Overclocking my 1080ti, thoughts and suggestions!

[kdmasmkda]

Below I've linked my Unigine Superposition scores on extreme 1080p and I was wondering where they fall in line with the others! I was also looking for some advice on overclocking and "troubleshooting". My current OC currently stays around 1920MHz all the way up to 2000MHz for a few seconds. As far as voltage goes, it's pretty steady at 950mV but also reaches a max of a little over 1000mV. If I wanted the clock speed to be a steady 2000MHz during the benchmarking, should I raise my boost clock? Raise my voltage? Are these results perfectly fine for similar setups? Thanks! Feel free to share your results to.

 

[kdmasmkda]

If it means anything, the GPU utilization reaches 100% for just seconds before dropping down to 99% most of the time (hence the voltage dropping to 950mV and clocks dropping to ~1920MHz). I'm assuming this is normal and not due to voltages? Obviously 80MHz on the clock will have little difference in the real world (probably less than 2FPS), I'm aware.

[Sound_Barrier]

Just crank up your core clock until your game crashes, then decrease it until it stops for optimal performance.

[kdmasmkda]

11 minutes ago, Sound_Barrier said:

Just crank up your core clock until your game crashes, then decrease it until it stops for optimal performance.

My main issue is that the GPU doesn't reach 100% utilization, rather it only hits 99%, I don't see what raising the clock would do

Edit: Despite what I said, raising the clock did result in a slight increase. Don't wanna push it too far and need to uninstall the program in safe mode again! lol. 

[PokiDaSpitz]

52 minutes ago, kdmasmkda said:

My main issue is that the GPU doesn't reach 100% utilization, rather it only hits 99%, I don't see what raising the clock would do

Edit: Despite what I said, raising the clock did result in a slight increase. Don't wanna push it too far and need to uninstall the program in safe mode again! lol. 

sometimes it's cause of the software but 99% is the same as 100% so don't worry there shouldn't be any bottlenecks of the cpu as you're using a i7-8700k.

If you actually pushed your gpu too far you would see artifacts (like weird colours, glitchy screen)

[kdmasmkda]

2 minutes ago, PokiDaSpitz said:

sometimes it's cause of the software but 99% is the same as 100% so don't worry there shouldn't be any bottlenecks of the cpu as you're using a i7-8700k.

Yeah the CPU does just fine @ 5GHz but at 2000MHz on the 1080ti it seems a bit unstable as far as keeping it steady. It either goes up to 2000MHz and goes right back down to 1920-1950MHz or if I raise the clock speed it will stay at 1900-2000 and shoot up past 2000 causing it to crash, either way, it's not like it ever stays steady.

[PokiDaSpitz]

2 minutes ago, kdmasmkda said:

Yeah the CPU does just fine @ 5GHz but at 2000MHz on the 1080ti it seems a bit unstable as far as keeping it steady. It either goes up to 2000MHz and goes right back down to 1920-1950MHz or if I raise the clock speed it will stay at 1900-2000 and shoot up past 2000 causing it to crash, either way, it's not like it ever stays steady.

Sometimes the speed goes down cause it's getting hot. what are the temps?

[kdmasmkda]

1 minute ago, PokiDaSpitz said:

Sometimes the speed goes down cause it's getting hot. what are the temps?

Stays below 75 at all times as you see in the results, it's usually at 70-72c on full load.

[PokiDaSpitz]

1 minute ago, kdmasmkda said:

Stays below 75 at all times as you see in the results.

oh sorry my bad... I guess it's not that stable haha.

[Spotty]

2 hours ago, kdmasmkda said:

Below I've linked my Unigine Superposition scores on extreme 1080p and I was wondering where they fall in line with the others! I was also looking for some advice on overclocking and "troubleshooting". My current OC currently stays around 1920MHz all the way up to 2000MHz for a few seconds. As far as voltage goes, it's pretty steady at 950mV but also reaches a max of a little over 1000mV. If I wanted the clock speed to be a steady 2000MHz during the benchmarking, should I raise my boost clock? Raise my voltage? Are these results perfectly fine for similar setups? Thanks! Feel free to share your results to.

