Is it possible that the RX 6800XT is faster than the 3090? - Page 7 - Video Cards - HWzone Forums
Skip to content
  • Create an account
  • About Us

    Hello Guest!

    Please note - in order to participate in our community, comment and open new discussions, you must join as a registered member.

    Our members enjoy many advantages, including the ability to participate in discussions, enjoy raffles and promotions for members of the site, and receive our weekly content directly by email.

    Do not like being harassed by email? You can register for the site but do not submit your registration to the weekly email updates.

Is it possible that the RX 6800XT is faster than the 3090?


Recommended Posts


Meanwhile lanzar reports that it has read the technical review which details that the new version of Aegesa from today improves the efficiency of

Risen 5000 in front of the platform. Therefore, performance improvements are expected across the board. So this is the background that explains the issue for us.

By the way this means that all of the Risen 5000 reviews on the internet should actually undergo a re-execution. Let's see

With overseas reviewers enough professionals are worthy and wise to do so. And if they do not, then not bad, we will bring to the local community

Our knowledge in love and out of mission.


Now back to our topic, we had a debt to you guys at Tombraider in So we ran it today too (all thanks to the caveman)

Who remarked to us at all that that should be checked as well . So again a lot of appreciation for his wise advice because without it we would probably be too lazy to carry the TV into the room).


Of course we ran again at 1440P only to find a significant improvement thanks to the new aegesa version. Today is a day of news for friends : )


So here are fresh 1440P results from today, we went up from 166FPS yesterday to 176FPS today after the Aegesa update 



Here are the results on That we had to = 101 FPS



And the graph for reference:

At 1440P we went up from 148FPS to 176 (19% improvement)

B- We increased from 89FPS to 101 (13.5% improvement)



Edited By nec_000
Link to content
Share on other sites

I agree with you that if the price is the same, the performance is the same, the post-performance is the same, buy Smarter. Drivers mature in day one, most of the architecture gets most of the performance already in the first drivers. Features that get in nvidia that sometimes arrive late at , And despite all this. 80-90 percent of gamers prefer net performance over performance under ray tracing.


So here the choice is your net.




Another thing, do not forget that the vast majority of titles do not support ray tracing. And worse Misa, not every game that supports ray tracing supports dlss. In such cases the ray tracing is unnecessary because the frame rate you will get would be unreasonable. This is where the dlss was born. Compensate for the resources that ray tracing requires. And guess what? The vast majority of gamers run dlss on their own. Without enabling ray tracing. To get a great gaming experience at high resolutions like .


Edited By Lanzar
Link to content
Share on other sites

Another interesting thing, we realized that RDNA2 cards (unlike Nvidia cards in width - not just empire) and there is material on the net

Also, gain (slightly) performance when working on the 4th generation pci-e bus. Something like 1-3% or so. We have seen this replicate on RDNA1

Using the rx5700 that lanzar has. He too enjoys 4th generation pci-e.


Therefore when performing the measurements on a Risen 5000 platform that supports 4th generation pci-e, part of the improvement is due to this.

A network of reviewers testing RDNA2 cards on the 3rd generation pci-e channel, which is what Intel 10s has, detracts a bit from performance.

This is a point to keep in mind.


Now we see how complex the business is, and let's not build a whole thesis from online reviews, without understanding the running environment

Fluently. Enough going through a generation Outdated for the modern Generation Risen 5000, the business looks different.


Also clearly see that there are bottlenecks mainly in the outdated platform of , In some places when such cards are strong

Even at 1440P they fail to reach their full potential. The Risen 5000 did release some of the neck of this bottle,

But it may not be the end yet. We have seen that improving aegesa helped more in 1440P than in , Which indicates that there is a bottleneck in her name

At the processor / platform level that has been released a bit.


** Probably only in Empiric and RDNA2 can be adequately saturated today. And from moment to moment it seems that this is the resolution

Correct for load testing while isolating and taking the CPU and platform out of the equation.


Edited By nec_000
Link to content
Share on other sites

Quote of Lanzar

I agree with you that if the price is the same, the performance is the same, the post-performance is the same, buy NVIDIA Smarter. Drivers mature in day one, most of the architecture gets most of the performance already in the first drivers. Features that get in nvidia that sometimes arrive late at AMD, And despite all this. 80-90 percent of gamers prefer net performance over performance under ray tracing.


So here the choice is your net.


Another month or so lanzar, another month. Then the 3080ti will shuffle all the cards again and will be required to perform a reevaluate of each feasibility equation.

It's not much longer.


** Although needless to say, that Did you learn the lesson at this launch for a change? 🤔

Its day-1 drivers for RDNA2 were relatively fine, not only in terms of stability, but for the best viewing, 

Also in terms of performance in the games themselves.

Edited By nec_000
Link to content
Share on other sites

The choice is indeed my net. I do not agree that the DXR is particularly negligible given the push it does in the consoles and I doubt that the 3080 ti will change anything in the overall picture. It will probably be announced next month, will not be available (because for how many products can the same core be divided? This is also the problem of Producing 100 products on 7nm of TSMC), as long as it is based on 8nm of The potential for haste will be limited and at current pricing I guess it will cost an MSRP of a thousand dollars and will be available at 1200 at best. 


