AMD 6000 reviews thread

[BFG10K]

AMD Radeon RX 6800 XT Review - NVIDIA is in Trouble

The Radeon RX 6800 XT is a worthy opponent to the NVIDIA GeForce RTX 3080. It offers similar performance at better power efficiency and lower noise levels. In our Radeon RX 6800 XT review, we also take a closer look at raytracing and "Rage Mode", a new 1-click method for overclocking.

www.techpowerup.com

Wow, even the number 3 card has 16GB VRAM and is faster than the 2080TI. And the $1000 6900XT matches the $1500 3090 in performance.

The 3000 parts don't look so hot now.

Post reviews edit:
It's astonishing what AMD have managed to achieve with both the Ryzen 5000 and the Radeon 6000, especially given the absolutely minuscule R&D budget and resources compared to nVidia/Intel. Lisa Su is definitely the "Steve Jobs" of AMD with such a remarkable turnaround.

6900XT:
(It's absolutely amazing to see AMD compete with the 3090)

AMD Radeon RX 6900 XT Review - The Biggest Big Navi

AMD's Radeon RX 6900 XT offers convincing 4K gaming performance, yet stays below 300 W. Thanks to this impressive efficiency, the card is almost whisper-quiet, quieter than any RTX 3090 we've ever tested. In our review, we not only benchmark the RX 6900 XT on Intel, but also on Zen 3, with fast...

www.techpowerup.com

 

[antihelten]

15-20% where?They said 18% average with SAM enabled.And SAM working only on ryzen 5000+B550/X570 and add another 5-11% performance.So it will be like 8-10% in wide game test without it.

SAM isn't that much performance. Based on the slide AMD showed, SAM plus RAGE is an average of 5.5%, or 6% if you ignore the outliers of DOOM and Forza.

However this number includes Rage, which AMD has stated is responsible for 1-2% of the performance uplift (source). So SAM on it's own would bring about 4-4.5% on average.

The slide from AMD of the 6800 showed it roughly 15% ahead of the 2080 Ti with SAM, so it would be about 10% without SAM.

So not 15-20% faster, but not $200 more either.

NVIDIA doubled the price for 10% higher performance and double the VRAM

But basically no one agrees with that price though, so that's hardly relevant.

 

[etrin]

I hope they sell a lot of these and it keeps nvidia inline on pricing. BUT I have head too many horror stories about drivers and SLOW updates. That scares me too much to try amd.

 

[DrMrLordX]

And I REALLY miss HBM lol...sooo much easier to cool/watercool than GDDRX cards. The 5700XT I have is a nightmare to cool compared to the Vega I have because the GDDR6 vram runs so hot.

Considering the prices they're charging for these cards, why NOT use HBM2e? Little confused about that one.

 

[AMD64Blondie]

I'm in a quandary trying to decide(I'm waiting until prices drop on both the Radeon RX 6800,and the RTX 3070),which company to go for.

My current card is a Nvidia GTX 1080 8 GB, hooked up to a 32-inch 1440p Samsung LCD monitor.

The last AMD video card I had was a Radeon X1900XT...IIRC.

 

[GrumpyMan]

So will anyone actually have these in stock to buy in Nov?

 

[Kenmitch]

Looks like we go the 1st official paper launch from PowerCooler

PowerColor teases Radeon RX 6800 XT Red Devil - VideoCardz.com

PowerColor preps Radeon RX 6800 XT, announces giveaway PowerColor is teasing its upcoming Red Devil, a custom variant of the just-announced Radeon RX 6800 XT graphics card. In a very short video posted on social media, the company reveals that the Red Devil will be a triple-fan design with a...

videocardz.com

 

[igor_kavinski]

AMD may not be in any hurry to release the 6700 and lower SKUs. They will first let their partners clear out the 5000 series stock. Come November 18, who anticipates that they deliver the exciting news to current 5000 series owners that a new driver will make their cards fully "DX12 Ultimate"-compliant?

 

[insertcarehere]

Come November 18, who anticipates that they deliver the exciting news to current 5000 series owners that a new driver will make their cards fully "DX12 Ultimate"-compliant?

Don't hold your breath, DX 12 Ultimate requires hardware RT (among other things) which RDNA1 almost certainly doesn't have.

