Video Card Issues- GTX 1070 Horizontal Lines, please help!

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Hi,

I have a brand new (2 week old) MSI GTX 1070 Gaming 8GB. It has worked great since I got it, but yesterday I started intermittently getting horizontal black and white lines flickering across the screen. It was only occurring on one of my monitors (ASUS 144Hz 1080p), so I switched out the cable for a different DisplayPort Cable. It seemed to work fine but then it started again. Today, it has gotten much worse. It has also started doing the same thing on my second monitor, an LG 1080p ISP display.

I got the card through amazon, so I can still easily exchange it. My question is whether there is another cause of this besides the video card. I'm on a completely new build and all items can still be exchanged through the vendor, so I'd like to exchange anything that may be the cause before that exchange period ends. I'm linking a video to what the screen looks like. A few minutes ago I tested the graphics card using Furmark, and there were zero issues. However, when I turned off furkmark it started doing the screen tearing thing continuously.

Basically, is it possible this issue is caused by the motherboard, CPU, or RAM? I am planning to run memory test overnight tonight to find any issues that may come up. I have until October 12 to exchange any items and would like to solve this issue before then.

https://goo.gl/photos/BnGAmcq7Gmhe2rEV9

PC Specs are as follows:
-MSI GTX 1070 Gaming 8G
-Asus Z170-A Motherboard
-G.Skill DDR4 2666MhZ 16GB (2x8GB)
-Rosewill Capstone 450W PSU **This PSU is from my old build and I ordered a 650W EVGA SuperNOVA G2 to replace it in case it is causing issues**
-Samsung 850 EVO 240GB SSD Boot drive, 850 EVO 500GB Application drive, 1TB WD Black Storage drive
-ASUS VG248QE 144Hz 1080p monitor, connected with Displayport 1.2 Cable
-LG 24MP56 IPS 1080p 60Hz Monitor, connected with HDMI
-Windows 10 64-bit w/ Anniversary Update

All display and motherboard drivers were updated today.

Thanks in advance!!!

-Nikolai H
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Good idea. Running the memory test now but in the morning I'll try that. It seems weird to me that furkmark wouldn't cause issues but after stopping the test it would get really bad.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Another question- this card is very heavy. It has a thick backplate to keep it sturdy but is there anything else I should be doing to keep the card supported? Maybe extra stress on the PCI slot is causing communication issues?
 

master_shake_

Diamond Member
May 22, 2012
6,430
291
121
Another question- this card is very heavy. It has a thick backplate to keep it sturdy but is there anything else I should be doing to keep the card supported? Maybe extra stress on the PCI slot is causing communication issues?

try propping it up if you think it's causing an issue.

my 980's have back plates and i have never had that issue.
 
  • Like
Reactions: Niksterberg123

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Okay! It doesn't look like it's drooping I just didn't know if that was common practice.

Is there any chance this could be CPU related?
Forgot to mention I have an i5 6600k
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Okay thank you.

I put the card in the second slot of the motherboard and so far it has not had any issues. I'm beginning to believe the motherboard PCI slot is the issue. Memory tests came up with no issues. I'm going to replace the motherboard and hope it resolves all issues. I also found a graphics card support bracket I may purchase to help keep it supported.

Thank you all for the quick replies and help!!!!
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
I am now thoroughly confused.

I replaced the motherboard and graphics card. Amazon for the win on same day shipping and no hassle replacements!

However, my monitor is again flickering. New PSU as well. I don't really know where to go from there. I have seen the flickering on both monitors and used two cables. CPU? RAM? Ram tested overnight and didn't have errors. Maybe it wasn't tested long enough? Is it possible that the wall power is unstable and causing it?

Any suggestions would be great. Thank you for all the help!

Edit: Is it possible that the newest NVIDIA drivers are the problem? I imagine more would have this issue. I may try to revert back. Maybe I'm just senile and believing my second monitor is being affected when its really just the ASUS 144Hz monitor. I'm going to look into an RMA with ASUS.
 

[DHT]Osiris

Lifer
Dec 15, 2015
14,003
12,067
146
Power from the wall is pretty unlikely, the PSU will smooth over any bumps, and if it's more than a bump it's gonna turn your system off. Ram doesn't generally cause graphical corruption (especially when not being taxed) so I'd leer away from that, and CPU... those generally cause fatal crashes. First thing I'd do is take one monitor out of the equation, test both of then independently with different connection types if you can (DP, HDMI, DVI) and see if it's reproducible like that. If it isn't, there could be an issue with that card running dual display (as it's reproduced with two units). You could also test a few diff driver versions (go back a full release version or two, test a few) to see if it goes away with the dual monitor setup on separate versions.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Okay, so I'm not sure if anyone cares but I finally found the issue. Apparently many of the 1070 manufacturers started using Micron GDDR5 memory in their graphics cards in lieu of Samsung. Apparently my issue is known to occur with micron memory not being optimized. There SHOULD be a firmware update from MSI soon.

I have until November 2nd to return my current graphics card to Amazon. Would you recommend I replace it or wait for a hopeful fix with the Bios update? I could go with a different brand that only uses Samsung memory. I read earlier that MSI had a major recall in China on their 1070s and 1080s, not sure if it was the same reason. If I do replace it, do you have any model recommendations?

Thanks!

-Nikolai
 

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
Argh....TBH I am always cringing when people randomly replace hardware based on blind guessing. (No offense).

So, if this is related to the "Micron Memory" issue, I would not RMA/exchange the card. Why?

* Chances are slim you will get a card with Samsung memory
* The problem is not so much "inferior" ram (not even remotely, Micron is actually good), but how the BIOS/card supplies voltage to this memory when overclocking.
* It can be solved by a bios update, and also using workarounds with MSI Afterburner, Precision X etc. setting idle voltage higher.
I personally would NOT go through the hassles of replacing w/ this particular issue. I would use the MSI Afterburner workaround (check MSI forum, someone there explained how it's done) and in the meantime wait until they release a final bios.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
No offense taken, haha. I went crazy with replacements because Amazon is extremely easy to exchange parts with, and before I found out this micron memory issue, I was unable to properly diagnose it myself.

Good thing to know I can just increase voltage and fix the problem for the time being. I'll read more into that. Thanks for the tips. Hoping for an update shortly.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
In light of not wanting to sound like an idiot after your first post flexy, I failed to mention that I had already replaced the graphics card with an EVGA model. It arrived today. It has Samsung GDDR5 memory.

I am having the exact same issues as with the MSI card(s) but now more frequent. I have updated the BIOS of the graphics card to the newest version. I am honestly completely dumbfounded about what could be causing the issue. It is not MOBO, not GPU, not PSU. I HAS to be the RAM or CPU. I cannot imagine a CPU causing these issues (please correct me if I'm wrong). So, RAM? Does this sort of issue make any sense that it could be caused by the RAM? I am frustrated beyond belief at this point. I just want my $1000+ dollar computer to work.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
The results of monitor tests:

Only ASUS 144Hz monitor plugged in:
No flickering, ever

Only LG IPS monitor:
No flickering, ever

Both monitors:
Pretty bad flickering on ASUS
Occasional, rare flicker on LG monitor

If I unplug my LG monitor's HDMI cable and plug it in, the flicker stops on both screens.
If I unplug my ASUS monitor's DP cable and plug it back in, flickering continues.

Is it possible that the LG monitor is going bad and it is causing some sort of feedback through the HDMI cable to the graphics card, therefore messing up my ASUS monitor?

I have tried two display port cables and one DVI-D cable on the ASUS monitor, same issue with all 3 cables.

I have tried two different HDMI cables on the LG monitor, same issue.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
I'm not sure you can just disable the display port on a monitor like that...
My LG monitor has only HDMI and VGA, as it is older.

