• 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.

Rtx 3080 vs Rx 6900 XT

Price being equal, I'd go with a 6900 XT for your needs. You can use the added rendering power for VR, and you won't see the benefits of Raytracing in those titles. But prices are not equal... if you can find the 3080 for ~£200 less, go for that.

Regarding the PSU, I'd move to a higher capacity CPU >=750W. You could run a 3600 and 3080 or 6900XT on a quality 650W PSU, but you'll likely not get much/if any OC headroom, and you'll always be a bit worried. Always cracks me up when people don't want to spend money on the cheapest component that could basically make or break (literally) your rig.
 
You could run a 3600 and 3080 or 6900XT on a quality 650W PSU, but you'll likely not get much/if any OC headroom, and you'll always be a bit worried. Always cracks me up when people don't want to spend money on the cheapest component that could basically make or break (literally) your rig.

Yep, don't cheap out on the PSU lol

 
'our friend here', don't be a patronising ******** :rolleyes:
I thought we were friends?
Yep, don't cheap out on the PSU lol
I'd rather have the lower 50-75W average under full load to be honest as that means you can run a lower fan speed, with less noise and heat that the cooler has to dissipate.

You don't really hear any 6800 XT and 6900 XT owners complaining about heat, fan noise or memory chips running at or near throttling temperatures on a daily basis without modification or water cooling.

Ooof, never seen this before.. couple that with an 11900k and your looking at a 1200w PSU to cover potential spikes !!
It's largely meaningless tbh. Any good quality 850W unit can handle a 6900 XT with ease and as has been noted in this thread, we have people using 650W units with a 6900 XT. Not that I would recommend lower than 850W personally.
 
Why would you recommend a GPU that can spike to over 600W :eek::cry:
Well, there goes also the 3080.

Igor'sLab measured one of the non-reference 3080 cards drawing 600W level sub-ms transients continuously in gaming with first driver.
https://www.igorslab.de/en/kfa2-geforce-rtx-3080-sg-10-gb-in-test-not-so-quiet-but-true-cool/4/
And even reference card draws 500W transients during gaming.
So pretty sure non-reference factory OC models routinely go over 500W during gaming.
 
Ooof, never seen this before.. couple that with an 11900k and your looking at a 1200w PSU to cover potential spikes !!
Reference 6900 XT's gaming transients are little above 450W:
https://www.igorslab.de/en/grasps-a...with-benchmarks-and-a-technology-analysis/15/
And that includes also finer analysis at 100 samples per ms accuracy showing power draw during individual frames!

Even longer sequence analysis has quarter higher sampling rate than used by TPU.
Who make big number of that single FurMark initiliazing surge while not checking gaming transients.
Then again TPU's 40 samples per second accuracy wouldn't be enough for per frame transients.
Even at Hollywood's stone age fps...
 
Surely you would just pay a little more for a 3090, more vram and all the nvidia features

Thing is it's not only a little more currently, the Liquid cooled Sapphire 6900XTX with the 360mm rad is £1800-£1900, one of the most expensive models, where 3090's start from around £2500.
 
Thing is it's not only a little more currently, the Liquid cooled Sapphire 6900XTX with the 360mm rad is £1800-£1900, one of the most expensive models, where 3090's start from around £2500.

Ah yeah I suppose. If NZXT release a kraken g13 that would help costs, but there doesn't seem to be a product like that for rtx 3090
 
Well, there goes also the 3080.

Igor'sLab measured one of the non-reference 3080 cards drawing 600W level sub-ms transients continuously in gaming with first driver.
https://www.igorslab.de/en/kfa2-geforce-rtx-3080-sg-10-gb-in-test-not-so-quiet-but-true-cool/4/
And even reference card draws 500W transients during gaming.
So pretty sure non-reference factory OC models routinely go over 500W during gaming.

Probably ironed out since then in various driver and BIOS updates as the TPU inclusion of the spike data is recent.
 
Probably ironed out since then in various driver and BIOS updates as the TPU inclusion of the spike data is recent.
It's AMD who stands chance for eliminating that spike by making driver impose automatic hard cap for frequency/volts the moment it detects Furmark.
That spike looks exactly like waiting for temperature to rise before forcing limits.
Nvidia has done such automatic limiting in Furmark after previous records breaking power draw and serious inefficiency of Fermi.

And TPU doesn't measure per frame transients during gaming.
Their low 40 samples per second accuracy couldn't even tell anything about those... Unless playing at few fps performance.
 
Reference 6900 XT's gaming transients are little above 450W:
https://www.igorslab.de/en/grasps-a...with-benchmarks-and-a-technology-analysis/15/
And that includes also finer analysis at 100 samples per ms accuracy showing power draw during individual frames!

