Well i was running my 7970 at 1125/1625 since april 12 with various drivers until last friday when my system started crashing cause of the overclock. I reinstalled the drivers, msi afterburner then the whole OS only to have it still crash. So I remove the overclock and it has worked fine all week. But now i want to overclock again and its doing it all again it will not overclock no matter what voltage. I am running on the 12.11 beta 11 and msi afterburner 2.3
I'm having a similar problem too: 2500k 4.7 ghz 7950 crossfire 850TX windows 8 I had my 7950 crossfire running at 1100mhz for a good month without any issue. I then started playing a game that only used one gpu for a week and when I went back to play something using both gpus my games lock up instantly if I use crossfire and that overclock. If I turn off overclock and use cards by themselves they work fine. Crossfire stock 900 mhz - Fine Crossfire 1000 mhz oc - semi stable, locks at times. can close by forcing task manager Crossfire 1000-1100 = instant ingame lock, can close by forcing task manager I took the cards out, swapped psu cables, placed the top card in other slot but it still does it. Reinstalled drivers and tried older 12.11 beta 8 but it stills occured. The reason why I think it might be the PSU: if i turn crossfire off and use the 1100hz overclock it doesn't matter which card is in the primary slot, they both work fine and I tested them for hours at 100% gpu usage. It's only if I turn the OC crossfire that it locks up (it doesn't lock the entire pc up, just the game. I can alt tab out and force it to quit).
@ Colt M4 Try uninstalling all drivers/overclocking utilities and then running genetix_cleanup, do not restart the pc after running the cleaning software or you'll have to enter safe mode and uninstall the standard vga drivers to avoid a black screen at boot. @ nsegative Have you disabled ULPS? The TX 850 is sufficient for an overclocked crossfire 7950 based system, assuming they use reference based 2x 75w PCI-e power connectors.
Yes, from here Genetix Cleaner v5.01. It should clean everything, just be sure to uninstall all third party overclocking utilities.
It turned out that it's my secondary IPS screen that's causing the problem. It's connected via hdmi into gpu and dvi into screen. As soon as I unplug it and play my 1100 clocks work fine. Plug it back in and games lock up. Not sure if ULPS would do anything but i'll look into it.
You'll find that ULPS is the first thing you'd want to disable for trouble free crossfire overclocking. The simplest way is to use Sapphire's Trixx utility as it has the option built in.
Disabling the ULPS with trixx solved my lock ups completely even with the 2nd screen using hdmi/dvi. Much thanks.
I'm glad it got fixed. I should also clarify that the TX 850 should be more than sufficient for an overclocked crossfire 7950 based system as the reference design only uses 2 x 75w PCi-e power cables rather than a single 75w and 150w that 79770 and some non-ref 7950 cards use.
Well I didn't use the genetix cleaner but I used driver sweeper 3.2 unistalled everything including ab then reinstalled 12.11 beta 11 but it still crashes. Its like the voltage falls off or isn't being applied. Going to try trixx Edit trixx doesn't work. According to tech up it isn't applying the right voltage or any voltage.
Maybe try a WHQL driver, like the 12.10 ones. Or have you already had AB 2.3.0 already running great with 12.11 BETA 11s?
I had it running with the beta 4 for about 4 or 5 weeks. I am going try the 12.10 then. How much slower is the 12.10 than the 12.11?
I don't know mate, I run 6000 series cards, so all the good stuff in the latest drivers doesn't really do anything for me as I believe its more for the GCN architecture. Your 7970 is a beast anyway, it'll run anything really well with or without 12.11 At least if you try with 12.10 and its OK it could be because AB 2.3.0 isn't cool with the BETA11 version of 12.11 and you may have to wait for say 2.3.1 to be released.
Assuming driver cleaner cleared out the system (I can not be sure as I do not use it) and 'tech up' is able to read the voltages being applied correctly and you understand the difference between actual voltage and set voltage (vdroop) and are using the applications accordingly then either ovp is kicking in or there is still some software conflict as 'applying' voltage should work. Whether it's stable and works in games is another matter. What is the stock voltage on the card? Perhaps providing a gpu-z sensors tab screenshot with the windowed render test could prove useful. Something like this. Also, if you have any doubts it's better to ask. You can not expect people to provide assistance if it means going around in circles because you are not willing to follow instructions.
Stock voltage is 1.112 https://www.dropbox.com/s/e78zzhsx7k1p4ey/Untitled.png I have always used driver sweeper it cleans out everything including the registery
I don't think either of us wants to discuss the pros and cons of what driver cleaning software to use and all such utilities are required to clean out the registry, though many do not recommend driver sweeper. A quick goggle search will likely highlight why, as my recommendation was based on experience. In any case, according to that screenshot all is well at stock clocks i.e voltages/temps etc. Another similar screenshot, but with the higher clocks and voltages applied through afterburner or trixx would be helpful.
Well it crashed to quick for me to get a screenshot but this is what it did. I applied 1.2 volts stock clocks started render test. Starts at 1.15 then 1.146 then 1.145, 1.143, 1.141, 1.136, 1.135, 1.134, 1.127, 1.126, 1.125, 1.124, 1.121, then last one i saw 1.107 crash all in about 45 seconds. Also i didn't use the genetix cleaner cause I don't feel comfortable using a command line program.
Being GUI based or command line based does not effect the utility's effectiveness, it's also only a matter of pressing a single key. I am not sure what clocks you applied for that test but that is less relevant, that does sound like there is a hardware error of sorts or the card is gradually limiting power as the load increases. Did you increase the power limit slider to +20%? I am also assuming you did not make any changes to the bios or drivers between the time when all worked well and the issues appeared? Edit: It could also be caused by overheating components as it does sound a lot like a protection of some sort kicking in.
yeah last friday it completely stopped working I had been using this version of afterburner with the 12.11 beta 4 since they came out. https://www.dropbox.com/s/6n8nzw5ptp82irn/Untitled1.png Not sure if this could be a cause since I never see my temps over 60 and my vrm temps don't go above 70 but I have a Prolimatech MK-26 installed but i was using that cooler for 5 weeks before this started.
That is most likely the cause if I was to guess. Something is probably overheating, try remounting or using the old cooler.