 

[dzoni2k2]

Looks like we go the 1st official paper launch from PowerCooler

PowerColor teases Radeon RX 6800 XT Red Devil - VideoCardz.com

PowerColor preps Radeon RX 6800 XT, announces giveaway PowerColor is teasing its upcoming Red Devil, a custom variant of the just-announced Radeon RX 6800 XT graphics card. In a very short video posted on social media, the company reveals that the Red Devil will be a triple-fan design with a...

videocardz.com

Literally.

 

[thilanliyan]

I'm in a quandary trying to decide(I'm waiting until prices drop on both the Radeon RX 6800,and the RTX 3070),which company to go for.

My current card is a Nvidia GTX 1080 8 GB, hooked up to a 32-inch 1440p Samsung LCD monitor.

The last AMD video card I had was a Radeon X1900XT...IIRC.

Wait for reviews then decide.

 

[n0x1ous]

so 6900xt is not being offered by AIBs ?

 

[tviceman]

I'm happily shocked at the jumps AMD is claiming. Waiting for final reviews, but everything looks great. RT and their answer to DLSS is still up in the air, but efficiency looks fantastic.

A 6800 XT just may be in my future!

 

[B-Riz]

Kinda bummed no RDNA2 cards $500 or less right now.

Hopefully launch day forward there are not major driver issues.

 

[ModEl4]

TPU do not test Bordelands with Badass mode and they use the same API for both cards. AMD said they used the best API for each card.

Cross comparing FPS numbers for different setups is a folly, don't do it. Far too many variables to account for.



They also showed the 5700XT as being faster than the 2070 and then when the real benchmarks came out the 5700XT was ahead by a few % more than AMD claimed in their event.

On top of that they avoided games like Death Stranding, F1 2020, Horizon:Zero Dawn where the 5700XT is faster than the 2070S @4k so they could have cherry picked a lot harder if they had wanted to.

That's why I didn't mention a specific fps result for Borderlands 3 in contrast with Gears 5. Still think the Borderlands result is around 2-3% off and the Gears result is even more off.
in any case I didn't take into account these differences, I think that the results will be at least -2.5% in relation with what AMD's slides are suggesting based on a more fair testing with more games. I just mentioned it because it's not out of the realm of possibility 6900XT to be just 1% better than 3080 in 4K in a 10900K without Rage which will be a huge fail at $999...
Based on Navi 22 rumors (2.5GHz, 64RBEs, 4 primitives, 40CU, 192bit bus, infinite cache) probably will be just 30% better than a 5700XT at most (meaning 2070S will be just -17% & 2080S just -4% vs 6700XT) Guys here said it will be around +5% vs 2080Ti, then it was at 2080Ti level, a few days after it was 2080Ti -5%, now we are at 2080S +10%, lol when does it stop?

 

[Hitman928]

That's why I didn't mention a specific fps result for Borderlands 3 in contrast with Gears 5. Still think the Borderlands result is around 2-3% off and the Gears result is even more off.
in any case I didn't take into account these differences, I think that the results will be at least -2.5% in relation with what AMD's slides are suggesting based on a more fair testing with more games. I just mentioned it because it's not out of the realm of possibility 6900XT to be just 1% better than 3080 in 4K in a 10900K without Rage which will be a huge fail at $999...
Based on Navi 22 rumors (2.5GHz, 64RBEs, 4 primitives, 40CU, 192bit bus, infinite cache) probably will be just 30% better than a 5700XT at most (meaning 2070S will be just -17% & 2080S just -4% vs 6700XT) Guys here said it will be around +5% vs 2080Ti, then it was at 2080Ti level, a few days after it was 2080Ti -5%, now we are at 2080S +10%, lol when does it stop?

AMD's 3080 numbers in Borderlands 3 are right in line with what Hexus got when testing Badass quality.

Review: Nvidia GeForce RTX 3080 Founders Edition - Graphics - HEXUS.net

m.hexus.net

Their Gears 5 numbers line up too:

Obviously different setups may swing fps a little bit one way or another and we'll get the full picture when reviews come out, but at least in the two games you mentioned, it doesn't look like AMD was doing anything to decrease the 3080's performance in their comparison and several of the titles they chose weren't exactly AMD favorable titles, at least not for RDNA1 cards. I'd be pretty shocked if the 6900XT (without rage or SAM) is only 1% better than the 3080 at 4K, that would mean there would be no appreciable difference between a 6800XT and 6900XT.

