The Ampere RTX 3080 Owners Thread

[Sarge2255]

Yes each card has different quality silicon, so this may not be a good profitable experiment sadly.

My profile 1 I didnt change anything, I just saved it, so I had an easy way to get back to my card's stock. So either your card has a different voltage curve to mine at stock or maybe you previously had a higher power limit or clock boost dialled in? as it would reset those also.

I would not want to overclock my card , that's why I got the overclocked version. It runs at 73c max with 70% ish fan speed so I'll just stick with this. And with 40% fan speed 76c. It's ultra silent.

But the card doesn't like to be given low voltage and and lower clocks speed and perform better than normal. But it makes sense to me.

Oh well, was looking forward to an undervolt with some boost

 

[heatonpkmassive]

I would not want to overclock my card , that's why I got the overclocked version.

I credited you with more smartness than that Sarge, the world and his dog knows that paying extra for a factory OC is money down the pan. You literally up the power limit to the max on the non-OC models and see what clocks you get, less than 10 minutes work and often, you achieve a better OC than the factory OC.

But, it's your money..

 

[chrcoluk]

scores for stock timespy.

profile 1 15542 graphics 16447 cpu 11850
profile 3 15373 graphics 16185 cpu 11973
profile 4 15751 graphics 16686 cpu 11955

attaching saved scores for people to examine in 3dmark.

https://drive.google.com/file/d/1Y6kkgA_j5rgajM50Mi4c2y03qjMjW4uD/view?usp=sharing

profile 1 FE stock
profile 3 target clock 1830 voltage 0.850, power limit 80% but doesnt go above about 70% usage
profile 4 target clock 1905 voltage 0.882, power limit 100% but doesnt go above about 90% usage.

 

[Sarge2255]

I credited you with more smartness than that Sarge, the world and his dog knows that paying extra for a factory OC is money down the pan. You literally up the power limit to the max on the non-OC models and see what clocks you get, less than 10 minutes work and often, you achieve a better OC than the factory OC.

But, it's your money..

Ha ha, I'm a good overclocker too. But tbh the card looked like it would run cool with the beefy cooler and with all the limited choice I planned not to overclock it and brick it with all this shortage of cards/ void any warranty etc at any point. So glad it comes OC out of the box.

You would have to look at benchmarks to justify the OC out of the box to the extra £80 or so you would pay. I found it to be like 3 - 5 fps here and there but, is it all a gimmick? I'll let you decide. Personally, I would prefer just to have a card in this climate.

scores for stock timespy.

profile 1 15542 graphics 16447 cpu 11850
profile 3 15373 graphics 16185 cpu 11973
profile 4 15751 graphics 16686 cpu 11955

attaching saved scores for people to examine in 3dmark.

https://drive.google.com/file/d/1Y6kkgA_j5rgajM50Mi4c2y03qjMjW4uD/view?usp=sharing

profile 1 FE stock
profile 3 target clock 1830 voltage 0.850, power limit 80% but doesnt go above about 70% usage
profile 4 target clock 1905 voltage 0.882, power limit 100% but doesnt go above about 90% usage.

I did additional tests.
I found out that I got identical performance at 850mV with 1815Mhz core speed as to with 850mV with 1615Mhz core speed (-200Mhz core speed). This could explain why people are getting the same/similar performance after reducing core speed and undervolting the gpu.

The issue is from a stock overclocked card reducing the core would result in a small performance loss, all though be it 1-3 fps but significant power reduction like 100w . I couldn't run the card undervolted at the stock overclocked speeds either, the card would not agree to it.

 

[heatonpkmassive]

Ha ha, I'm a good overclocker too. But tbh the card looked like it would run cool with the beefy cooler and with all the limited choice I planned not to overclock it and brick it with all this shortage of cards/ void any warranty etc at any point. So glad it comes OC out of the box.

You would have to look at benchmarks to justify the OC out of the box to the extra £80 or so you would pay. I found it to be like 3 - 5 fps here and there but, is it all a gimmick? I'll let you decide. Personally, I would prefer just to have a card in this climate.

Well as I say, it's your money to do with as you please Sarge. I also pointed that out in case anyone else reading this wasn't aware.

Plus, you won't brick anything with Afterburner, that's why it has that 12% or whatever it is power limit. So if you have a good cooler, I'd be pushing for whatever that 12% power limit will allow.

 

[chrcoluk]

I see now why me and doug are seen as having low 3dmark scores, reviewers reported 17k graphics score on these cards (not overclocked) and average on 3dmark is about 17.6k. I have no idea whats going on there, whether I got a big loser on silicon lottery or if other parts of the system affect graphics scores in 3dmark, I suspect its the latter. The average on 3dmark is also probably skewed by most on there overclocking, but still mine seems about 500 points down.

 

[Sarge2255]

I see now why me and doug are seen as having low 3dmark scores, reviewers reported 17k graphics score on these cards (not overclocked) and average on 3dmark is about 17.6k. I have no idea whats going on there, whether I got a big loser on silicon lottery or if other parts of the system affect graphics scores in 3dmark, I suspect its the latter. The average on 3dmark is also probably skewed by most on there overclocking, but still mine seems about 500 points down.

I'd say the cpu speed and cores affect the score when benchmarking with a 3080.

 

[Joxeon]

I'd say the cpu speed and cores affect the score when benchmarking with a 3080.

A 9900k is still a very good CPU though so I doubt that's holding it back, I was getting a higher graphics score with a ryzen 3600 which has less and slower cores.

