AMD RDNA3 unveiling event

[arknor]

if you select the highest preset you are not truly running max.

it doesn't even say what settings AMD used and their cyberpunk settings aren't the max either.
Dying light 2 high RT? why the game defaults to ultra rt on my pc so amd manually turned the settings down after turning on raytracing it seems


I like how they have the Nvidia ard to compare ports and price but not fps... pretty telling imo. also turning down settings in 2year old games lol

 

[LtMatt]

it doesn't even say what settings AMD used and their cyberpunk settings aren't the max either.
Dying light 2 high RT? why the game defaults to ultra rt on my pc so amd manually turned the settings down after turning on raytracing it seems


I like how they have the Nvidia ard to compare ports and price but not fps... pretty telling imo. also turning down settings in 2year old games lol

I'm merely commenting on the Nvidia result and posting the true score at 4k max settings, not using a resolution slider to guess 4k performance as it's not accurate.

 

[uscool]

it doesn't even say what settings AMD used and their cyberpunk settings aren't the max either.
Dying light 2 high RT? why the game defaults to ultra rt on my pc so amd manually turned the settings down after turning on raytracing it seems


I like how they have the Nvidia ard to compare ports and price but not fps... pretty telling imo. also turning down settings in 2year old games lol

Because the 4080 isn't released?

 

[arknor]

Because the 4080 isn't released?

then compare to a 4090 and let people guestimate

 

[uscool]

then compare to a 4090 and let people guestimate

Because it will probably lose in every chart that's what people look at regardless of the price difference

 

[gpuerrilla]

 

[RavenXXX2]

4080 is going to get demolished by both the xtx and xt if it's only 19% faster over a 3090ti.

NVIDIA GeForce RTX 4080 is 19% faster than RTX 3090 Ti in the first gaming review - VideoCardz.com

NVIDIA RTX 4080 tested Big Hardware Player (硬件大玩家) delivers another pre-launch review of GeForce graphics card. This content creator and reviewer was the first to share RTX 3080 Ti and 3070 Ti benchmarks, so this is not the first time we are getting early results from him. This time around he is...

videocardz.com

 

[Joxeon]

1) We don't know what the performance will be exactly.
2) The 6900XT won't be in stock forever, especially at that price
3) the 7800XT hasn't even been announced yet so we have no idea how long it'll take to arrive.

The leaks say the 7800XT will have 60 CUs on a navi 32 die which is 12 less than the 6800XT and 24 less than the 7900XT.

From these new slides the 84 cu 7900XT looks around 30% faster than the 6950XT which had 80cu's

If the 7800XT does indeed have 60 CUs the same as the 6800 non XT then we would expect the same +25-30% uplift which would put it around 10-15% faster than the 6800XT which is exactly where the 4070ti is going to sit.

If the leaks are wrong which is unlikely and AMD use the navi31 die we would likely see 72CUs and the card would then be around 25-30% faster than the 6800XT.

 

[Besty]

worth re-stating that the 6900 series is too hot and power hungry to fit inside a laptop, the mobile 7800XT part will be a game changer.

therefore, AMD can afford to be a bit cheeky with the 7800XT on the basis that it will play 99% of all <2021 games at over 90fps on a laptop.

for some people the 7800XT is a grail GPU because its fast enough, wont melt your laptop and wont cost over a grand.

 

[gpuerrilla]

If the 7800XT does indeed have 60 CUs the same as the 6800 non XT then we would expect the same +25-30% uplift which would put it around 10-15% faster than the 6800XT which is exactly where the 4070ti is going to sit.

and make it tie with the 6900xt?..

 

[Besty]

re: 4090 vs 7900XTX, it looks like the 7900XTX is about 12% slower than a 4090.

 

[arknor]

re: 4090 vs 7900XTX, it looks like the 7900XTX is about 12% slower than a 4090.

if you turn down settings in games like AMD

 

[LtMatt]

It's annoying mate as the game often resets the graphics settings to a lower value for no apparent reason. It would be so much easier to apply the Extreme pre-set (the highest) and have it truly apply the maximum settings. Instead, if leaves a couple of settings below max and you have to manually change them every time the game decides to reset settings. This can make comparisons tricky, unless the pre-set is used. From the MW2 benchmark results screen though, you don't know what settngs were used.

 

[Kirby Wurm]

It's fair to say that some of the extreme settings in many games are hard to distinguish from the high settings, for a very disproportionate hit to FPS. But still, for apples to apples comparisons getting them consistent is obviously a must.

 

[Joxeon]

and make it tie with the 6900xt?..

The 7800XT will be around 5% faster than a 6900XT and the 4070ti if it's a 60 CU card.

The 7900XT looks like it's the same speed as a 4080.

 

[EastCoastHandle]

I would not pay too much attention to that Warzone benchmark he is using render resolution so it's not running at true 4k native.

Here's my 4090-running stock, 4K max settings in Modern Warfare 2. This is true max settings btw, if you select the highest preset you are not truly running max.

Spoiler

Thanks for that. What was it about that review that told you it wasn't native res?

 

[LtMatt]

Thanks for that. What was it about that review that told you it wasn't native res?

I'm talking about the screenshot that says DLSS off, that's the one to ignore. Not running true 4K.

 

[EastCoastHandle]

I'm talking about the screenshot that says DLSS off, that's the one to ignore. Not running true 4K.

Gotcha. That 4k result is not running true 4k because it's using render resolution. Can you elaborate on what this in the benchmark results?

 

[arknor]

It's annoying mate as the game often resets the graphics settings to a lower value for no apparent reason. It would be so much easier to apply the Extreme pre-set (the highest) and have it truly apply the maximum settings. Instead, if leaves a couple of settings below max and you have to manually change them every time the game decides to reset settings. This can make comparisons tricky, unless the pre-set is used. From the MW2 benchmark results screen though, you don't know what settngs were used.

apparently, cyberpunk has/had a bug where the settings wouldn't actually change, I experienced it when I upgraded to a 4090, even launching the game again they weren't applied I had to use geforce experience crap which must overwrite the games settings file

Reddit - Dive into anything

www.reddit.com

brainfreeze77
·2 mo. ago

There seems to be a bug in the graphics settings. If you make a change and then run the benchmark the new setting does not get applied. Click Video and then make a change to Vsync, click the apply button that pops up and then change your vsync option back so apply pops up again and click it again. Then run the benchmark.

so that makes comparison's even more fun

I'm talking about the screenshot that says DLSS off, that's the one to ignore. Not running true 4K.

wouldn't that be the opposite if DLSS was on it's not actual 4k but upscaled

 

[gpuerrilla]

The 7800XT will be around 5% faster than a 6900XT and the 4070ti if it's a 60 CU card.

Yeah I see that, but considering (if I was needing/wanting a GPU right now) then a 6900XT is quite reasonable at £650 if you cant wait a few months. Pretty much on par with a soon to be 7800XT.