 

[alexruiz]

Looks like we go the 1st official paper launch from PowerCooler

PowerColor teases Radeon RX 6800 XT Red Devil - VideoCardz.com

PowerColor preps Radeon RX 6800 XT, announces giveaway PowerColor is teasing its upcoming Red Devil, a custom variant of the just-announced Radeon RX 6800 XT graphics card. In a very short video posted on social media, the company reveals that the Red Devil will be a triple-fan design with a...

videocardz.com

I am looking forward to that one!
In my eyes, the Red Devil was the top RX 5700 XT.

 

[Timorous]

AMD's 3080 numbers in Borderlands 3 are right in line with what Hexus got when testing Badass quality.

Review: Nvidia GeForce RTX 3080 Founders Edition - Graphics - HEXUS.net

m.hexus.net

Their Gears 5 numbers line up too:

Obviously different setups may swing fps a little bit one way or another and we'll get the full picture when reviews come out, but at least in the two games you mentioned, it doesn't look like AMD was doing anything to decrease the 3080's performance in their comparison and several of the titles they chose weren't exactly AMD favorable titles, at least not for RDNA1 cards. I'd be pretty shocked if the 6900XT (without rage or SAM) is only 1% better than the 3080 at 4K, that would mean there would be no appreciable difference between a 6800XT and 6900XT.

I did some maths on 6800XT 'stock' vs an estimated 6900XT 'stock' and the 6900XT was about 8% faster on average. If we say stock 6800XT = 3080 (FE?) and use the computerbase 3090 FE review we get the 3090 FE being 11% faster than the 3080 FE on average so I expect the 3090 FE would be around 3% faster than the 6900XT without rage and SAM turned on. It seems like Rage should probably be turned on all the time so that will go down to a virtual tie and then if you have Ryzen 5k you get SAM to push you ahead.

Apparently a SAM like feature has been available in Linux for a long time now and can be used on most platforms with many GPUs so I would not be surprised to see it supported with other CPUs in the future but for now using it to push more Ryzen 5k systems is a smart business play.

 

[Geranium]

Looks like we go the 1st official paper launch from PowerCooler

PowerColor teases Radeon RX 6800 XT Red Devil - VideoCardz.com

PowerColor preps Radeon RX 6800 XT, announces giveaway PowerColor is teasing its upcoming Red Devil, a custom variant of the just-announced Radeon RX 6800 XT graphics card. In a very short video posted on social media, the company reveals that the Red Devil will be a triple-fan design with a...

videocardz.com

How come a teaser a paper launch??? Are you comparing apples to oranges.

 

[amrnuke]

I have some qualms with the pricing to an extent. That being said, the 6800XT and 6900XT look like great values. It appears these cards will equal (with Rage and SAM) their competitor. The 6900XT particularly is looking like a killer for the 3090. The smart buy seems to be the 6800XT if it really does roughly parallel the 3080 in 4K (and 1440p) performance, as it's only $70 more than the 6800 (a touch over 10%), yet offers a 30% gain in 4K performance.

Unless the 6800 drops to $479 or so I can't see it being viable against the 6800XT or the 3070. I think that at $479 you can justify that it represents a massive gain in performance (34%) at 4K over 5700XT with only a 17% increase in cost, which is what we all want - performance gains that exceed price increase.

Card	Price	VRAM	TFLOPS	Est. 4K
RX 5700	$349	8GB	7.9	59%
RX 5700XT	$399	8GB	9.75	66%
2080 Ti	$999-$1199	11GB	13.4	99%
3070	$499	8GB	20.4	100%
RX 6800	$579	16GB	13.9	~113% (SAM)
3080	$699	10GB	39.8	131%
RX 6800XT	$649	16GB	18.6	~131%
3090	$1499	24GB	35.7	144%
RX 6900XT	$999	16GB	20.6	~144% (RAGE, SAM)

 

[Z15CAM]

Man, after the RTX 3070 and RX 6800 introduction, I really bleed for those that bought a $1200 RTX 2080 Ti resonantly. Do not believe they could sell their used one for over $400 now.

On the other hand, I doubt anyone who can afford a $1200 GPU much cares !

It's unbelievable a GPU worth $1200 a month ago is worth less then $450 now !

That $70 charge difference between the RTX 3070 Vs the RX 6800 is just away AMD is rubbing nVidia's nose in the mud after all the years nVidia price gouging - LOL

I'm tossed between the RTX 3070 and RX 6800 or spend $100 more for an RX 6800XT that bests the RTX 3090 for less money; wishing I could afford $1000 to beat the $1200 RTX 3950 ;o(

 

[ModEl4]

AMD's 3080 numbers in Borderlands 3 are right in line with what Hexus got when testing Badass quality.

