looked over the lights on the back of the card to see what they might show, but I get the proper three lights [X] [ ] [X] [X] every time.
Lastly, and this part has always confused me, whenever this issue occurs, it seems something within the system changes. This is made apparent prior to starting Windows. After the initial POST screen, memory test, drive detection, etc., the Device List (each device's IRQ) is shown (within a box). Following that, I see "Verifying DMI Pool Data .......... Update Success!" Now, I only see the "Update Success!" portion of that message following a "stalled" POST. If I allow the system to start, then shutdown and restart it, the message is displayed again, as if it's reverting to conditions prior to the first "Update Success!" message. Subsequent shutdowns and restarts display no further "Update Success!" messages. My system seems to run fine, regardless of whether I'm using it after the first "Update Success!" message or the second. I can see no apparent hardware changes within Windows itself, and the Windows Experience program shows no indication that my system hasn't been rated due to a hardware change.
We have another AMD system (Gigabyte MA785G-UD3H, X2 550 Calisto, HD4850, 4GB DDR2-1066) here, and it has never experienced this issue.