Overclocking on the 1080ti is pretty pointless in my opinion. The GPU Boost does a pretty good job of overclocking things by itself without you needing to do anything, and will still hit 2000Mhz as long as temperatures are not an issue.

This is my GTX 1080ti with a i7 6700k with stock settings on the card.

[jones177]

I go along with Spotty.

GPU EVGA SC2

Going by these samples, the average frame rate is what's important.

[Spotty]

9 minutes ago, jones177 said:

GPU EVGA SC2

You have an 8086k as well? Nice pairing. You should drop your PC specs in your sig and show it off.

And yeah, the average is probably the best measurement. The min/max FPS readings may have only been for a couple of frames. Anything like another program running in the background could cause a minor drop in the low FPS as well.

[kdmasmkda]

4 minutes ago, Spotty said:

You have an 8086k as well? Nice pairing. You should drop your PC specs in your sig and show it off.

And yeah, the average is probably the best measurement. The min/max FPS readings may have only been for a couple of frames. Anything like another program running in the background could cause a minor drop in the low FPS as well.

Like I said, it's truly only a couple fps difference at most in the real world, just look at @jones177! Thank you both for sharing your results, It's definitely interesting

[MkaiL]

Here are my results with ambient 27c. First is gpu boost at 2000mhz and second is oc'ed to 2126mhz (on master bios of my ftw3, haven't pushed max oc on slave bios yet). Fps are almost the same. Pascal cards lower their clocks depending on temps, I believe it downclocks every 8c. My card is watercooled btw.  

[jones177]

55 minutes ago, MkaiL said:

Here are my results with ambient 27c. First is gpu boost at 2000mhz and second is oc'ed to 2126mhz (on master bios of my ftw3, haven't pushed max oc on slave bios yet). Fps are almost the same. Pascal cards lower their clocks depending on temps, I believe it downclocks every 8c. My card is watercooled btw.  

 

My computer would blow up with a ambient temp 27c & I would melt. 

 

I didn't get the ftw3 because I was not planning on overclocking. If I get a third GTX 1080 ti it will be a ftw3.

 

4 hours ago, Spotty said:

You have an 8086k as well? Nice pairing. You should drop your PC specs in your sig and show it off.

And yeah, the average is probably the best measurement. The min/max FPS readings may have only been for a couple of frames. Anything like another program running in the background could cause a minor drop in the low FPS as well.

I will give it a go. 

 

5 hours ago, kdmasmkda said:

Like I said, it's truly only a couple fps difference at most in the real world, just look at @jones177! Thank you both for sharing your results, It's definitely interesting

Now that there is a post with a score that starts with a 6 we have some work to do.

 

[Froody129]

This is what you show people when they ask you what you bench 

[kdmasmkda]

10 hours ago, MkaiL said:

Here are my results with ambient 27c. First is gpu boost at 2000mhz and second is oc'ed to 2126mhz (on master bios of my ftw3, haven't pushed max oc on slave bios yet). Fps are almost the same. Pascal cards lower their clocks depending on temps, I believe it downclocks every 8c. My card is watercooled btw.  

Those temps are insane!I also have a 26-27c ambient temp but my card could never reach those temps at load. What's your watercooling setup out of curiosity? (I wouldn't be able to fit it in my case anyways)

 

Edit: After seeing these results, I definitely considered it but It's just not worth the extra cost when there's little to have other than a couple fps and lower temps. Another day

[MkaiL]

1 hour ago, kdmasmkda said:

Those temps are insane!I also have a 26-27c ambient temp but my card could never reach those temps at load. What's your watercooling setup out of curiosity? (I wouldn't be able to fit it in my case anyways)

 

Edit: After seeing these results, I definitely considered it but It's just not worth the extra cost when there's little to have other than a couple fps and lower temps. Another day

Yeah, the performance gain is not that big in terms of fps but I primarily did it for silence and also to keep me busy with my pc.

 

My setup: hardline loop in a modded phanteks evolv (for better airflow) with 360 + 240 gts hardwarelabs rads, ek full nickel cpu block + gpu block + integrated vrm block on motherboard (asus formula). If you want to see the entire build, just click on "The Black Pearl" in my signature.