Review: Nvidia GeForce RTX 3080 Founders Edition - Graphics - HEXUS.net

m.hexus.net

Their Gears 5 numbers line up too:

Obviously different setups may swing fps a little bit one way or another and we'll get the full picture when reviews come out, but at least in the two games you mentioned, it doesn't look like AMD was doing anything to decrease the 3080's performance in their comparison and several of the titles they chose weren't exactly AMD favorable titles, at least not for RDNA1 cards. I'd be pretty shocked if the 6900XT (without rage or SAM) is only 1% better than the 3080 at 4K, that would mean there would be no appreciable difference between a 6800XT and 6900XT.

Let's hope you're right, my assumption was that Borderlands was 2-3% off and Gears 3-4% off, averaging 2.5% in the low end 3.5% in the upper end. I'm not going to nit pick reviews trying to prove my case (I don't care), trying to find 2.5%, it doesn't natter so much, the difference is very small anyway.
77fps on 3950X in Nexus, 77fps on 5900X at 4K, maybe 5900X vs 3950X doesn't gain even 1fps at 4K in Gears 5.
We should wait for reviews, in any case AMD had a good showing, I'm not even worrying too much about the absence of ray tracing results, it will be a little bit slower but it will be just fine, every game eventually will be optimized for consoles, no big deal, what I want to see (when is ready) is the quality of SuperRes vs DLSS, this would be interesting.

 

[Timorous]

I have some qualms with the pricing to an extent. That being said, the 6800XT and 6900XT look like great values. It appears these cards will equal (with Rage and SAM) their competitor. The 6900XT particularly is looking like a killer for the 3090. The smart buy seems to be the 6800XT if it really does roughly parallel the 3080 in 4K (and 1440p) performance, as it's only $70 more than the 6800 (a touch over 10%), yet offers a 30% gain in 4K performance.

Unless the 6800 drops to $479 or so I can't see it being viable against the 6800XT or the 3070. I think that at $479 you can justify that it represents a massive gain in performance (34%) at 4K over 5700XT with only a 17% increase in cost, which is what we all want - performance gains that exceed price increase.

Card	Price	VRAM	TFLOPS	Est. 4K
RX 5700	$349	8GB	7.9	59%
RX 5700XT	$399	8GB	9.75	66%
2080 Ti	$999-$1199	11GB	13.4	99%
3070	$499	8GB	20.4	100%
RX 6800	$579	16GB	13.9	~100%
3080	$699	10GB	39.8	131%
RX 6800XT	$649	16GB	18.6	~131%
3090	$1499	24GB	35.7	144%
RX 6900XT	$999	16GB	20.6	~144%

The 6800 is more like ~115% in your chart. No idea where you got 100% from.

 

[antihelten]

AMD's 3080 numbers in Borderlands 3 are right in line with what Hexus got when testing Badass quality.

As can be seen from this roundup from 3Dcenter.org, Hexus isn't the only site that got numbers like this.

In fact it's pretty clear that Techpowerup is the outlier here (since they didn't use Badass settings).

The same thing can also be seen for Gears 5, where Techpowerup again got the highest results. Looking into this a bit more, it looks like TPU runs their Gears 5 bench without motion blur enabled, which apparently can make a difference of about 5%. So with motion blur enabled TPU would have gotten somewhere around 80 FPS

With all that being said though, there is definitely something weird going on with AMDs Gear results. During the presentation they reported that the 6800XT gets 78 FPS, however if you go to their website, they report that it gets 84 FPS. One possible explanation might be that the latter result is from enabling Rage and SAM.

All in all, none of this really matters though. Either the 6800 XT is equal to the 3080 (78 versus ~77 FPS going with the low results) or it is still equal to the 3080 (84 versus 84,4 FPS going with the high results).

 

[amrnuke]

The 6800 is more like ~115% in your chart. No idea where you got 100% from.

I think 110-115% is fair, updated. I was trying to estimate with SAM off, but we truly don't have data so I updated chart to reflect enhancements AMD used.

 

[Hitman928]

With all that being said though, there is definitely something weird going on with AMDs Gear results. During the presentation they reported that the 6800XT gets 78 FPS, however if you go to their website, they report that it gets 84 FPS. One possible explanation might be that the latter result is from enabling Rage and SAM.

If you look at the footnotes on the website results, SAM and rage mode are on versus off in the presentation.