Another update to testing:
Today I ran Prime95 and FurMark. When I ran the Prime95 blended test (lots of ram tested), I started the test and immediately started furmark. FurMark ran at <5 fps and both screens went completely fuzzy with flickering. After about a minute I stopped both test and it went back to normal, with occasional flickering.

Next, I ran Prime95's small FFT test (little RAM tested), and it got slightly worse flickering for the first 10 seconds with <10 fps on Furmark, then it stabilized and FurMark ran 100+ FPS and Prime95 ran without issue... and little to no flickering.

This makes me think it is the RAM. Who I am to know anymore, though. I'm going to run another series of RAM tests tonight and see what happens.
 

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
A) Stop using Furmark. It is really worthless except as a tool to "heat up" your GPU. Furmark lost its usefulness long ago. For stability testing your GPU, use Unigine Heaven

B) Don't use Furmark and any other bench AT THE SAME TIME. Test your memory with memtest, CPU/mem with Prime or with OCCT. If Prime or OCCT passes, your CPU is ok. If memtest passes, your memory is ok. If Heaven runs for some time without crapping out, your GPU is ok.

C) make a custom resolution with NV control panel or CRU (custom resolution utility) in case your current res is a "HDMI - HDTV" resolution. Just curious what happens if you force the PC to see it as a "PC resolution" which you can with a custom resolution SHOULD it currently be listed as "HDMI".

Add:

Did you already replace the PSU? Which one do you use? Do you have all required cables on the GU connected, PCIE etc. I mean, just wondering.
 
Last edited:

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Yep, brand new PSU. EVGA Supernova G2 650W. It is the exact same issue before and after replacing the PSU. All cables are correct.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Also thank you for the testing methodology. I think in my stupor and frustration to solve this I'm not thinking in a logical manner.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Individually removing ram and testing one at a time, i still have the issue. Going to test as advised.
 
Last edited:

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
I will look into changing display port/HDMI protocol.

I ran memtest86 overnight (~7 hours) and it ran 4 full passes with zero issues. Is this enough time to effectively rule out memory issues?

I am currently running prime95 while at work. I watched it for the first hour and temps stayed below 60C and it didn't have any faults for the first hour or so.

I attempted to make a custom resolution but it keeps saying that a 1920x1080 144hz resolution already exists.

In the graphics settings for my LG monitor (HDMI), I only have an option for 1080p in the TV category in Nvidia control panel, not in the 'PC' category. The highest resolution in the PC category is 1800x1050 or something thereabouts. I tried to make a custom PC resolution for it but once again Nvidia said the PC resolution already existed.

When I unplug my LG monitor, the 144hz monitor reverts to 1920x1080 60hz (from 144hz) and the flickering stops. Maybe the 144hz setting is causing the issues, although with my last PC (gtx660, i5 4670k) I never had issues at 144hz w/ Nvidia control panel.

Maybe my Asus monitor is in fact going bad. Seems odd that it wouldn't cause issues with the LG unplugged (when I switched it back to 144hz it still seemed fine). I guess I can run it for a few days as a single monitor at 144hz and see if it works okay. I also find it odd that this issue immediately cropped up when I got a new computer, the monitor worked great before that.

Ill run uniengine graphics test tonight and see how it works out.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Okay thanks, I'll look through the monitor settings.
I ran prime95 blended test for 9 hours today with no issues. Seems RAM and CPU are fine. Now to test GPU.
 

Niksterberg123

Junior Member
Oct 1, 2016
18
0
6
Just ran Unigine Valley for ~1 hour and it kept framerates between 75 and 120 on ultra w/ 8x MSAA. Card didn't break 51C. I think the GPU is fine.

So far today I haven't experienced the issue again after a few modifications. I switched to the default Windows 10 theme and everything seemed to go away. I haven't restarted the computer, so they may come back, but I read on another thread that someone had a similar experience where using the default theme fixed it. I have no idea why that could be.

Edit: It is back. I am just going to run only the 144hz monitor for a few days and see if it ever has issues. If it does, I'm going to probably RMA the monitor. If it doesn't, well I'm not sure what to do then.
 
Last edited: