• 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.

10GB vram enough for the 3080? Discuss..

Status
Not open for further replies.
There is always going to be some subjectiveness to it - but few are going to buy a 3060/70 class card with the intention of playing the latest and greatest games at 4K with quality setting which compliment that. Sure they can play older games or newer ones at reduced quality setting at 4K but it is always a relative position - it isn't about the exceptions.

Well it kinda is, because if I google some 3070 benchmarks and look at performance in a sample of modern AAA games it performs quite nicely at 4k in almost everything. So you have to go to exceptions to claim otherwise. I just looked at Tomsharware 15 game average for a quick average across a wide array of modern games (in 4k Ultra), and the 3070FE is 61.2FPS average.

I'm not particularly trying to defend the 3070 as "a 4k card", I've never made that claim. But to the degree it's not, you have to pretty much go to exceptions with it, it's not the rule even if you ignore the vast swath of games on the market that aren't demanding AAA games and you narrowly focus apriori on the one that are, you still only see a small fraction of them struggle and even then struggle is by a trivial amount. If you look at steam stats of games played, they list the top 100, I just took those peak numbers put them in a spreadsheet and the top 10 games account for 55% of the peak players. Everything in the top 10 a 3070 can play in 4k easily. And even of the entire top 100 games there's probably only a couple it'd struggle with, Cyberpunk, Maybe Ark because it's so horribly optimized.

It's a problem with interpretation of modern bench marking. Using a suite of all the latest and greatest games is good for stress testing different hardware against each other to see relative speeds, but it's NOT representative of what people play. Most sane people aren't going to spend hard earned money on buying more expensive hardware to eliminate a few exceptions which they might not even play! It's why a standard of a few exceptions prevent a card being "4k" has almost zero real utility. Like when I bought my 3080, I saw there were exceptions in games like MS FS2020, 4k Ultra isn't possible and it was kind of a standalone game in that regard, but I don't even play FS2020 and never will, so what? Nor will most people for that matter. it's a misleading and useless metric, only if you condition it in some sensible way can you really make it useful.
 
Well it kinda is, because if I google some 3070 benchmarks and look at performance in a sample of modern AAA games it performs quite nicely at 4k in almost everything. So you have to go to exceptions to claim otherwise. I just looked at Tomsharware 15 game average for a quick average across a wide array of modern games (in 4k Ultra), and the 3070FE is 61.2FPS average.

60 FPS average probably means you are dropping below and/or in some games below ideal performance. Benchmarks will include a range of games some last gen, etc. where performance at 4K will be much higher. The Division games for instance I'll rarely drop below 100 FPS with the 3070.
 
Well it kinda is, because if I google some 3070 benchmarks and look at performance in a sample of modern AAA games it performs quite nicely at 4k in almost everything. So you have to go to exceptions to claim otherwise. I just looked at Tomsharware 15 game average for a quick average across a wide array of modern games (in 4k Ultra), and the 3070FE is 61.2FPS average.

I'm not particularly trying to defend the 3070 as "a 4k card", I've never made that claim. But to the degree it's not, you have to pretty much go to exceptions with it, it's not the rule even if you ignore the vast swath of games on the market that aren't demanding AAA games and you narrowly focus apriori on the one that are, you still only see a small fraction of them struggle and even then struggle is by a trivial amount. If you look at steam stats of games played, they list the top 100, I just took those peak numbers put them in a spreadsheet and the top 10 games account for 55% of the peak players. Everything in the top 10 a 3070 can play in 4k easily. And even of the entire top 100 games there's probably only a couple it'd struggle with, Cyberpunk, Maybe Ark because it's so horribly optimized.

It's a problem with interpretation of modern bench marking. Using a suite of all the latest and greatest games is good for stress testing different hardware against each other to see relative speeds, but it's NOT representative of what people play. Most sane people aren't going to spend hard earned money on buying more expensive hardware to eliminate a few exceptions which they might not even play! It's why a standard of a few exceptions prevent a card being "4k" has almost zero real utility. Like when I bought my 3080, I saw there were exceptions in games like MS FS2020, 4k Ultra isn't possible and it was kind of a standalone game in that regard, but I don't even play FS2020 and never will, so what? Nor will most people for that matter. it's a misleading and useless metric, only if you condition it in some sensible way can you really make it useful.
He is saying needs to be able to do 120fps to be a 4K card :p
 
He is saying needs to be able to do 120fps to be a 4K card :p

Not 240fps?

This is why this is so stupid, the subjectivity is so extreme in its scope, that you can take exceptions and make claims that no card is capable of anything. It's meaningless. This is why I've repeatedly joked about making a Unity asset flip game on steam that has insane requirements and just putting the whole stupid argument to bed, it's so tempting to actually do that.
 
you dont need 60 fps all the time, vrr is there for a reason

as long as it doesnt drop below 40 fps most TPS games will feel smoth and be playable

fps games are a different story and 3070 can push 100+ frames on most of them already
 
Not 240fps?

