According to Videocardz’s internal sources some leaked specs for AMD’s next-gen 7nm RDNA 2-based NAVI GPUs have been shared. Even though most of the information posted here may turn out to be true, you should still exercise caution with any early leak or rumor until there is some official confirmation from the company. With that being said, let us focus on the specs.
The GPU clock speeds listed below may be for the Reference Model from AMD, since custom AIB cards will have different clock speed values based on the custom PCB/board.
As per Videocardz, we have 5 new SKUs. The 7nm Navi 21 silicon chips will power the high-end flagship cards, the Radeon RX 6900 XT, followed by the RX 6800 XT and RX 6800, respectively. These are then followed by the RX 6700 XT and RX 6700 SKUs, which are based on the Navi 22 silicon.
The flagship top-end GPU is the AMD Radeon RX 6900 XT, based on the Navi 21 XTX GPU silicon. The RX 6900 XT maxes out the chip, enabling all 80 CUs (compute units) thus amounting to a total of 5,120 stream processors/SPs. This card will be sold in limited quantities and will be exclusive to AMD at least initially.
The RX 6900 XT will have 16 GB of GDDR6 memory across a 256-bit wide memory interface bus, and this GPU would feature a 2040 MHz game clock and a 2330 MHz boost clock speed value.
Next SKU in the lineup is the RX 6800 XT GPU based on the Navi 21 XT silicon. This card is a cut-down variant featuring 72 Compute Units, which amounts to a total of 4608 Stream Processors. This card is having the same 16 GB 256-bit GDDR6 memory configuration as the flagship RX 6900 XT card, while its engine clocks go all the way up to 2.25 GHz.
The game clock is expected to be 2015 MHz and the boost clock should be under 2250 MHz.
Next we have the AMD Radeon RX 6800 GPU based on the Navi 21 XL silicon. According to the data the RX 6800 non-X SKU will feature 64 Compute Units, which gives a total of 4096 Stream Processors.
And similar to the 6900XT and 6800XT, this variant will also come with 16 GB GDDR6 memory and a 256-bit wide memory bus interface. As per the report, the clock speeds for this SKU are lower. The card is expected to have 1815 MHz as the game clock and 2105 MHz as the boost clock speed value.
The Radeon RX 6900XT, 6800XT, and 6800 SKUs will all be officially unveiled on October 28th.
Lastly, we have two more SKUs, the AMD Radeon RX 6700 XT and 6700, respectively, both based on the Navi 22 silicon. AMD plans to launch these cards this coming January. The Radeon RX 6700 XT is expected to feature the full Navi 22 silicon with 40 CUs (or 2560 Stream Processors), having 12 GB of GDDR6 VRAM based on a 192-bit wide memory interface bus.
Since the RX 6700 XT maxes out the Navi 22 chip, we expect the RX 6700 non-XT variant to feature fewer CUs, most probably 36. The memory configuration should remain the same though, featuring 12 GB of GDDR6 VRAM on a 192-bit wide memory interface bus, but this hasn’t been fully confirmed yet.
Similar specs have also been shared by Coreteks as well, with some minor changes to the clock speeds. According to Videocardz, AIBs will release their custom cards next month, but the AMD reference design card will get the first priority. The reference cards will even have a separate review embargo date.
We also have some info to share on one AMD Radeon RX 6800XT board partner custom card. According to Igor Wallossek from Igor’s LAB, an exclusive custom AIB graphics card based on the Navi 21 GPU will feature a boost clock of 2577 MHz.
Igor used the MorePowerTool app to read the data from the BIOS/firmware, which also reveals that the maximum allowed GPU boost clock speed is 2800 MHz, which is an extremely high clock. This frequency can only be achieved via a Liquid Nitrogen cooling solution, so this value makes little sense for most of the gamers.
But nonetheless, it gives us an idea about the efficiency and refinement of the new RDNA 2 GPU architecture powering these cards, assuming this finding from Igor is accurate.
RDNA 2 GPUs are expected to be fabbed on an optimized 7nm process node, and they will support hardware-level ray tracing as well, and are expected to deliver a 50% increase in performance per watt over previous-gen RDNA arch.
Along with an improved performance-per-clock (IPC), and logic enhancement that helps reduce the design complexity and switching power. The GPU clock speeds are also going to get a boost.
AMD plans to showcase its Radeon RX 6000 NAVI GPU lineup on October 28th, based on the new RDNA 2 architecture. John will cover this live event showcase.
Stay tuned for more!
Hello, my name is NICK Richardson. I’m an avid PC and tech fan since the good old days of RIVA TNT2, and 3DFX interactive “Voodoo” gaming cards. I love playing mostly First-person shooters, and I’m a die-hard fan of this FPS genre, since the good ‘old Doom and Wolfenstein days.
MUSIC has always been my passion/roots, but I started gaming “casually” when I was young on Nvidia’s GeForce3 series of cards. I’m by no means an avid or a hardcore gamer though, but I just love stuff related to the PC, Games, and technology in general. I’ve been involved with many indie Metal bands worldwide, and have helped them promote their albums in record labels. I’m a very broad-minded down to earth guy. MUSIC is my inner expression, and soul.
Contact: Email
wants 120- 150 w TBP … gets 300 400 w TBP :/
The power draw has increased this whole generation of GPUs from both Nvidia and AMD unfortunately.
I was expecting more efficiency from the new architecture though.
You don’t even know how powerful they are though. Your comment makes no sense. They could be faster than the 3080 and honestly from the looks of it probably will be in non rtx dlss games.
Assuming these specs turn out to be true, then I’m glad we have minimum 12 GB of VRAM in the lineup, which is still sufficient.
Obviously 16 GB seems to be sweet spot, I doubt a lot of games are going to eat up this much amount of VRAM though. But the Ampere cards currently only have 10 GB so this makes AMD cards a better buy.
The more VRAM a company puts, the expensive the card will become, but gamers still don’t realize this. They just want more and more VRAM on the same price bracket, which isn’t possible imo.
part of more vram could be from the new unreleased microsoft directStrorage api to where they are pre loading textures and such directly from the pcie 4.0 ssd? just a thought.
it’s a shame these cards are not as powerful as Xbox Series X!
These are actually more powerful than the consoles
Yeah, they’re nothing special. These days, you can buy an iPhone 13 for like $299 and play Player Unknown’s Far Crysis at silky smooth 8K 30fps without resorting to gimmicks like AMD’s DLSS or nVidia’s RDNA. Samsung really nailed it with the new iPhone.
Good Read, Thanks Homie ? now all I gotta think about is which card out of the line up skills be more economical. Guess only pricing can tell me that at this point.
So flagship card gets at least 16GB Vram ,.,,,Nice article btw. Will wait till AMD makes these official though.