Wednesday, October 21st 2020

AMD Radeon RX 6000 Series Specs Leak: RX 6900 XT, RX 6800 XT, RX 6700 Series

AMD's Radeon RX 6000 series graphics cards, based on the RDNA2 graphics architecture, will see the introduction of the company's first DirectX 12 Ultimate graphics cards (featuring features such as real-time raytracing). A VideoCardz report sheds light on the specifications. The 7 nm "Navi 21" and "Navi 22" chips will power the top-end of the lineup. The flagship part is the Radeon RX 6900 XT, followed by the RX 6800 XT and RX 6800; which are all based on the "Navi 21." These are followed by the RX 6700 XT and RX 6700, which are based on the "Navi 22" silicon.

The "Navi 21" silicon physically features 80 RDNA2 compute units, working out to 5,120 stream processors. The RX 6900 XT maxes the chip out, enabling all 80 CUs, and is internally referred to as the "Navi 21 XTX." Besides these, the RX 6900 XT features 16 GB of GDDR6 memory across a 256-bit wide memory interface, and engine clocks boosting beyond 2.30 GHz. The next SKU in AMD's product stack is the RX 6800 XT (Navi 21 XT), featuring 72 out of 80 CUs, working out to 4,608 stream processors, the same 16 GB 256-bit GDDR6 memory configuration as the flagship, while its engine clocks go up to 2.25 GHz.
A notch below the RX 6800 XT is the RX 6800 (Navi 21 XL), which cuts down the "Navi 21" further, giving it 64 compute units or 4,096 stream processors; the very same 16 GB of 256-bit GDDR6 memory interface, and up to 2.15 GHz engine clocks. The RX 6900 XT, along with the RX 6800 series, will be announced in the October 28 presser.

The next chip AMD is designing is the 7 nm "Navi 22" silicon, which features 40 compute units. On paper, this count looks similar to that of the "Navi 10," and it remains to be seen if this is a re-badge or a new silicon based on RDNA2. The RX 6700 XT maxes this chip out, featuring 40 CUs or 2,560 stream processors; while the RX 6700 features fewer CUs (possibly 36). The interesting thing about these two is their memory configuration—12 GB of 192-bit GDDR6.
Source: VideoCardz
Add your own comment

191 Comments on AMD Radeon RX 6000 Series Specs Leak: RX 6900 XT, RX 6800 XT, RX 6700 Series

#26
Ibotibo01
I think that RX 6800 would be better value than RTX 3060/3070 and Nvidia doesn't deliver their GPUs. AMD will take advantage if they deliver before Q1 2021.


RX 6900 XT's core count is 2 times of RX 5700 XT so i show GTX 1060 VS GTX 1080 (performance gap is same with between RX 6900 XT and RX 5700 XT but RDNA2 will probably give 1.2 performance boost over RDNA1.

100/57 = 1.754, 1.66*1.2*0.91(Bandwidth) = 1.81 I think 6900 XT is faster than RTX 3080 and 6900XT's performance is between RTX 3090 and RTX 3080.

My prediction about prices and performances
RX 6900 XT $799 faster than RTX 3080 but weaker than RTX 3090
RX 6800 XT $629 faster than RTX 3070 and almost same with RTX 3080's performance
RX 6800 $479 faster than RTX 2080 Ti by %5-10
RX 6700 XT $379 same performance with RTX 3060 Ti but it has 12GB VRAM
RX 6700 $329 same performance with RTX 3060/2080
RX 6600 XT $229 faster than RTX 2060/S/3050 Ti
RX 6500 XT $179 same performance with RTX 3050 Ti/GTX 1660 Ti
RX 6500 $129 same performance with RTX 3050/GTX 1660

All of GPUs are capable of Ray Tracing and even RX 6500 and RTX 3050 give RTX 2060's Ray Tracing performance. All in all, it is my prediction.
Posted on Reply
#27
Ja.KooLit
Ibotibo01I think that RX 6800 would be better value than RTX 3060/3070 and Nvidia doesn't deliver their GPUs. AMD will take advantage if they deliver before Q1 2021.


RX 6900 XT's core count is 2 times of RX 5700 XT so i show GTX 1060 VS GTX 1080 (performance gap is same with between RX 6900 XT and RX 5700 XT but RDNA2 will probably give 1.2 performance boost over RDNA1.

100/57 = 1.754, 1.66*1.2*0.91(Bandwidth) = 1.81 I think 6900 XT is faster than RTX 3080 and 6900XT's performance is between RTX 3090 and RTX 3080.

