- Joined
- Jan 11, 2010
- Messages
- 41 (0.01/day)
System Name | Armor |
---|---|
Processor | AMD FX-8350 Eight Core CPU @ 4.1 Ghz |
Motherboard | Asus Sabertooth 990fx Rev. 2.0 |
Cooling | Corsair H100 Liquid Cooler |
Memory | Corsair Dominator Platinum DDR3 1866 CAS 9 |
Video Card(s) | 2x Gigabyte R9 290xOC Windforce 3 PCIe 3.0 |
Storage | Samsung 840 EVO 120 Gb SSD (OS) Seagate Barracuda ST1000dm003 1Tb SATA (Storage) |
Display(s) | LG 27' 1080p AH-IPS pannel LED Monitor. |
Case | NZXT Phantom 410 Black |
Audio Device(s) | on board Realtek Audio Processor |
Power Supply | Seasonic X 1250w 80+ Gold |
Software | Windows 7 Ultimate 64bit (Genuine |
Hello all, this will be my first post here since before 2009. I've got a new system up and running, but I've met some massive difficulties in getting it to run properly. System specs are posted on my profile!
First off, I'm running (Trying to, anyways....) two Gigabyte R9 290x 4 gig cards in crossfire. I've got an Asus Sabertooth 990fx Revision 2.0 motherboard. Currently, I am only running one of my cards, as the other seems to be bad. But, since the 290x's aren't the first cards I've ran on this system, lets start from the top, shall we?
When I purchased this system, I started out with two R9 280x cards, they worked in crossfire for a few months, however, after discovering that my motherboard manual recommended using slot 1 and slot 3 for crossfire, I switched the cards to these slots. It took me a long time and a lot of tinkering to get it working again (frequent system shut downs). After a while the secondary card started giving me a lot of problems. So, I pulled the primary card and tested the secondary card in OCCT, during the video card tests it showed a lot of errors (in the thousands) after the first 30 seconds and then the system would freeze. It seemed the card couldn't hold the 3d mode clocks and the Core clock was frequently fluctuating during 3d rendering and eventually would stop the system. So, I RMA'd the card, Gigabyte sent me a new Rev 1.0 (Original cards were both Rev 1.0) and I tried it, which also froze the system.
I started to think, perhaps my PSU wasn't working properly, so I tested it, and showed 12.16v on all PCIE external power connectors even though OCCT and GPUz were showing 11.7v on my 12v rail. I wasn't sure if that was just the software reporting the incorrect values or if the motherboard is bad.
Anyway, I RMA'd the second card and they sent me a brand new Rev. 2.0 R9 280x, and I tried it, but this time, I put the new Rev. 2.0 card in the primary slot and the working Rev 1.0 in the secondary slot. Oddly enough, the Rev 2.0 card died which eliminated the possibility that the motherboard had a bad slot or something. After I RMA'd the Rev 2.0 they sent me another refurb Rev 1.0, I placed it in the secondary card slot and it was also producing errors with my system.
After the third RMA I told Gigabyte that I wanted a model upgrade, so I ended up with my R9 290x cards. Here's the kicker, When I connected the cards, one in slot 1 and the other in slot 3, I booted my system and immediately the system locked up on windows log in, I assumed this was a driver problem. Went into safemode and removed old driver install then reinstalled 14.9 catalyst drivers, it was fine at that point.
I started Firefall and tried that, it doesn't support Crossfire so there was a lot of flickering and horizontal lines scrolling down the screen vertically, after a while the game froze and I was able to recover by ending the process. I started OCCT and ran a video test in full screen mode, it reported both cards were running x16 2.0 which was accurate, but after a minute or two I started to see errors and the test froze, once again, I was able to get back to the desk top and recover.
So, I pulled the working 290x out and ran the presumed bad one by itself, it was fine in Windows, but as soon as I started running a 3d test the screen started getting horizontal black lines flashing here and there, eventually the system froze and I had to reboot. Funny thing is, after the reboot, my system thought there was something wrong with my D: drive and required a checkdisk. Every time the bad card crashed my system, I had to perform a Checkdisk on my D: drive (event viewer reported NTFS errors on that drive at shudown). Also, after system crashes, my WNA3100 USB 802.11N adapter would reinstall itself at adapter #2 like it had been connected to a different USB port but it wouldn't work after that (Code 10) in dev manager.