 

[chrcoluk]

A 9900k is still a very good CPU though so I doubt that's holding it back, I was getting a higher graphics score with a ryzen 3600 which has less and slower cores.

What score did you get? assuming 3080 FE at stock.

 

[AyeAyeAhoy!]

I see now why me and doug are seen as having low 3dmark scores, reviewers reported 17k graphics score on these cards (not overclocked) and average on 3dmark is about 17.6k. I have no idea whats going on there, whether I got a big loser on silicon lottery or if other parts of the system affect graphics scores in 3dmark, I suspect its the latter. The average on 3dmark is also probably skewed by most on there overclocking, but still mine seems about 500 points down.

I'll fire it up again tonight, but I'm sure when mine run it said I was over the average for a 3080. I can't remember exactly what settings I used either - will look properly later

 

[Sarge2255]

A 9900k is still a very good CPU though so I doubt that's holding it back, I was getting a higher graphics score with a ryzen 3600 which has less and slower cores.

I agree, Core i9 9900k is a good cpu for gaming! I get 17,800/18,000 on the graphics for Timespy and I'm using core i7 4.5Ghz cpu

Overclock your 9900k, that makes a difference.

 

[Sarge2255]

I'll fire it up again tonight, but I'm sure when mine run it said I was over the average for a 3080. I can't remember exactly what settings I used either - will look properly later

run it at default settings.

 

[heatonpkmassive]

Overclock your 9900k, that makes a difference.

That's more like it Sarge

 

[Joxeon]

What score did you get? assuming 3080 FE at stock.

I was getting 17300 at stock with the ventus which should be similar to the FE, CPU score with a 3600 was around 7800, I'm now running a 5800X which scores 12700 but graphics score is about the same so I don't think the CPU effects it that much so long as the CPU is reasonably modern.

 

[ltron]

I see now why me and doug are seen as having low 3dmark scores, reviewers reported 17k graphics score on these cards (not overclocked) and average on 3dmark is about 17.6k. I have no idea whats going on there, whether I got a big loser on silicon lottery or if other parts of the system affect graphics scores in 3dmark, I suspect its the latter. The average on 3dmark is also probably skewed by most on there overclocking, but still mine seems about 500 points down.

Have you got virtualisation enabled in the BIOS? If Windows has automatically enabled Memory Integrity in the Core Isolation section of Windows Security (which can happen with virtualisation on) or you've installed Hyper-V this can significantly reduce gaming performance. A quick way to test is to disable virtualisation in the BIOS.

I get about 17200 for my Timespy graphics score with a stock 3080 FE and Ryzen 3900X.

 

[chrcoluk]

Have you got virtualisation enabled in the BIOS? If Windows has automatically enabled Memory Integrity in the Core Isolation section of Windows Security (which can happen with virtualisation on) or you've installed Hyper-V this can significantly reduce gaming performance. A quick way to test is to disable virtualisation in the BIOS.

I get about 17200 for my Timespy graphics score with a stock 3080 FE and Ryzen 3900X.

Its on, when I flash the bios for rebar support, I will disable virtualization then and rebench.

Also i might put this in my ryzen machine temporarily and try a bench, wondering if its a ryzen advantage. But will need to be careful as that has a weaker PSU.

This is worth a read, a couple also getting scores similar to me. https://www.reddit.com/r/nvidia/comments/j58bm7/3080_time_spy_gpu_score_at_stock/

 

[ltron]

Its on, when I flash the bios for rebar support, I will disable virtualization then and rebench.

Also i might put this in my ryzen machine temporarily and try a bench, wondering if its a ryzen advantage. But will need to be careful as that has a weaker PSU.

This is worth a read, a couple also getting scores similar to me. https://www.reddit.com/r/nvidia/comments/j58bm7/3080_time_spy_gpu_score_at_stock/

I don't think it's a Ryzen advantage, I'd be surprised, but if disabling virtualisation doesn't improve things then it is worth trying the card in the Ryzen machine.

 

[chrcoluk]

I don't think it's a Ryzen advantage, I'd be surprised, but if disabling virtualisation doesn't improve things then it is worth trying the card in the Ryzen machine.

I did already check that memory integrity setting and its off. But I will still disable in the bios as well.

With the clocks people are reporting, I dont think its clock speed related. So I am super curious now what happens in my ryzen rig.

 

[chrcoluk]

I don't think it's a Ryzen advantage, I'd be surprised, but if disabling virtualisation doesn't improve things then it is worth trying the card in the Ryzen machine.

Is fixed

I went in the bios disabled virtualization, it wasnt that, but whilst I was in there i checked everything over and fclk was on 800mhz, boosted it to 1ghz and that fixed it.

Now how do these scores look? will also attach the saved 3dmark files.

Also this time I let gpu cool down before running profile 3 and 4 this time so they are a bit more favourable vs last time compared to stock.

profile 1 score 16299 gpu 17402 cpu 11993
profile 3 score 16245 gpu 17270 cpu 12160
profile 4 score 16665 gpu 17842 cpu 12133

https://drive.google.com/file/d/1zlNUzCdzuAm6xk9VT7oSTKt6-to5aX_O/view?usp=sharing

 

[ltron]

I did already check that memory integrity setting and its off. But I will still disable in the bios as well.

With the clocks people are reporting, I dont think its clock speed related. So I am super curious now what happens in my ryzen rig.

Yes, disable virtualization in BIOS and check to be sure.