• 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.

4k 120hz CPU impact?

Tweaktown have am article on AC Valhalla, Just barely 4k 60fps on a 3090 and latest Intel cpu. Some games you must just wait two years games are mastered around 60fps thats how it is. Its no big deal to me two years by then every dlc is out and the price halves.

Win win :)
 
No, a 6900XT trades blows with a 3090 with rage mode on (small overclock) and SAM, this also hasn’t been independently verified and is only the manufactures data. Trades blows is not better, though the 6900 is cheaper of course.

The 3090 also doesn’t need to be paired with the latest processor to get that performance and at 120fps you can get away with a much older and cheaper CPU to get the same performance.

Once you start overclocking the 3090 to match the rage mode as well it will start to pull ahead again.

If a 3090 can’t do 4K 120 in the latest games at max settings, neither can a 3900XT without sacrificing quality.

If you drop down the image quality you can get 4K 120. That’s what the consoles do.

And before you say doom eternal, yes you could do 4K 120 in that game but it’s just one extremely well optimised game, the vast majority do not run like that.

Don’t get me wrong the 6900/6800 range look like really powerful cards and seriously good value but let’s step off the hype train, take a breath and wait for the claims to be independently verified.

The 6900XT isn’t available until December so I can’t see how anyone has verified anything yet. AMD say they are seeing 4-13% extra performance on an all AMD system without optimisation or the AGESA update. We’ll have to wait and see, but it’s looking as if AMD will be faster.
 
No, a 6900XT trades blows with a 3090 with rage mode on (small overclock) and SAM, this also hasn’t been independently verified and is only the manufactures data. Trades blows is not better, though the 6900 is cheaper of course.

The 3090 also doesn’t need to be paired with the latest processor to get that performance and at 120fps you can get away with a much older and cheaper CPU to get the same performance.

Once you start overclocking the 3090 to match the rage mode as well it will start to pull ahead again.

If a 3090 can’t do 4K 120 in the latest games at max settings, neither can a 3900XT without sacrificing quality.

If you drop down the image quality you can get 4K 120. That’s what the consoles do.

And before you say doom eternal, yes you could do 4K 120 in that game but it’s just one extremely well optimised game, the vast majority do not run like that.

Don’t get me wrong the 6900/6800 range look like really powerful cards and seriously good value but let’s step off the hype train, take a breath and wait for the claims to be independently verified.


If you are turning all settings to max @ 4k then you are doing it wrong. Many games settings are there to make potato resolutions (1080p) to look better so @4k you turn them off/low, freeing up valuable processing power. The only reason you see benchmarks from reviewers with 4k and low fps figures is becuase it's an apples to apples comparison of processing power, they have to turn on all the potato settings too, which @ 4k requires takes up a significant amount of available performance.
 
I'd image that 1080p 120hz, 1440p 120hz, and 4k 120hz all have pretty much the same CPU impact. So just look up 1080p since there's tons of data regarding CPU impact.

I agree with Woodsta888, settings like FSAA for example are for removing jagged edges from lower resolutions. It's a waste of time setting these in 4k.
 
Jaggies are still an issue at 4K with more or less of an impact depending on screensize/viewing distance. Distant shimmering can be annoying too.
 
We have a lot of resolutions between 2560x1440 and 4K.

3.7 to 8.4 megapixels is a huge jump. We have all the ultra wide resolutions between.
 
Last edited:
The 6900XT isn’t available until December so I can’t see how anyone has verified anything yet. AMD say they are seeing 4-13% extra performance on an all AMD system without optimisation or the AGESA update. We’ll have to wait and see, but it’s looking as if AMD will be faster.

That’s true, but if you look at the AMD slides they already have both of those modes enables to get the two cards to trade blows, a 6900X isn’t faster than a 3090, it’s significantly better value though and doesn’t really change that no one should be buying either over the 3080/6800XT unless you absolutely must have the best as both are poor value compared to the next two cards down the stack.
 
That’s true, but if you look at the AMD slides they already have both of those modes enables to get the two cards to trade blows, a 6900X isn’t faster than a 3090, it’s significantly better value though and doesn’t really change that no one should be buying either over the 3080/6800XT unless you absolutely must have the best as both are poor value compared to the next two cards down the stack.

But the 6900XT seems faster. Even without much (any?) optimisation.
 