So, after disconnecting the USB device and using cmd to configure devmgr to show nonconnected devices and removing device #2, then reconnecting the USB adapter, it worked fine again. This also happened with the 280x's when I was running them. Anyway, If I ran BF4 with the bad card, all over the screen I'd see little blue splotches in a "^^^^^^" pattern all the way down my screen in horizontal lines. They would start to show up and then become more noticeable after a few frames went by, eventually (within 10 seconds) freeze the machine and I'd either have to close BF4 or hard reboot the PC.
I assumed the card was bad, so I removed it and put in the working 290x, which was fine the first day, then on the second day, I was playing Firefall and noticed little black squares in a "^^^^^" pattern that were showing up one or two rows at a time, but would blink in and out at different elevations on the monitor. I close Firefall and saw that the card was running at 1.18v in OC Guru, so I over volted it to 1.24v and the problem went away. But I'm still seeing some artifacting, like black sections in the game where there should be geometry, they don't have sharp edges, they're very smooth black areas on the screen, like the texture is missing or something like that. But it's not as bad, I had it running at 1.25 at that point, so I downed it to 1.21 and it seemed to go away.
Also, GPUz says under VDDC Current in 11.2A under load, but VDDC Current out says 94A. My 12v rail is 104 A, I'm concerned here as I'm not sure where this 90A reading is coming from, I thought these cards had a TDP of 290W, 94A would put it way over that. Anyway, I find it odd that I go through two Rev 1.0, one Rev 2.0 and a brand new 290x with the same or similar problem without there being something else wrong with the system.
Ohh, by the way, when I use Crossfire, or when I'm running the bad card by itself, most of the time now it forced an immediate shutdown without a crashdump or error report of any kind. After the system immediately shuts down, the power button doesn't work, I have to reset the switch on the PSU to get the system back up and running. Could this be a problem with the Motherboard that's frying the cards? Would it be relevant to say that it's not the motherboard as it always kills the secondary card in my crossfire configuration regardless of whether or not I put it in the secondary slot or the primary slot or that it only kills one of the cards and not the other?
I'm so confused here and I'm not sure how to proceed! I'm sorry for the long post as I'm sure 90% of you won't bother to read it, but those of you who do, any help would be greatly appreciated as I'm completely worn out on RMA processes and I'd like to avoid it entirely. Hopefully someone has had a problem like this before!
First off, I'm running (Trying to, anyways....) two Gigabyte R9 290x 4 gig cards in crossfire. I've got an Asus Sabertooth 990fx Revision 2.0 motherboard. Currently, I am only running one of my cards, as the other seems to be bad. But, since the 290x's aren't the first cards I've ran on this system, lets start from the top, shall we?
When I purchased this system, I started out with two R9 280x cards, they worked in crossfire for a few months, however, after discovering that my motherboard manual recommended using slot 1 and slot 3 for crossfire, I switched the cards to these slots. It took me a long time and a lot of tinkering to get it working again (frequent system shut downs). After a while the secondary card started giving me a lot of problems. So, I pulled the primary card and tested the secondary card in OCCT, during the video card tests it showed a lot of errors (in the thousands) after the first 30 seconds and then the system would freeze. It seemed the card couldn't hold the 3d mode clocks and the Core clock was frequently fluctuating during 3d rendering and eventually would stop the system. So, I RMA'd the card, Gigabyte sent me a new Rev 1.0 (Original cards were both Rev 1.0) and I tried it, which also froze the system.
I started to think, perhaps my PSU wasn't working properly, so I tested it, and showed 12.16v on all PCIE external power connectors even though OCCT and GPUz were showing 11.7v on my 12v rail. I wasn't sure if that was just the software reporting the incorrect values or if the motherboard is bad.
