This is going to be a lenghty post, sorry about that but english is not my mother language and I'm not particularly good in summing things up.THIS
is one thread, there were others but I can't find them right now, I probably have them bookmarked at home. It's only 2 pages, please read all the posts because the thread starts talking about a different thing.
Talking about me I've been performing tests since last week, when I received my new PC. Won't list all the components but the main one are 3770 CPU and Asrock Z77 Extreme 4.
The issues is about appearent driver reset issues, BSODs, freezes and crashes whenever you use the videocard. They happen randomly even in idle, but if you run a game or a 3D benchmark they happen 100% of the time. The rest of the system is a bit slowed down anyway even without the errors, you notice that by moving windows around with Aero active etc.
I started testing every single bit in my PC, let me name some of the tests I performed in this week
1) Multiple clean re-installations of Windows 7 Professional x64 (including all winupdate patches)
2) Multiple clean re-install of drivers, both WHQL and beta
3) Used only the latest chipset and periferal drivers from Asrock website, not the CD
4) Tested CPU with Prime95, CPU-Z and superpi, multiple cycles
5) Tested RAM with Prime95, Memtest windows and Memtest86+ Iso boot
6) Tested multiple power cables and multiple PSUs
7) Tested the videocard on the second videocard PCIE slot of my motherboeard
Tested my videocard in an another PC (my old one, E8400 + Asus P5Q-Pro)
9) Tested my old system videocard in my PC (GTX 460, Gigabyte Windforce)
10) Tested another GTX 670 (a friend's) in my system
11) Tested my GTX 670 on the first slot but capping the speed of the PCIE to gen1 (PCIE 1.1)
None of the test showed any sort of error.
The strange part is that my GTX 670 Videocard was working perfectly on my older system and on the second slot, but I couldn't understand why.
I started to think maybe my first PCIE slot on the motherboard was broken, but then the older 460 and my friend's 670 were working perfectly on that slot.
Then I started to ask what was the difference, and I found it out.
My GTX 670 in the second slot runs at PCIE 3.0 8x
On the older system instead it runs on PCIE 2.0
While with the capped gen1 bus of course it runs on the first slot at PCIE 1.1
Nobody in the typical HW forums I surf knew anything about this issue, and I have some friends with my same CPU, my same motherboard and my same GPU who experienced no issues at all, showing that it wasn't an issue affecting ALL MODELS of GTX 670 Windforce.
Then I found that topic on Anandtech and other topics, and it was clear, the issue happens only when the videocard runs at PCIE 3.0 16x and appearently only on Z77 chipset motherboards.
My reseller is offering me RMA, altough they are talking about 2-3 months time to send the card back to China (actually Taiwan I guess?)
In all this time I won't have a videocard to run my system, and when it will come back to me I won't be sure that it's not another defective model.
Now... if someone told me It will be fixed in the near future with a new Bios
it wouldn't be a big deal for me to wait until then and keep it in PCIE 1.1 16x in the meantime. But the thing is that I'm not even sure that such a strange issue can be fixed with a new bios release, or that a new GPU bios will ever be released at all.
So yeah, this is why I was wondering if it could be possible for me to ask for a faster and more efficient RMA directly to Gigabyte.
I'm willing to give further information by phone or email if you want me to.
I already submitted a ticket yesterday of course
Thanks for remindin me though.