No it doesn’t, get off the hype train.

According to AMD they trade blows in a carefully selected list of games chosen by AMD with all the optimisations turned on. They also put out zero ray tracing numbers, it might be good at RT, likewise they might be terrible.

I don’t see how you can say a 6900XT seems ‘faster’ than a 3090. It’s certainly cheaper and a better value but I just can’t see how you can say it’s faster unless your reading different slides to everyone else. Either way they are both bad value compared to the 6800XT/3080.

I think the argument your trying to say but getting your facts wrong is the 6800XT (according to AMD) is faster than a 3080 with all the optimisations turned on which very well could be the case. They showed it trading blows in the same carefully selected games with the optimisations turned off. Therefore it’s fair to assume that with them switched on it may nudge ahead.
 
No it doesn’t, get off the hype train.

According to AMD they trade blows in a carefully selected list of games chosen by AMD with all the optimisations turned on. They also put out zero ray tracing numbers, it might be good at RT, likewise they might be terrible.

I don’t see how you can say a 6900XT seems ‘faster’ than a 3090. It’s certainly cheaper and a better value but I just can’t see how you can say it’s faster unless your reading different slides to everyone else. Either way they are both bad value compared to the 6800XT/3080.

I think the argument your trying to say but getting your facts wrong is the 6800XT (according to AMD) is faster than a 3080 with all the optimisations turned on which very well could be the case. They showed it trading blows in the same carefully selected games with the optimisations turned off. Therefore it’s fair to assume that with them switched on it may nudge ahead.

I’m saying it because it looks as if the 6900XT will be faster than the RTX 3090. For that not to be the case the 6800XT would have to be only a few percent slower, yet we know the 6900XT has a much better performance per watt improvement.
 
Performance per watt improvement doesn't necessarily translate to frame rate directly - things like transistor gating, etc. can save a considerable amount of power but not necessarily mean you can turn the frequency up significantly.
 
Performance per watt improvement doesn't necessarily translate to frame rate directly - things like transistor gating, etc. can save a considerable amount of power but not necessarily mean you can turn the frequency up significantly.

Of course, but what are the chances the 6800XT and 6900XT will both fall between the 3080 and 3090 when that margin is barely double digit percentages.

It looks like

3070
6800
3080
6800XT
3090
6900XT
Ampere Titan - 6900XTX
 
Personally waiting for the reviews - the 3090 shows performance at the top doesn't necessarily pan out as expected (and that can go either way).
 
Personally waiting for the reviews - the 3090 shows performance at the top doesn't necessarily pan out as expected (and that can go either way).

Well it looks as if Nvidia has rushed Ampere to market and AMD have turned its attention to Radeon. It’s staggering that in 18 months AMD have caught Nvidia after having pretty zero interest in the highend desktop graphics market.
 
Of course, but what are the chances the 6800XT and 6900XT will both fall between the 3080 and 3090 when that margin is barely double digit percentages.

It looks like

3070
6800
3080
6800XT
3090
6900XT
Ampere Titan - 6900XTX

Going off of AMD's pricing and guessing that the pricing reflects some level of confidence;

3070
6800
6800XT
3080
6900XT
3090w

I still think the 6800XT will be close enough to the 3080 that I will be happy with whichever I can actually buy.
 
Not sure why @jigger seems to keep peddling that a 6900xt is faster than a 3090 when AMDs own slides show it at best equaling it when overclocked and with SAM enabled. I get it’s significantly cheaper and a far better value proposition but that doesn’t make it faster.
 
Certainly going to make for some interesting reviews - especially if performance at the top end depends heavily on having an AMD ecosystem in play.
 
Not sure why @jigger seems to keep peddling that a 6900xt is faster than a 3090 when AMDs own slides show it at best equaling it when overclocked and with SAM enabled. I get it’s significantly cheaper and a far better value proposition but that doesn’t make it faster.
The 6900 XT was not overclocked, Rage mode slightly increases the power slider which allows for a higher average clock speed, it does not increase the slider to its maximum value.
 
Not sure why @jigger seems to keep peddling that a 6900xt is faster than a 3090 when AMDs own slides show it at best equaling it when overclocked and with SAM enabled. I get it’s significantly cheaper and a far better value proposition but that doesn’t make it faster.

Ive explained why several times now.
 
Back
Top Bottom