I just recently (5 days ago) got my Crossfire rig set up.
GPUs: 2x Gigabyte Windforce Radeon R9 270x
CPU: AMD fx 6300
Memory: 8gb G.SKILL Sniper RAM
Mobo: MSI 970A-G43
PSU: Raidmax Vampire 800w Gold PSU
Windows 8.1 Pro
(Nothing is OC'd by the way. Everything is stock speeds.)
So, everything was working fine with my single 270x until I decided to buy another 270x and try Crossfire for the first time. It worked pretty well for the first few days until about two days ago. I was playing Mass Effect 2 (definitely not a graphicly intensive game) when all of a sudden, my screen goes black, monitor goes into power save mode, and my fans go to 100% (at least thats what it sounded like). So I say, "Ok, no problem, I'll just reboot. Maybe MSI Afterburner or CCC (catalyst control center) messed up the fan speeds and let them overheat causing them to shutdown." So I boot up (about 10sec after I was forced to manually shut down the PC as I had no screen image) and look at my GPU temps. They were at a very resonable 30C on one and 34C on the other. So I started up another game. Game started ok, loaded the first level, played for about ten seconds and the shutdown happened again. Reboot, check temps, feel GPU (before I rebooted of course), everything seems fine. I reinstall my game, drivers, etc. Again same result. Tried a different game, same result. So I start to re-download my game to see if Origin or the game had something to do with the issue. I start to watch a recent Linus Tech video and again, about 10 sec in, the display goes black and my monitor goes into power save. I go into my PC, check all connections (they're ok), clean up hard drive, check CPU temp (again ok), try again. Same results. So I finally decide to disable Crossfire (both GPUs are still in) and then it works. No crash or anything, it works just fine. Re-enable X-Fire, crash. I've seen some people say it could be my RAM or CPU messing up. (It was working fine before though.) Maybe my CPU is bottlenecking? (It only gets to 78% Crysis 3 max settings). Please help. I've done all I can other than do extensive RAM and CPU error checking, taking it to a store, or calling AMD support. This is extremely frustrating for my first time doing it. I would like to mention I am also broke and would prefer not to spend any money unless absolutely neccessary.
Thanks so much guys, I love you
Check every one of your power connections and you may even want to check your power supply. At mas load these cards can potentially draw 300W or more since they have 2 6pin connectors and the 75W available via pcie slot. Checking drivers is good but you need to also go into either amd overdrive or msi afterburner and check your cards power usage and make sure your second card isn't in powersaver mode because iv'e heard that sometimes that can cause a crash(an otherwise helpfull feature if fixed). After checking those things I don't know what else could be the problem if you've made sure your drivers and games were good. If you didn't before, you could try reinstalling the drivers again but making sure to do a clean install of them unless that't what you did before when you reinstalled them. I hope this helps but if not i'm sure someone with more knowledge is on here.
And to clarify; 800W should be enough but if your power supply is older or doesn't carry enough amps on the rail(s) your using then that could be your problem. Good Luck
Maybe i'm a little off topic, but isn't the pci-e layout x16, x4 ? That's not exactly crossfire friendly, from what i remember, but that's in regards to performance, not stability.
80+ gold, but Raidmax...i don't know.
Google up how to disable ULPS, that may help.
I have disabled/reenabled ulps and tried forcing the voltage. Also tried to underclock them by about 40hz to see if the capacitors were maybe a little old and couldn't handle the full voltage.
pci-e layout is x16 for what I'm using. Unless you want a legacy card from 5+ years back, pci-e x16 is the standard. I know corsair/nzxt/cooler master are the PSUs to go for, but Raidmax has always been good to me and I got the thing on sale with a 30$ rebate. (it's also about a week old) I did a test on the throughput and it was getting 795-815watts over the course of a few hours.
I haven't tried looking at power save mode....that makes sense because my second gpu is running about 500mhz lower frequently
I literally bought it last week lol.
Here's the full specs if it helps
Type
ATX12V / EPS12V
-
Maximum Power
- Continuous 800 watts
-
Fans
- 135 mm Fan
-
PFC
- Yes
-
Main Connector
- 20+4Pin
-
+12V Rails
- 2
-
PCI-Express Connector
- 4 x 6+2-Pin
-
SATA Power Connector
- 8
-
SLI
- SLI Ready
-
CrossFire
- CrossFire Ready
-
Modular
- Modular
-
Efficiency
- up to 90.95%
- Energy-Efficient
- 80 PLUS GOLD Certified
-
Over Voltage Protection
- Yes
-
Overload Protection
- Yes
- Input Voltage
- 110 - 240 V
- Input Frequency Range
- 50/60 Hz
- Input Current
- 12A
- Output
- +3.3V@24A, +5V@20A, +12V1@35A, +12V2@35A, [email protected] ,5VSB@3A
-
MTBF
- >100,000 Hours
- Approvals
- UL, CUL, TUV, FCC, CE,
- Dimensions
- 3.50" x 6.00" x 6.50" (H x W x D)
- Weight
- 6.30 lbs.
What drivers are you using and what are your graphics setting in ccc set at? Every once in while i get a card thats set on the shelf so long that the tim is no good. You need to watch load temps verses idle. Hwin64 is a good tool. Could be alot of things? That psu causes me pause but in theory it Should be fine. I have seen a bad crossfire bridge cause problems before. Maybe try a different one?
After some reading, pci-e 2.0 x4 has enough bandwidth, especially for the 270 / 270x, crossfire gains should be ok, as the performance loss on x4 is very small. So my previous point is moot now :p