My prediction about prices and performances
RX 6900 XT $799 faster than RTX 3080 but weaker than RTX 3090
RX 6800 XT $629 faster than RTX 3070 and almost same with RTX 3080's performance
RX 6800 $479 faster than RTX 2080 Ti by %5-10
RX 6700 XT $379 same performance with RTX 3060 Ti but it has 12GB VRAM
RX 6700 $329 same performance with RTX 3060/2080
RX 6600 XT $229 faster than RTX 2060/S/3050 Ti
RX 6500 XT $179 same performance with RTX 3050 Ti/GTX 1660 Ti
RX 6500 $129 same performance with RTX 3050/GTX 1660

All of GPUs are capable of Ray Tracing and even RX 6500 and RTX 3050 give RTX 2060's Ray Tracing performance. All in all, it is my prediction.
I hope youre right. Then Amd would be my next card :)
Posted on Reply
#28
bug
I see nobody noticed the seemingly large gap between the 6700XT and 6800. If nothing else, it means the price difference is most likely >$100 between the two.
One more week to go, fingers crossed.
Posted on Reply
#29
Turmania
I probably get a card that has a 150-160w avg. Gaming usage. That means I have to wait for a while longer, not bad, they would fix more driver/stability issues by then.
Posted on Reply
#30
ShurikN
Here's hoping that the 6700 is no more than 300€
Posted on Reply
#31
bug
ShurikNHere's hoping that the 6700 is no more than 300€
With 192 bit bus, I'd hope the 6700XT is €300 or less.
Posted on Reply
#32
WeeRab
Solid State Soul ( SSS )Performance isn't always the full picture, stable drivers also matters, look how the RX 5000 series launched, terrible drivers for months.
You said that with a straight face too...even after the manifest and well documented problems with Nvidia drivers for the 3800/3900.
I'll settle for products you can actually buy - and not some semi-vapourware, crash-to-desktop, bug-ridden Nvidia cards.
Posted on Reply
#33
RedelZaVedno
RX 6900 XT = 23.55 TFLOPS-32 (75% more than 2080TI)
RX 6800 XT = 20.74 TFLOPS (54% more than 2080TI)
RX 6800 = 17.61 TFLOPS (31% more than 2080TI)
RX 6700XT (if clocked at 2.15GHz) = 11 TFLOPS (18% less than 2080TI/ on pair with 2080S )

Not bad at all... RDNA2 should give Ampere good run for it's money IF AMD chooses to be aggressive on pricing.
Posted on Reply
#34
bug
WeeRabYou said that with a straight face too...even after the manifest and well documented problems with Nvidia drivers for the 3800/3900.
I'll settle for products you can actually buy - and not some semi-vapourware, crash-to-desktop, bug-ridden Nvidia cards.
Well, Nvidia's problems were at least confirmed and worked around quickly ;)
As for availability, I don't care for cards that expensive, so it's all the same to me.
Posted on Reply
#35
kapone32
Nothing but roses for AMD. If all of these cards clock past 2 GHZ then it means that the 6700XT should be about 15% faster than the 5700XT with more VRAM. That is traditional though I want to know more about the rumoured cache on these cards.
Posted on Reply
#36
bug
kapone32Nothing but roses for AMD. If all of these cards clock past 2 GHZ then it means that the 6700XT should be about 15% faster than the 5700XT with more VRAM. That is traditional though I want to know more about the rumoured cache on these cards.
What do you honestly expect from a cache? It will speed-up access as long as you don't exceed its cache and it will do nothing when you do.
NB If it's a cache for VRAM, having to stay within the cache's size would defeat the purpose of more VRAM. But this may be some other type of cache (if it exists at all, I think AMD never confirmed its existence).
Posted on Reply
#37
RH92
QuicksLooking GOOD hope the prices look good as well.
Depends whats your definition of ''prices looking good'' . If by looking good you are expecting an AMD to SKU to perform similarly to an Nvidia counterpart and cost much less then with all respect due you are fooling yourself . AMD doesn't work like this anymore , nowadays when AMD manages to compete with nvidia on performance prices are almost the same , so if you were complaining about Nvidia pricing then im afraid it won't look good for you !
Posted on Reply
#38
ebivan
RedelZaVednoRX 6900 XT = 23.55 TFLOPS-32 (75% more than 2080TI)
RX 6800 XT = 20.74 TFLOPS (54% more than 2080TI)
RX 6800 = 17.61 TFLOPS (31% more than 2080TI)
RX 6700XT (if clocked at 2.15GHz) = 11 TFLOPS (18% less than 2080TI/ on pair with 2080S )

