Hi.
I submitted an RMA for the Gigabyte 7850 OC Windforce edition in August and had it returned, tested as "not faulty" but I still get continual crashes when gaming. I'd like to point out at this point that it was an item I brought from ******, not OCUK. I'm just posting here for help and advice as the forums here are much, much better (and come next upgrade, I know where I'll be shopping too!)
The only time this PC will crash is under graphically intensive programs.
Anything from 5/10 minutes, to an hour of playing a modern game (League of Legends, Skyrim, Metro 2033, XCOM, Witcher 2 and Guild Wars - all have this problem) i'll get one of two crashes.
The most common one is when the screen turns all of one colour, and has vertical stripes running down the screen. This can be grey, blue, pink etc. This crash is usually accompanied by a tone, which i'm guessing is the last sound to come out of the speakers being stretched out continually.
The other crash is a simple black screen of death, and it cycles through the input on the monitor. Probably 90% of crashes are the first, the remainder are the second.
Only a hard reset will get the pc back again.
This ALWAYS happens, I have yet to get an hour in a game with this graphics card, yet my previous one, an HD 5770, I will not have this problem. I have switched the cards over and found no issue with the 5770, other than a diminished graphical performance.
My current PC set up is
Gigabyte GA-Z77-DS3H
Gigabyte HD 7850 OC Windforce X2 975MHz
Intel Core i5 3570K
OCZ 256GB Vertex 4 SATA III SSD
G-Skill 8GB (2x4GB) DDR3 1600Mhz RipjawsX Memory Kit CL9 (9-9-9-24) 1.5V
Antec TruePower Trio 650W PSU
All of these are brand new components with the exception of the PSU which is around 5 years old, and ran my previous rig perfectly.
Temperatures run hottest on the GPU at around 50 degrees before the crash, this I have used GPU-Z to log on a number of crashes and is the hottest it got, previous crashes have all been in the 40-50 degrees mark.
I have run Memtest x86+ and it showed no problems.
I have run superPi for around an hour with no problems
I have run furmark for around an hour, also with no problems
I have updated all drivers, and the motherboard BIOS to the most recent versions, and have used previous versions with the same crashing occuring. I have wiped drivers and reinstalled as well, and still the same error.
The only way I have been able to drop the frequency of the crashes is to underclock the card with the Catalyst Control Centre. By dropping the clocks on it the crash can usually be put back a few minutes, but they have always crashed in under an hour of gaming. Another way I have found to drop the frequency is to frequently alt-tab to the desktop for 30 seconds or so. I noticed this when alt-tabbing to check temperatures of the card as I gamed, and found that this also prolonged the period I could play in.
Could anyone offer any suggestions at this point? I'm wondering if it might be the older PSU, but it should still be rated for modern graphics cards I would have thought? Its still a good 650w supply after all.
I submitted an RMA for the Gigabyte 7850 OC Windforce edition in August and had it returned, tested as "not faulty" but I still get continual crashes when gaming. I'd like to point out at this point that it was an item I brought from ******, not OCUK. I'm just posting here for help and advice as the forums here are much, much better (and come next upgrade, I know where I'll be shopping too!)
The only time this PC will crash is under graphically intensive programs.
Anything from 5/10 minutes, to an hour of playing a modern game (League of Legends, Skyrim, Metro 2033, XCOM, Witcher 2 and Guild Wars - all have this problem) i'll get one of two crashes.
The most common one is when the screen turns all of one colour, and has vertical stripes running down the screen. This can be grey, blue, pink etc. This crash is usually accompanied by a tone, which i'm guessing is the last sound to come out of the speakers being stretched out continually.
The other crash is a simple black screen of death, and it cycles through the input on the monitor. Probably 90% of crashes are the first, the remainder are the second.
Only a hard reset will get the pc back again.
This ALWAYS happens, I have yet to get an hour in a game with this graphics card, yet my previous one, an HD 5770, I will not have this problem. I have switched the cards over and found no issue with the 5770, other than a diminished graphical performance.
My current PC set up is
Gigabyte GA-Z77-DS3H
Gigabyte HD 7850 OC Windforce X2 975MHz
Intel Core i5 3570K
OCZ 256GB Vertex 4 SATA III SSD
G-Skill 8GB (2x4GB) DDR3 1600Mhz RipjawsX Memory Kit CL9 (9-9-9-24) 1.5V
Antec TruePower Trio 650W PSU
All of these are brand new components with the exception of the PSU which is around 5 years old, and ran my previous rig perfectly.
Temperatures run hottest on the GPU at around 50 degrees before the crash, this I have used GPU-Z to log on a number of crashes and is the hottest it got, previous crashes have all been in the 40-50 degrees mark.
I have run Memtest x86+ and it showed no problems.
I have run superPi for around an hour with no problems
I have run furmark for around an hour, also with no problems
I have updated all drivers, and the motherboard BIOS to the most recent versions, and have used previous versions with the same crashing occuring. I have wiped drivers and reinstalled as well, and still the same error.
The only way I have been able to drop the frequency of the crashes is to underclock the card with the Catalyst Control Centre. By dropping the clocks on it the crash can usually be put back a few minutes, but they have always crashed in under an hour of gaming. Another way I have found to drop the frequency is to frequently alt-tab to the desktop for 30 seconds or so. I noticed this when alt-tabbing to check temperatures of the card as I gamed, and found that this also prolonged the period I could play in.
Could anyone offer any suggestions at this point? I'm wondering if it might be the older PSU, but it should still be rated for modern graphics cards I would have thought? Its still a good 650w supply after all.