Anyway, once my card arrives, I want 144fps in everything 🤣. Ahh, first world problems. 


post Scriptum. And because of the lack of more or less everything, it seems to me that I can sell the previous hardware at an excessive price? 

PS Maybe it would have been better for them to postpone this generation for half a year instead of knocking themselves the reputation?👹

Edited By DrShawarma
Link to content
Share on other sites

In the meantime, while I happen to be wandering through the learning process in favor of gathering knowledge and materials, I can't help but get upset about the lack

The professionalism and integrity of one of the most famous sites in history in the field of reviews, is not known as techpowerup.


In the following link we can see what I'm talking about, he reviews one of the tickets we highly recommend here in the thread,

God- Merc 319, here see the relevant page at the following link:


This is a screenshot of what I'm talking about. The genius named the one who did the reviews on the 6800xt cards didn't even bother

Check and learn, you hurried Knocks about 10% off the performance, and what he did there, it raised the memory frequency

To the maximum that he succeeds before his business collapses, and for him this is the way he presents the performance of these cards

In stock and after , Just see and cry:



In each case it demonstrates the maximum speed, as such where the memory works at 2140mhz instead of 2000mhz.

And I ask, you review not like that, you are supposed to be a professional and experienced not like that, why did you not do homework to study the

The new architecture on Borya thoroughly, and that it does not render at all in front of VRAM but in front of cache?

Why haven't you studied topology until now, 12 years ago Launched in 2008 for generations (transformed

Later to GDDR6) and its error correction protocol, which requires checking, at what operating frequency the performance is optimal,

And when do they go down because of the errors (because there is no other way to know, the collapse occurs long after there are already errors), and as long as there are no

Crash and you are within range of errors, retransmits of transmissions slow down performance. Every experienced reviewer I have ever known

Know it. And this particular site with the old name isn't it?

In fact it is a mistake that these days seems to cross , So TPU are not alone in the cauldron.


So all he sees is her name in the column on the right is about 10% less than what the card is capable of doing.

And it's a site with tens of thousands of visitors, if not over 100 readers. This is disinformation that he and many other reviewers

Too across the grid produce due to ignorance.


Continuation of the test shows that the testing platform of the TPU site is outdated in terms of the domain and still stands on

9900K rushed to 5.0ghz:


We have no doubt at this stage, because we have already seen, that this system impairs the performance of today's newest and most powerful cards.

Even Generation 10 does not allow to maximize the issue.


This is a huge waste of , For not employing PR professionals, who would explain to reviewers, that they do

Wrong, teach them in advance training and after-practice feedback training, activate the card when the frequency is memory

At the right speed and display proper performance.


I have said many times and I will answer, our forum here in the land of blue and white is one of the highest quality professionally I

Recognizes a horizontal international level. It's respect and pride. There are more of these of course, but they are few and among the drops should be found.


All in all, a consumer who does not do sufficient testing work, sees the reviews of reference models for all the ailments in which,

For him this is what this product knows how to do, he does not see a satisfactory improvement in speed due to the mistake in setting VRAM no

Correct, thus closing the gap for him.


A consumer / amateur who delves into and examines things thoroughly, who cares and interests him, arrives at the right essence.

There is at least 20 percent (or more - depending on the scenario) of product performance. Just need to know how to extract them correctly.

Both in terms of choosing the right card (we brought two models for example), and in terms of the proper canopy (not to rush VRAM).

The advanced enthusiast understands that it is better to work with Risen 5000 if possible, including his direct access to card memory, including pci-e

4th generation, and includes all the latest aegesa updates to motherboards (only today came out one that improves and we saw some).

Beyond those the advanced consumer understands that the computer must be optimized, that is, to speed up our hardware properly, to configure infinity

Fabric for optimum which is 2000mhz, to calibrate a synchronization ratio against memories 1: 1 with - 4000mhz, and can also be tightened

Predatory timings, it goes without saying. to choose dual rank also not to be forgotten. All in order to give the GPU the

The best infrastructure we can provide, so that no bottleneck is created at the CPU and platform level. Let the GPU

The so-called, the chance to work at its best. 

Edited By nec_000
Link to content
Share on other sites

Quote of Jabberwock

What memories are installed on the test system?

We will leave it to lanzar to reply.

In the meantime as far as I know and with the necessary caution (he will correct if I am wrong), Which runs at 4000mhz is relatively risky

1: 1 against the infinity cache that works at 2000MHz. Of course dual channel and also dual rank. This is the optimum in the Risen system

5000 we know.

* The processor has a Risen 5900X at 4.8ghz all (Can do a little more, but do not want to bring it to an end).

Edited By nec_000
Link to content
Share on other sites

Join the discussion

You can then join the discussion and then join our community. If you already have an account with us, please Log in now To comment under your username.
Note that: The comment will appear to the surfers after approval by the board management team.

Add a comment

×   The content pasted is with formatting.   Remove formatting

  Only 75 emoji are allowed.

×   Your link has been automatically assimilated.   Show as regular link

×   Your previous content has been automatically restored.   Clear all

×   You can not paste images directly. Upload or insert images from URL.


  • Create new ...

At the top of the news:

new on the site