• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

3080 Undervolting

Is everyong using a custom curve or just alterting core voltage?

My original curve -
3080-Undervolt-Original.jpg


And after reading the sucess of lower voltages I will start nudging the lower nodes up -
3080-Undervolt-Current.jpg


What do your curves look like? ;)

I forgot that I also have a +450 on the memory. I'll need to prove to myself again that it does as intended.

You shouldn't overclock memory whilst trying to find a stable core.

Do core only. Find a stable core. Then try clocking the ram.

You may find one makes an impact on the other. As an example. Your plus 400 on the memory could be making your core need more volts to be stable.

Plus memory makes such a small difference it's best leaving it stock anyway until you have your core sorted then see if it's even worthwhile.
 
When I undervolted my 3090 it was constantly hitting the power limit and occasionally crashing unless I supplied it enough volts to make the underclock pretty pointless. Overclocking and raising the power limit made little real world difference apart from more heat and fan noise. As you say, stock works no matter the scenario. It's almost like they have a bunch of quite bright people who figure this stuff out.
:D - Yeh, exactly

I expect mine was doing the same, not enough volts when required. I also tried overclocking but it seemed pointless too and only created added heat and noise

They don't turnover $10 billion for no reason
 
:D - Yeh, exactly

I expect mine was doing the same, not enough volts when required. I also tried overclocking but it seemed pointless too and only created added heat and noise

They don't turnover $10 billion for no reason

Overclocking is pointless which is why this thread is about undervolting.

The video explains this pretty well.

I'm sure you are in the minority that thinks that this isn't useful. You got poor silicone nobody can really make up for that.

I was having issues using CTR for my ryzen CPU. Ended up just using PBO and it's stable. Undervolting CPU with CTR for me wasn't stable
 
Overclocking is pointless which is why this thread is about undervolting.

The video explains this pretty well.

I'm sure you are in the minority that thinks that this isn't useful. You got poor silicone nobody can really make up for that.

I was having issues using CTR for my ryzen CPU. Ended up just using PBO and it's stable. Undervolting CPU with CTR for me wasn't stable

Undervolting the GPU is also not stable in certain games. As we both said, if enough voltage cannot be supplied when required, it will crash. I lock my FPS to 97 so most of the time the card sits at 60c, probably way less silicone use than someone running much higher FPS with an undervolt and also more stable
 
Undervolting the GPU is also not stable in certain games. As we both said, if enough voltage cannot be supplied when required, it will crash. I lock my FPS to 97 so most of the time the card sits at 60c, probably way less silicone use than someone running much higher FPS with an undervolt and also more stable

100hz monitor?

My monitor is 240hz so I can't really lock it to 97hz. My undervolt uses 80w less power so it runs pretty cool thanks to that. I get more power than stock too and yes it's stable enough for me.

I have ran several runs of 3D mark port royal. I have also played 3 different games so far for a few hours each and no crashes in any of them.

That's good enough for me. Unfortunately no I don't have cyberpunk to test that out. But I'm pretty sure the many threads I looked at on Reddit some of them must have it to test it out and they all seemed happy with their undervolting too.
 
I'm sure you are in the minority that thinks that this isn't useful. You got poor silicone nobody can really make up for that.

Depends on your goal and what games or software you're running.

Stable in synthetic benchmarks is different to stable running RTX, which is different to stable running RTX and DLSS which is different again to running an old game or a VR game with neither at higher boost clocks. Then you've got rendering with CUDA and rendering with Optix.

Basically we're men/women/attack helicopters in a shed playing with the exposed settings and never being 100% sure what's going on under the hood.

Then, how do you define stable? Mostly didn't crash? I run the GPU in the high 90s to 100% ultilisation whether gaming or rendering and I've found after many tedious hours of tinkering that bouncing off the power limit promotes instability or variances in frame time that can be seen as stutter in game.

How much benefit can you get from undervolting while keeping the same performance? I managed to save 30w total system draw. Medium whoop.

If you're prepared to sacrifice performance for lower heat and noise then go for it but you're going to have to go down into the 800s to do it. Or you could not spend the hours I've spent dicking about, run it stock and play some games or make naked elf pictures.:)
 
Depends on your goal and what games or software you're running.

Stable in synthetic benchmarks is different to stable running RTX, which is different to stable running RTX and DLSS which is different again to running an old game or a VR game with neither at higher boost clocks. Then you've got rendering with CUDA and rendering with Optix.

Basically we're men/women/attack helicopters in a shed playing with the exposed settings and never being 100% sure what's going on under the hood.

Then, how do you define stable? Mostly didn't crash? I run the GPU in the high 90s to 100% ultilisation whether gaming or rendering and I've found after many tedious hours of tinkering that bouncing off the power limit promotes instability or variances in frame time that can be seen as stutter in game.

How much benefit can you get from undervolting while keeping the same performance? I managed to save 30w total system draw. Medium whoop.

If you're prepared to sacrifice performance for lower heat and noise then go for it but you're going to have to go down into the 800s to do it. Or you could not spend the hours I've spent dicking about, run it stock and play some games or make naked elf pictures.:)

Port royal is a ray tracing benchmark fyi.

