Asus Strix 1070 OC - Memory Issue?

[Swinder]

Back in October of 2016 I picked up a 1070 OC with Samsung memory (dodged the micron bullet). I have had numerous issues which come and go but I am struggling now so a bit of help or advice is much needed.

On stock settings and with the gaming profile enabled via GPU Tweak2 I had issues playing Battlefield 1. I would get frequent driver crashes at varying lengths of played time. The fix was to drop the factory overclock and run the non-overclocked settings.

Since then I have dabbled in overclocking using the stickied thread by Nate152 and the benchmarks and stress testing were okay. I reinstalled BF1 and the issues have returned, once again the only way to player BF1 or F1 2016 is to drop down to the equivalent of a non-overclocked 1070 which isn't really ideal considering it cost more...

Win 7 x64.
Latest nvidia drivers (doesn't matter if I run old or new).
Bios: 86.04.50.00.63
PCIe x16 3.0.
Temps all good, 30c at idle, pushes circa 61c at load. (Custom fan curve to ramp up to 100% by 60c).

[Nate152]

Hi Swinder

Welcome to the ROG forum !

The driver crashing is usually because of too high of a core overclock.

Watch dogs 2 is kinda the same way for me with the memory, my memory overclock settings for the unigine valley test pass but for watch dogs 2 I have to lower the memory 200MHz or it crashes to the desktop.

Have you tried increasing the voltage to 100% ? Try Gaming Mode and OC Mode with the voltage at 100% and see if this helps. Increasing the voltage helps with the core clock, it has no effect on the memory.

Let me know how it goes with the voltage at 100%.

[Swinder]

Okay, ran as OC mode with following settings;

GPU Boost Clock - 1860MHz (+25Mhz)
GPU Voltage - 100%
Memory Clock - 8008MHz
Fan Speed - Custom Profile (60% @ 40c and 100% @ 70c)
Power Target - 110% (+10%)
GPU Temp Target - 83c
Frame Rate Target - 0

Load up F1 2016 and crash before going through menus to the game play itself. Nvidia driver crash and check device removed error.


Ran as Gaming mode with the following settings;

GPU Boost Clock - 1835MHz
GPU Voltage - 100%
Memory Clock - 8008MHz
Fan Speed - Custom Profile (60% @ 40c and 100% @ 70c)
Power Target - 100%
GPU Temp Target - 83c
Frame Rate Target - 0

Managed to get to the game play itself, completed a qualifying lap and then crash to desktop due to nvidia driver crash as I reached the next race.

[Nate152]

It's sounding like you might have a faulty gpu.

What psu do you have ?

You can fill in your pc specs and they'll show under your name like mine. Click the head in the white circle at the top right of this page then click inbox. On the left under "My Settings" click "Edit Profile", scroll and list your parts and click save.

[Swinder]

Spec detail added.

PSU is a Be Quiet Dark Power Pro 1000W - Overkill I know but I won it in a competition.

F1 2016 now seems to be playing nicely with those gaming settings posted above, getting some flickering across various menus and textures in game. Battlefield 1is still crashing but not as frequently.

[Nate152]

Thanks for your specs and you have plenty of power. Before we go calling the gpu faulty it could be a BF1 issue since you can pass the benchmark, I see alot people reporting the same issue you're having on the geforce forum and battlefield forum.

Do other games play ok ?

Try the hotfix driver 384.80 and see if this helps, after the license agreement check the "custom (advanced)" install and uncheck everything except the graphics driver and physx driver, then check the box that says perform a clean installation.

http://nvidia.custhelp.com/app/answers/detail/a_id/4522/~/geforce-hotfix-driver-version-384.80

[Korth]

What are your GPU temps when your game crashes? Maybe it's just a cooling issue.

And how does your BF1 play with HT disabled (in BIOS)?

[Nate152]

There is also an option in the game settings to disable directx 12, try disabling that.

[Swinder]

Okay, back to crashing in F1 2016 on gaming settings. This time the game has threw up a error relating to out of memory. The hotfix driver didn't change anything, the issue still persists.

The only other games I play are Counter Strike Global Offensive, Player Unknown's Battlegrounds and Diablo 3. These seem to run okay, if not a little down on FPS. All games I play default to DX11 as I am on Windows 7. I double checked the options for each game to make sure it isn't trying to utilise something that isn't available.

With the custom fan curve the GPU temps hover around 55c whilst the GPU is under load.

My girlfriends PC is identical with the exception of PSU and she has no problem with any of the games I have being playing. At some point I am going to switch the GPU over and see if the problem persist, at least then I can rule out any other hardware being the cause.

[Swinder]

One thing I want to try before filing an RMA is flash the older bios that had the micron issues, the catch is getting a copy of that which is legit.

If anyone has it could they upload/link it?