Even longer sequence analysis has quarter higher sampling rate than used by TPU.
Who make big number of that single FurMark initiliazing surge while not checking gaming transients.
Then again TPU's 40 samples per second accuracy wouldn't be enough for per frame transients.
Even at Hollywood's stone age fps...
I didn't realise TPU is measuring with Furmark, good info. That's a worse case scenario then.
 
Looks like folks here are leaning slightly towards the 6900XT.
But now I have recently had a look at the 6800XT, I'm considering saving a couple of hundred pounds by going for the 6800XT.

I'm just popping over to youtube to compare 6900xt vs 6800xt.
Oh, the quandry. But what a nice quandry: )

Sorry, I know I am late, but most of the advice that you have received is from People who obviously don't play VR games and are basing their advice purely on Pancake games.

Somebody should have asked what VR headset you have first before handing out advice on which card to buy. Now I think I remember you from the VR forum, You either have a Rift S and thinking about buying a Quest 2 or had a Quest 2 and went back to a Rift S? I can't remember which way around.

Which way was it? Or are you thinking of buying something else? Or am I confusing you with someone else?

As for VR racing sims, most work better on Nvidia cards, so it's also important to know which sims you play most often?
 
Sorry, I know I am late, but most of the advice that you have received is from People who obviously don't play VR games and are basing their advice purely on Pancake games.

Somebody should have asked what VR headset you have first before handing out advice on which card to buy. Now I think I remember you from the VR forum, You either have a Rift S and thinking about buying a Quest 2 or had a Quest 2 and went back to a Rift S? I can't remember which way around.

Which way was it? Or are you thinking of buying something else? Or am I confusing you with someone else?

As for VR racing sims, most work better on Nvidia cards, so it's also important to know which sims you play most often?

You're correct. It's probably me you have in mind.
I tested out oculus Go, Rift S and Quest 2.
Finally settled on a Quest 2,

At the moment I'm currently playing Automobilista 2. That could well alter depending on the weather ;). Project cars, assetto corsa too. It's not a good strategy to get much better at these games. Hey, I have fun and frustration with all of them. :)

I've already committed to an rx 6800 xt. It'll be good enough for me, I'm sure. I'm a casual gamer, not an avid gamer.

I posted earlier in this thread the reasons for my decision.

Eliminated 3090 on price.
Eliminated 3080 on price and availability
Eliminated 6900 on price v performance value.

So that more or less led me to an rx 6800 xt.
 
You're correct. It's probably me you have in mind.
I tested out oculus Go, Rift S and Quest 2.
Finally settled on a Quest 2,

At the moment I'm currently playing Automobilista 2. That could well alter depending on the weather ;). Project cars, assetto corsa too. It's not a good strategy to get much better at these games. Hey, I have fun and frustration with all of them. :)

I've already committed to an rx 6800 xt. It'll be good enough for me, I'm sure. I'm a casual gamer, not an avid gamer.

I posted earlier in this thread the reasons for my decision.

Eliminated 3090 on price.
Eliminated 3080 on price and availability
Eliminated 6900 on price v performance value.

So that more or less led me to an rx 6800 xt.

Actually in this case VR doesnt really make a difference to what card you have as much as PC dedicated VR. With the Quest 2 you also have to mention how you are streaming the game to the Quest 2. Using Virtual Desktop and Oculus own (nicked) version is a bit like Google Stadia or Geforce now. Your game is "streamed" to the Quest and for decent play is as dependant on a good network connection especially if you play via WIFI. So you might have the best card in the world but if your wifi is poor then so is the game experience. If you use a USB cable then its better (but limiting your movement etc) but you can still tinker with the settings to get a very decent experience even on something like a GTX 1070. Just speaking about the Quest 1 from experience but the Quest 2 has 90hz displays still and you dont even need a current gen card to get the most out of it. The new software update gives 120hz but I dont know whether the software inc Virtual Desktop allows for this yet.

What I am trying to say is you settled on the Quest 2 but you need to tell us how you are playing the games. Using Supersampling ? Wifi ? USB Cable ? people have reported your game plays great on a 1080ti for instance which is 2 gens old. Its not quite the same as say getting a Rift S.
 
I didn't realise TPU is measuring with Furmark, good info. That's a worse case scenario then.
And it's only some kind initial start up surge before rising temperature makes driver limits clocks/volts down.
In TPU's measurements gaming spikes are actually at 350W.
https://www.techpowerup.com/review/gigabyte-radeon-rx-6900-xt-gaming-oc/35.html

But that 40 samples per second just isn't fast enough to catch transients inside rendering of frame.
Igor'sLab again takes that fine analysis at 100 000 samples per second rate.
Which actually shows that steady state draw in torture test (Furmark/MSI Combustor...) has lower transients than gaming!
 
Back
Top Bottom