Anyway, I RMA'd the second card and they sent me a brand new Rev. 2.0 R9 280x, and I tried it, but this time, I put the new Rev. 2.0 card in the primary slot and the working Rev 1.0 in the secondary slot. Oddly enough, the Rev 2.0 card died which eliminated the possibility that the motherboard had a bad slot or something. After I RMA'd the Rev 2.0 they sent me another refurb Rev 1.0, I placed it in the secondary card slot and it was also producing errors with my system.
After the third RMA I told Gigabyte that I wanted a model upgrade, so I ended up with my R9 290x cards. Here's the kicker, When I connected the cards, one in slot 1 and the other in slot 3, I booted my system and immediately the system locked up on windows log in, I assumed this was a driver problem. Went into safemode and removed old driver install then reinstalled 14.9 catalyst drivers, it was fine at that point.
I started Firefall and tried that, it doesn't support Crossfire so there was a lot of flickering and horizontal lines scrolling down the screen vertically, after a while the game froze and I was able to recover by ending the process. I started OCCT and ran a video test in full screen mode, it reported both cards were running x16 2.0 which was accurate, but after a minute or two I started to see errors and the test froze, once again, I was able to get back to the desk top and recover.
So, I pulled the working 290x out and ran the presumed bad one by itself, it was fine in Windows, but as soon as I started running a 3d test the screen started getting horizontal black lines flashing here and there, eventually the system froze and I had to reboot. Funny thing is, after the reboot, my system thought there was something wrong with my D: drive and required a checkdisk. Every time the bad card crashed my system, I had to perform a Checkdisk on my D: drive (event viewer reported NTFS errors on that drive at shudown). Also, after system crashes, my WNA3100 USB 802.11N adapter would reinstall itself at adapter #2 like it had been connected to a different USB port but it wouldn't work after that (Code 10) in dev manager.
So, after disconnecting the USB device and using cmd to configure devmgr to show nonconnected devices and removing device #2, then reconnecting the USB adapter, it worked fine again. This also happened with the 280x's when I was running them. Anyway, If I ran BF4 with the bad card, all over the screen I'd see little blue splotches in a "^^^^^^" pattern all the way down my screen in horizontal lines. They would start to show up and then become more noticeable after a few frames went by, eventually (within 10 seconds) freeze the machine and I'd either have to close BF4 or hard reboot the PC.
I assumed the card was bad, so I removed it and put in the working 290x, which was fine the first day, then on the second day, I was playing Firefall and noticed little black squares in a "^^^^^" pattern that were showing up one or two rows at a time, but would blink in and out at different elevations on the monitor. I close Firefall and saw that the card was running at 1.18v in OC Guru, so I over volted it to 1.24v and the problem went away. But I'm still seeing some artifacting, like black sections in the game where there should be geometry, they don't have sharp edges, they're very smooth black areas on the screen, like the texture is missing or something like that. But it's not as bad, I had it running at 1.25 at that point, so I downed it to 1.21 and it seemed to go away.
Also, GPUz says under VDDC Current in 11.2A under load, but VDDC Current out says 94A. My 12v rail is 104 A, I'm concerned here as I'm not sure where this 90A reading is coming from, I thought these cards had a TDP of 290W, 94A would put it way over that. Anyway, I find it odd that I go through two Rev 1.0, one Rev 2.0 and a brand new 290x with the same or similar problem without there being something else wrong with the system.
Ohh, by the way, when I use Crossfire, or when I'm running the bad card by itself, most of the time now it forced an immediate shutdown without a crashdump or error report of any kind. After the system immediately shuts down, the power button doesn't work, I have to reset the switch on the PSU to get the system back up and running. Could this be a problem with the Motherboard that's frying the cards? Would it be relevant to say that it's not the motherboard as it always kills the secondary card in my crossfire configuration regardless of whether or not I put it in the secondary slot or the primary slot or that it only kills one of the cards and not the other?
I'm so confused here and I'm not sure how to proceed! I'm sorry for the long post as I'm sure 90% of you won't bother to read it, but those of you who do, any help would be greatly appreciated as I'm completely worn out on RMA processes and I'd like to avoid it entirely. Hopefully someone has had a problem like this before!