GPU Overclocking Benchmark Crash? (Beginner)

Toggle sidebar Toggle sidebar
T

Turtles_321

Commendable
Aug 6, 2016
7
0
1,510
Hello, I have an MSI GTX 1080 Seahawk that i am trying to overclock on a 1440 resolution (CPU is i5 6600k). I am a beginner, so I am following a youtube video guide on overclocking the card with MSI Afterburner and Uniengine benchmarks (He used Valley). According to the instructor, all the 1080 cards should be able to hit 200+ on the core clock just fine. He suggest the person experiment with +20 till they start to see artifacts or black lines in the benchmark. I can run the core clock at 200+ just fine, but when I bump it to 220+ the benchmark crashes.

Should I treat this as the black-lines and artifacts and dial down the overclock? I'm aware that some graphics card overclock better than others, but it seems odd that I can't even run the benchmark by just adding +20 to the core clock.
 
Solution
VinkiVoble
Ignore the window test if i were you are attempt to run some games, benchmark programs like Valley and 3D mark tend to not reflect actual performance or stability in a real world application. I suggest maybe running some in game benchmarks like the built in ones in GTA V and Tomb Raider you can just set them to run without having to actually play the game.
Sort by date Sort by votes
VinkiVoble

VinkiVoble

Commendable
Mar 30, 2016
41
0
1,560
I would suggest starting with changing the power and temp limits to max and leaving the offset at +0 then seeing how the high the gpu clocks on its own first.

I am not sure if a custom fan curve benifits a hybrid cooler like the seahawk but i did notice that allowing the fan to go around 75% on my founders edition 1080 got it boosting to 1900+ mhz out of the box on its own.

Just Let GPU Boost do its own thing and see if you still need to go further.
 
Upvote 0 Downvote
T

Turtles_321

Commendable
Aug 6, 2016
7
0
1,510
Yeah, I tried increasing the power and it runs fine. But figuring out how the core clock and memory clock work is becoming frustrating. I'm not even sure what counts as stable since I only have experienced crashes or long stutters on my end. Everyone seems to get artifacts or black-lines on their screen to determine their limit.
 
Upvote 0 Downvote
T

Turtles_321

Commendable
Aug 6, 2016
7
0
1,510


Only raising the power limit and running Valley Benchmark:
Core Clock: 1987
Memory Clock: 5005
 
Upvote 0 Downvote
VinkiVoble

VinkiVoble

Commendable
Mar 30, 2016
41
0
1,560
Nice i would think that that would give you ideal fps in 1440p but if you really do need to overclock set the offset so its boosting around 2000 ish mhz not every card will reach the 2100mhz people suggest its really luck of the draw with these things. Here is a screenshot of my current settings in afterburner and i find this provides me with a solid 60fps in games like BF4 and GTA V at 4K.

http://i.imgur.com/azCLtPS.jpg
 
Upvote 0 Downvote
T

Turtles_321

Commendable
Aug 6, 2016
7
0
1,510
What I want to know is how you got there. I don't have experience with overclocking GPU, and just learned how to overclock CPU. When I run valley benchmark, the only thing I experience is stutter or a crash (just valley benchmark). From what I have seen most people are able to determine their limit from black lines or the blue screen of death. Does the stutter and benchmark crash count as the same thing?
 
Upvote 0 Downvote
VinkiVoble

VinkiVoble

Commendable
Mar 30, 2016
41
0
1,560
I would take any sort of crash as a sign of reaching a max stable clock the most common i find is a display driver crash. Rather than using specific benching programs like valley i always test in a real game to see the actual results of the clock and get a much better idea of how stable it is, if it crashes or you see any sort of artifact it is a sign you have gone too far.
 
Upvote 0 Downvote
T

Turtles_321

Commendable
Aug 6, 2016
7
0
1,510
Interesting. I tried your overclock on the core clock (+175) and it crashes valley on window mode (+150 is fine). However, if I run valley in full screen it runs fine with 1 random hiccups in between scenes. Should I ignore the window test? I know you recommend actually running games, but I find it a lot harder since I have to interact with it.
 
Upvote 0 Downvote
VinkiVoble

VinkiVoble

Commendable
Mar 30, 2016
41
0
1,560
Ignore the window test if i were you are attempt to run some games, benchmark programs like Valley and 3D mark tend to not reflect actual performance or stability in a real world application. I suggest maybe running some in game benchmarks like the built in ones in GTA V and Tomb Raider you can just set them to run without having to actually play the game.
 
Upvote 0 Downvote
Solution
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom