confirmed: AMD's big Navi launch to disrupt 4K gaming

According to tech YouTuber Corteks, who has shared information from sources in Asia, AMD’s high-end RDNA 2 GPUs are merely on par with green team’s GeForce RTX 2080 Ti.

“A best case scenario supposedly shared by AMD is that at most gamers should expect to get about 15% over the 2080ti in a select AMD ‘optimized’ titles.

According to the same sources AMD is aiming to launch the ‘Big Navi’ as a direct competitor to the soon to be launched RTX 3080, and not as a competitor to NVidia’s highest performing part as has been widely reported (let alone beat it). Some have suggested that ‘Big Navi’ would be up to 50% faster in traditional rasterization than the RTX 2080ti but according to AMD’s partners that will not be the case.”
Click to expand...
~ wrote Corteks.

https://www.thefpsreview.com/2020/0...navi-only-15-percent-better-than-rtx-2080-ti/

https://hardforum.com/threads/amds-...match-for-nvidia-amperes-second-tier.1999515/

erek said:
"Coreteks' AIB sources also suggest there will be a pair of Big Navi GPUs at launch, based on the same Sienna Cichlid silicon. That's not much of a stretch as this is classic AMD graphics card practice; it's very own rule of two. This typically sees a cut-down card launching alongside the full-fat option. Just think about the RX 5700 XT and RX 5700 last year, RX 580 and RX 570, R9 390X and R9 390… and on and on.

And traditionally we usually end up recommending the lower-spec one as its lower price, and generally only-very-slight specs cut, end up making it a better value card by comparison, offering often very similar performance.

It looks like those two Sienna Cichlid cards will be the only Big Navi GPUs we see this year too, if this report is to be believed. A second fish-based codename, Navy Flounder, has been tied to the new RDNA 2 architecture, and is reportedly a mid-range version of the Big Navi GPU slated for release in the first three months of 2021.

With only the high-end market set to be served at the tail-end of this year, it seems almost certain that most of us are going to have to wait until next year to see anything more affordable than the ultra-enthusiast, ultra-expensive GeForce and Radeon cards."

https://www.pcgamer.com/amd-big-navi-rdna-2-aimed-at-rtx-3080/
Click to expand...
 
Not too bad, considering that the 2080ti is a very good card.

AMD buyers never cared about having the best anyhow, so price it right and it will keep the fan base happy.
 
Marees said:
According to tech YouTuber Corteks, who has shared information from sources in Asia, AMD’s high-end RDNA 2 GPUs are merely on par with green team’s GeForce RTX 2080 Ti.
Click to expand...
Don't believe everything you read/hear on the internet. I like Corteks industry analysis, but I think he product information is a bit off this time.
 
variant said:
I simply heard that AIBs would be very late to the game which is why AMD has opted to skip the blower reference cards. I don't have a link.
Click to expand...
Wow. Can you make any more posts with zero sources and bad information?
 
variant said:
I simply heard that AIBs would be very late to the game which is why AMD has opted to skip the blower reference cards. I don't have a link.
Click to expand...
Seems weird for AIB to be late in the game when it is AMD steering the ship.
 
1_rick said:
It might be a good idea to provide at least one of those sources.
Click to expand...

🤷‍♂️

Is it that important? Should I create a website and release source information to compete with someone else's website that releases source information?

Having a URL doesn't really give anymore legitimacy to one rumor over another.
 
Last edited:
  • Like
Reactions: Axman
like this
Marees said:
AMD’s high-end RDNA 2 GPUs are merely on par with green team’s GeForce RTX 2080 Ti.
Click to expand...
"Merely"? The 2080 Ti is a beast of a card. If AMD can get that performance, at an affordable price, it would be a huge deal.
 
variant said:
Prove to me it's bad information, also there are definitely sources to it.
Click to expand...
I think you either misread or misunderstood:
"The new AMD Big Navi graphics card is expected to launch in late 2020 just as AMD CEO Dr. Lisa Su promised, but we're expecting the first batches of Big Navi cards to be in reference form -- no custom cards at first."
https://www.tweaktown.com/news/7396...ler-flagship-card-has-16gb-of-memory/amp.html

Reference form doesn't mean no AIB cards, just they won't be custom cards. Same as their other launches, reference design first, custom cards later.
 
Ready4Dis said:
Reference form doesn't mean no AIB cards, just they won't be custom cards. Same as their other launches, reference design first, custom cards later.
Click to expand...
Exactly. Pretty normal for AIB cards to follow the in-house cards nowadays. This way the MFG keeps control over the product information for just a bit longer. Lord knows as soon as it goes to China it leaks.
 
FrgMstr said:
That is not how it works. You are posting LIES/FUD/RUMORS. Stop making shit up and treating other forums like a source of correct information.
Click to expand...

This entire thread is based on rumors. My sources have nothing to do with any forums.
 
FrgMstr said:
Exactly. Pretty normal for AIB cards to follow the in-house cards nowadays. This way the MFG keeps control over the product information for just a bit longer. Lord knows as soon as it goes to China it leaks.
Click to expand...
Yup, as I said, same as last launches, which is why nobody is talking about it besides people who misread.
 
IdiotInCharge said:
Ouch...

I believe that AMD can do better, but have simply chosen not to (again).
Click to expand...
They have to keep their manufacturing limitations in mind, would it be better to produce a line that goes after the top end at the expense of supply for the mid and low end? I don't think that would be a good idea for AMD, they need volume more than a niche part of the market. Huge monolithic cores use large parts of a waifer, and then yeah you can take those failures and roll them down to some extent but would it be better to make 40 Top tier chips or 100 smaller ones for the mid tier. Mid and low move more product and that's where there is a lot more competition so I think it is totally a smart Idea to use their valuable TSMC GPU time on that segment instead. I however, have no doubt that AMD could release a product that would dethrone NVidia's top end I just don't think they should at this point, maybe in 3 years.
 
kirbyrj said:
Maybe by that time DLSS will have more than 10 games ;).
Click to expand...

Supposedly DLSS 3.0 will be on by default on any Title that uses TAA, so programmers won't have to do anything to enable DLSS any more. They implement TAA and on the new Ampere cards by using TAA you are automatically using DLSS 3.0, that will be huge if true.
 
variant said:
This entire thread is based on rumors. My sources have nothing to do with any forums.
Click to expand...
You are making some very specific statements. I would think that you would at least have some sort of information to back those up. Fair?
 
Lakados said:
Supposedly DLSS 3.0 will be on by default on any Title that uses TAA, so programmers won't have to do anything to enable DLSS any more. They implement TAA and on the new Ampere cards by using TAA you are automatically using DLSS 3.0, that will be huge if true.
Click to expand...
It still requires developer support. It does make it easier for them, but it's not automatic on games that weren't designed for it.
DLSS 3.0 will reportedly "work in any game with TAA" but it will require a Game Ready driver to do so, meaning developers will have to do some "specific programming per game to get it to work, but it should be easier than before".
Click to expand...
https://www.tweaktown.com/news/7388...uld-work-on-any-game-that-uses-taa/index.html
 
sabrewolf732 said:
My sources tell me big navi is 2080ti performance at $499 and 100 watt TDP

Sorry, I can't reveal my sources. I have zero history of accurate rumors. This is just what I've heard.
Click to expand...


The 2080ti has a die size of 754mm and is on 12nm

Big Navi is said to have a die size of 505mm and still on 7nm.

Does anyone know how much more dense 7nm is compared to 12nm on those two lines? Is it a one to one increase in density where one might be able to speculate on performance gains, or is it completely decoupled?
 
If it's a 1:1 shrink, which it's not, a 754mm 12mm die would be approximately 440mm on 7nm.

Math check? 7 / 12 = .5833 x 754 = 439.833

505 is 114.77% the area. Again, though, shrinks aren't straight across the board.
 
Axman said:
If it's a 1:1 shrink, which it's not, a 754mm 12mm die would be approximately 440mm on 7nm.

Math check? 7 / 12 = .5833 x 754 = 439.833

505 is 114.77% the area. Again, though, shrinks aren't straight across the board.
Click to expand...

7nm and 12nm are not true numeric representations of process size. You can't compare them mathematically based on the process name, anymore than if they named them purple and yellow.

You can look at the relative practical transistor densities. Which for GPUs is about 25m/mm2 for TSMC 16/12nm and about 40m/mm2 for TSMC 7nm.

754mm2 * 25/40 = 471mm2 if you die shrunk it to 7nm. Either way this is around 18 Billion transistors.

505mm2 * 40/25 = 808mm2 if you expanded that to 16/12nm. Either way this is around 20 Billion transistors.

So about 10% more transistor could translate into about 10% more performance if all else is equal, plus more swing on that depending on a bunch of factors. Like making sure the memory bandwidth is adequate, and the how clock speed changes, any new enhancements etc....

Of course that would only matter if the rumors were actually trustworthy, and not just made up, by someone essentially doubling everything from 5700Xt.
 
A 505mm² Navi 1 would be competitive with Nvidia's high end. That's over twice the size of the 5700 XT, and a good chunk of the die size is eaten up by the memory controller which wouldn't change going to a larger die size. If Navi 2 uses HBM2e, that is even more room freed up on the silicon for other things. Considering the 5700 XT has 40 CUs, it seems ridiculous to think the Navi X2 will only have 64. The rumored 80 CU seems more likely, but I would think we could see even more on the largest Navi, and the consoles have demonstrated that the RDNA2 easily clocks higher with the newer 7nm process.

I am going to take a guess that the AIBs have Navi 23, as opposed to Navi 21, which is rumored to be 340mm² die size. The Xbox Series X APU is 360mm², with the Zen 2 portion taking up 148mm², and it has 52 CUs, with I think 2 or 4 CUs disabled which takes up 220mm². If you simply scaled it up, a 505mm² would have way over 100 CUs.
 
Last edited:
Marees said:
Another "leak" from _rogame of Hardware Leaks .com

_rogame (@_rogame) Tweeted:


https://twitter.com/_rogame/status/1273792793144688640?s=20

View attachment 254864
Click to expand...

Looks like rogame of hardwareleaks.com is doubling down on his claim of Big Navi == 2x Navi 10 (5600 XT, 5700, 5700 XT)

Prominent leaker _rogame has shared a document (via VideoCardz) that reveals the purported shader engine (4), shader arrays per shader engine (2), and compute units per shader array (10) counts for AMD’s upcoming “Sienna Cichlid” GPU, otherwise known as “Big Navi.” Based on some simple math, these specifications suggest that red team’s next flagship Radeon card could boast 80 Compute Units (4 x 2 x 10), which would imply 5120 Stream Processors (80 x 64 cores).
Click to expand...

https://www.thefpsreview.com/2020/0...have-80-compute-units-5120-stream-processors/

Marees said:
A 6900XT that is exact doubling of non-overclocked 5600 XT. This should be 15%?? better than an RTX 2080 ti & again reasonable to expect $999 unless AMD implements fancy cooling solutions that push the power all the way up, in which case it could be $1200 or more
Click to expand...

Big Navi : Navi21
  • Up to 80 compute units/5120 GPU cores
  • A die size of around 505mm²
  • 50% better performance per watt
There are 4 Navi21 variants for the gaming market :

Screen-Shot-2020-05-22-at-22.59.18-1024x416.png

https://hardwareleaks.com/2020/05/23/exclusive-future-amd-gpu-stack-big-navi-navi10-refresh/

The data has been found and decoded by the leakers, to whom we promised not to reveal the original source in this article.
Click to expand...
https://videocardz.com/newz/amd-sienna-cichlid-navi-21-big-navi-to-feature-up-to-80-compute-units

Also DISCLAIMER:

Marees said:
When I started this thread in the beginning of the year, I expected some concrete information to be available by now

But it looks like we will be running on leaks & rumors till end of the year !!!
Click to expand...
 
variant said:
If Navi 2 uses HBM2e, that is even more room freed up on the silicon for other things.
Click to expand...

The default memory configuration for all Navi 2x series seems to be GDDR & not HBM

Power considerations suggest a doubling of 5600 XT for big Navi which means 12gb vram is likely. There could be a top-end / pro card with 24 gb ram too.

And if I were to take a guess, iff AMD designed big Navi to support both GDDR & HBM and iff big Navi is a hit and/or performs better than RTX 3080ti then AMD might be tempted to release a 16GB/24GB HBM card to take on RTX 390/titan !!?
 
Marees said:
The default memory configuration for all Navi 2x series seems to be GDDR & not HBM

Power considerations suggest a doubling of 5600 XT for big Navi which means 12gb vram is likely. There could be a top-end / pro card with 24 gb ram too.

And if I were to take a guess, iff AMD designed big Navi to support both GDDR & HBM and iff big Navi is a hit and/or performs better than RTX 3080ti then AMD might be tempted to release a 16GB/24GB HBM card to take on RTX 390/titan !!?
Click to expand...

There are rumors of both GGDR6 and HBM on RDNA2 cards, so that doesn't indicate any such default memory configuration. 'Big Navi' is the 505mm² behemoth, the others are just smaller configurations on a different die sizes.
 
You must log in or register to reply here.
Back
Top