I literally spent all of 5 mins not months overclocking. I watched the video. Did some reading on Reddit and just put the video settings in.

I used MSI afterburner and hwinfo and I'm getting 80w less power draw. It runs the games I play. That's stable enough for me. If it wasn't stable then sure I would just run stock or up the Voltage and try again but it's been fine so far.
 
My 3060ti undervolts like a champ and benchmarks / stresstests fine.. will not run CP2007 though, have to revert to stock !!

Anyone else have this issue ?
 
Port royal is a ray tracing benchmark fyi.

I literally spent all of 5 mins not months overclocking. I watched the video. Did some reading on Reddit and just put the video settings in.

I used MSI afterburner and hwinfo and I'm getting 80w less power draw. It runs the games I play. That's stable enough for me. If it wasn't stable then sure I would just run stock or up the Voltage and try again but it's been fine so far.

Furry Muff, if you're happy then you're happy. ;)

I ran a lot of passes of Port Royal.
1950/900 got me a 3% higher score. Overclocking got me 2% more. Neither translated to a better real world experience.
 
100hz monitor?

My monitor is 240hz so I can't really lock it to 97hz. My undervolt uses 80w less power so it runs pretty cool thanks to that. I get more power than stock too and yes it's stable enough for me.

I have ran several runs of 3D mark port royal. I have also played 3 different games so far for a few hours each and no crashes in any of them.

That's good enough for me. Unfortunately no I don't have cyberpunk to test that out. But I'm pretty sure the many threads I looked at on Reddit some of them must have it to test it out and they all seemed happy with their undervolting too.

Yep, 100Hz (Acer Predator x34)

Port Royal is fine, try Warzone on 3440 x 1440 with settings ramped up or Quake II RTX and play with screenshots. Same with Cyberpunk screenshots, but Quake II RTX is free.

Others may be happy with undervolting but I'm not happy when games crash mid / close to end game when playing with friends. Doesnt matter so much if a single player game or benchmark falls over. 922 hours playing Warzone/Modern Warfare and the only system crashes have been when I undervolted my 3080 FE
 
My 3060ti undervolts like a champ and benchmarks / stresstests fine.. will not run CP2007 though, have to revert to stock !!

Anyone else have this issue ?

I'm guessing you are using both RT and DLSS with CP2077. More of the chip is being used, which makes it a good title to test an undervolt. Just alter your settings by the smallest amount until it stops crashing. I had to come down from 2000 to 1920 for CP2077 to be stable at 900mV.
 
Yep, 100Hz (Acer Predator x34)

Port Royal is fine, try Warzone on 3440 x 1440 with settings ramped up or Quake II RTX and play with screenshots. Same with Cyberpunk screenshots, but Quake II RTX is free.

Others may be happy with undervolting but I'm not happy when games crash mid / close to end game when playing with friends. Doesnt matter so much if a single player game or benchmark falls over. 922 hours playing Warzone/Modern Warfare and the only system crashes have been when I undervolted my 3080 FE

I've got warzone. I'll give it a try tomorrow. Should I just max everything out?

Oh and games can crash sometimes for any reason. I had issues running stock settings before and had a random crash here and there. Alt tabbing out usually is usually the culprit I do that a lot and I run games full screen non windowed which always had issues with alt tabbing.

I don't think I have quake though. I have been running battlefield fine albeit low settings
 
I've noticed something odd, If I undervolt, Lower my clocks, Mem clocks etc... my GPU usage goes up under the same workloads, Guessing this is to compensate for lower clocks.
 
I've got warzone. I'll give it a try tomorrow. Should I just max everything out?

Oh and games can crash sometimes for any reason. I had issues running stock settings before and had a random crash here and there. Alt tabbing out usually is usually the culprit I do that a lot and I run games full screen non windowed which always had issues with alt tabbing.

I don't think I have quake though. I have been running battlefield fine albeit low settings

Here's one I prepared earlier :)
Also, games don't crash for "any reason" they crash for a very specific reason. Be it a software issue like a bug in game or driver incompatibility or hardware issue such as overclock or faulty hardware.
Note the drivers in use, i've stayed on these because upgrading to other caused some strange bugs in Warzone. I am be tempted in future to change to later drivers but if it aint broke...

Alt tabbing works fine for me in all the game I play. I run full screen too. I'm sure Quake II RTX is free on Steam, at least the shareware version will be. If not it'll be around £2

lU5eGRg.png
 
3070FE
1875@850mV mem +1000
105% performance at 90% power compared to stock

But also a super efficient mode
1500@700mV (and memory slider dropped to minimum -500).
When I play games that don't need much power.
And in this mode it is still faster than my old 2070Super overclocked.
 
Last edited:
Just ran Heaven 4 Benchmark. For comparison my stock score was 5327 and the temps came in at 68 degC

I undervolted to 1900 at 887mV on Afterburner and kept it running. Restarted PC and ran Heaven Benchmark again. I got a score of 5019 this time round but despite Afterburner showing 1900-1905MHz, Heaven was showing 2100MHz in top right corner. Temperature undervolted was 75 degC. Still plenty of coil whine.
I am on the latest driver too so maybe that's different. Or maybe I'm just doing something wrong. Will try Neon Noir benchmark now.
 
Back
Top Bottom