This is why this is so stupid, the subjectivity is so extreme in its scope, that you can take exceptions and make claims that no card is capable of anything. It's meaningless. This is why I've repeatedly joked about making a Unity asset flip game on steam that has insane requirements and just putting the whole stupid argument to bed, it's so tempting to actually do that.
Haha.

Yeah. Each to their own, but I know I will be fine with a 3070 at 4K. I have G-Sync also. Plus we are getting more and more DLSS in new games, so should be fine.

Would be awesome if they brought dlss to MS Flight Simulator.
 
The thing is most of the games are using 1spp, 2 if you are lucky. A high sample count that requires very little denoising is in the thousands of spp. :D
Most of the ray tracing is done through the denoising and not the power of RT units. It is an approximation of how the scenes will look like if enough rays are casted.
Here's a interesting thought.
Fast forward a few years/decades/centuries when GPUs can push out 100s of samples per frame.

We will end up in a situation where you walk into a building and your framerate just plummets from 100+ to just below 60 fps. Or your upscaler of choice starts working double duty and decimates IQ as you enter the building.

A tank comes rolling down the street but the metal has a rough finish so your FPS tanks because you made the mistake of looking at it as it rolls past.

The settings page will be interesting for such games. Sorry your GPU isn't powerful enough so everything will have a chrome finish rather than proper roughness, so you can maintain 60fps :p.

And popping smoke will be lethal if they use proper volumetrics. Your FPS drops to dial up speeds. you watch the frame load in one row at a time.
 
you dont need 60 fps all the time, vrr is there for a reason

as long as it doesnt drop below 40 fps most TPS games will feel smoth and be playable

fps games are a different story and 3070 can push 100+ frames on most of them already

Sure - but if over a bunch of games including older and newer you are only getting a 60 FPS average then most likely the older titles, which are going to be performing well over 60 FPS for the most part, are pulling the average up a fair bit and increasingly with contemporary titles you are going to see performance below ideal.
 
CB's Dealthloop review is interesting:
https://www.computerbase.de/2021-09..._10_gb_schnell_pflicht_sorglos_erst_mit_16_gb
(https://translate.google.ca/transla...nell_pflicht_sorglos_erst_mit_16_gb&sandbox=1)

It does looks like the game has VRAM management issues, but still. Even at 1080P with everything including raytracing turned on, they noticed slowdowns.

Frametimes are bit all over the place espeically on the 10GB card:
Jv6UtZ6.png
With Raytracing, the 6800 XT is still slower but at least more consistent.
Strange that their review didn't have a 24GB 3090 included.
 
Oh dear, here we go again :D :D :D

Trying to come to conclusions based on a game a few days after release with known technical problems is hardly representative.

Indeed.

If you read around and watch videos, you'll clearly see the game has issues on all platforms especially PC, no surprise being bethesada though..... For PC, the game has to be capped at 60 fps to resolve the frame pacing issues.

Seems like the 3080 does just fine against the 6800xt here..... although not a great comparison:


Basically best of to wait till it gets fixed before judging performance especially surrounding frame times....

FSR is also appears to be broken? Although could just be FSR not working well given the engine being used. Weird how there is no dlss as nvidias other tech is present, however, it may arrive later.....

https://www.reddit.com/r/Amd/comments/pp8ci6/deathloops_fsr_implementation_is_currently_broken/

With all that said..... Even though ray tracing is very limited in this, it is stunning in this game and really makes a difference.


udOrl8U.jpg

PS 5 based comparison:
bFJq0tj.png

olXyTIG.png

lIKcAiJ.png

inb4 people say it looks better with it off :o
 
Last edited:
Yeah, they still need to click the magical compress all button to compress 16+GB to under 10GB. Oh wait.

Unless they are suffering from some insane memory leak or bug I don't expect much to change.
 
Oh dear, here we go again :D :D :D



Indeed.

If you read around and watch videos, you'll clearly see the game has issues on all platforms especially PC, no surprise being bethesada though..... For PC, the game has to be capped at 60 fps to resolve the frame pacing issues.

Seems like the 3080 does just fine against the 6800xt here..... although not a great comparison:


Basically best of to wait till it gets fixed before judging performance especially surrounding frame times....

FSR is also appears to be broken? Although could just be FSR not working well given the engine being used. Weird how there is no dlss as nvidias other tech is present, however, it may arrive later.....

https://www.reddit.com/r/Amd/comments/pp8ci6/deathloops_fsr_implementation_is_currently_broken/

With all that said..... Even though ray tracing is very limited in this, it is stunning in this game and really makes a difference.


udOrl8U.jpg

PS 5 based comparison:
bFJq0tj.png

olXyTIG.png

lIKcAiJ.png

inb4 people say it looks better with it off :o

RTX 3080: 1440P VRam 7.7GB, Min FPS 43
RX 6800XT: 1440P VRam 9.7GB, Min FPS 87

HHmmmmm......?
 
Status
Not open for further replies.
Back
Top Bottom