Not bad at all... RDNA2 should give Ampere good run for it's money IF AMD chooses to be aggressive on pricing.
Well, TFLOPS say something about computing performance. Which dosn't really mean gaming performance.
Its like comparing sequential write speeds with real world ssd performance.
Posted on Reply
#39
RH92
MusselsI need some reliable performance leaks before my 3080 gets shipped, so i know if i should cancel or not :D
Im on the same boat here . Im waiting for Caseking to fulfill my TUF 3080 preorder but at this point it won't hurt me to see what AMD can offer . Although to be honest most peoples here judge those GPUs on pure rasterization performance , it's not my case so i will only go for AMD if they are both close on rasterization and raytracing perf .
Posted on Reply
#40
kapone32
bugWhat do you honestly expect from a cache? It will speed-up access as long as you don't exceed its cache and it will do nothing when you do.
NB If it's a cache for VRAM, having to stay within the cache's size would defeat the purpose of more VRAM. But this may be some other type of cache (if it exists at all, I think AMD never confirmed its existence).
It could be a mimic in some way of Infinity fabric. I have only ever seen rumours but I thought that something like that between the GPU and VRAM would be intriguing. In the article it was stating that it could mitigate the small bit bus vs the 3000 series cards and the speed difference between GDDR6 and 6X.
Posted on Reply
#41
Turmania
So 28th will be announcement, but when will they actually launch and up for reviews?
Posted on Reply
#42
kapone32
TurmaniaSo 28th will be announcement, but when will they actually launch and up for reviews?
Hopefully Nov 5 but most likely for CP 2077.
Posted on Reply
#43
bug
kapone32It could be a mimic in some way of Infinity fabric. I have only ever seen rumours but I thought that something like that between the GPU and VRAM would be intriguing.
Infinity Fabric is an interconnect solution, it's nothing like a cache.
kapone32In the article it was stating that it could mitigate the small bit bus vs the 3000 series cards and the speed difference between GDDR6 and 6X.
Again, a cache cannot do this, save for a few, limited scenarios.

I think you're waiting for magic, so you'll be disappointed.
Posted on Reply
#44
medi01
Ibotibo01RX 6800 XT $629 faster than RTX 3070 and almost same with RTX 3080's performance
There is no reason for AMD to undercut 3080, a missing in action card.
Posted on Reply
#45
Jism
I'm so excited about the times we live in. Both green and red camps releasing cards that do 4K sustained 60FPS.
Posted on Reply
#46
HD64G
Logical predictions based on the GPU market today and the leaks:

Full-die 80CU 6900XT for $700
Lightly cut-down die 72CU 6800XT for 600$
Heavily cut-down die 64CU 6800 for 500$

That would be great if 6900XT is above the performance of RTX3080 10GB. If it is on par or slightly behind deduct $50 from those prices. And remember that the 6800 will be almost surely faster on average than 3070 8GB.

As for the RX6700 series my guess is they will be at least 20-25% faster than RX5700 series and will cost less. So, 2080S level of performace for $250-350 (XT and not-XT into that range) would be great for us consumers me thinks.

My 5c.
Posted on Reply
#47
Vya Domus
kapone32In the article it was stating that it could mitigate the small bit bus vs the 3000 series cards and the speed difference between GDDR6 and 6X.
It's not "could", it's "does". That why caches exist, to mitigate the limitations caused by slow main memory. Caches work very well when the access patterns are sequential and guess what, the sort of things you need to compute for graphics, are always like that.
Posted on Reply
#48
bug
Vya DomusIt's not "could", it's "does". That why caches exist, to mitigate the limitations caused by slow main memory. Caches work very well when the access patterns are sequential and guess what the sort of things you need to compute for graphics are always like that.
Except that it doesn't work like that. Cache is much, much smaller than main memory. If you happen to frequently use only, say 16MB of the main memory, those MB will be retained in cache and used from there, resulting in dramatically improved performance. But if the video card wants to access say 8GB, the cache doesn't help much anymore.
Also, cache memory is so fast in part because it's very power hungry.
Posted on Reply
#49
RedelZaVedno
ebivanWell, TFLOPS say something about computing performance. Which dosn't really mean gaming performance.
Its like comparing sequential write speeds with real world ssd performance.
It scaled pretty well till Ampere launch aka 11.15 Tflops 2080S being 14% slower than 13.45Tflops 2080TI, 9,75 Tflops 5700XT being between 15-20% slower than 2800S... I have no doubt 23.55 TFLOPS GPU will compete well against 3080 if not 3090 in standard rasterization performance if we put AI upscaling and RT aside and that's all I really care about for now. Cheapest 4K/90fps capable and most power efficient GPU gets my money.
Posted on Reply
#50
AsRock
TPU addict
AnymalHow they both decided several years ago TBP would be 320W? Smaller node, no problem, lets make it more power hungry.
Yeah your talking a 500W just to play a game lmao.
Posted on Reply
Add your own comment
Apr 25th, 2024 13:52 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts