New Upcoming ATI/AMD GPU's Thread: Leaks, Hopes & Aftermarket GPU's

16GB, performance around RTX 3080 but with 300 wats top (still rumor but seems likely) hardware Ray Tracing and price point around RTX 3080.. I think I will leave nVidia again, I just hope AMD drivers won't disappoint me with some issues like black screens or whatever I'm reading in threads here

 

AMD Sienna Cichlid and Navy Flounder specifications

The next-generation desktop graphics processors based on AMD RDNA2 architecture are expected to debut on October 28th.
AMD did not confirm which GPUs will be unveiled but the manufacturer did announce it will reveal Radeon RX 6000 series, so there are at least two cards coming.

AMD is allegedly developing three Navi 2X GPUs: Navi 21, Navi 23, and Navi 23. Only Navi 21 (Big Navi) has been confirmed to carry a codename of Sienna Cichlid, while the Navy Flounder codename could refer to either 23 or 22.

It is expected that the higher the Navi suffix the lower the GPU will be positioned in the lineup. Thus, the Navi 21 is going to be the top tier GPU, while 23 is likely to land in an entry-level space.

Only leaks Sienna Cichlid and Navy Flounder leaks have been appearing in open-source software patches, such as Linux kernel or, like in this example, ROCm software.

Reddit stblr discovered that the recent ROCm update (3.8) features firmware for Navy Flounder.
This file contains variables defining the configuration of the graphics processor.

The Navi Flounder is believed to offer 40 Compute Units, just as many as Navi 10, but the memory bus width is rumored to be limited to 192-bit.
This is not confirmed, of course, but it is a result of the simple multiplication of three variables present in the firmware data.
Assuming that each Compute Unit carries exactly 64 Stream Processors, the Navy Flounder would feature 2560 unified cores.

-> https://videocardz.com/newz/amd-navy-flounder-to-feature-40-cus-and-192-bit-memory-bus

 

Where is the fun if you dont get a bsod or a black screen from time to time... we are not mere peasants! lolz

 

Speaking of BSOD... Just got one from nvidia driver.

Granted, I'm on dev win10 track so those things can happen

 

Shh dont speak about that someone from nvidia section will come here and call you a liar and how its your mistake. Nvidia drivers have no issues.

 

AHAHAHAHAHAHAHAHAHAHAHAAHAHAHAHAAHAHA

AHAHAHAHAHAHAHAHAh

AHAHhahaha

hah...

 

It just works right?

 

DPC Watchdog?

 

Nah, it was my first BSOD in like 2 years at least. On Win10 20211 build, launched WoW and boom, insta crash with something nvsomething. Didn't pay much attention so just rebooted and back to normal life goes.

 

Being quite serious I have to say that in all the years of using both AMD and NVidia graphic cards, I personally had more problems with NVidia cards / drivers, like broken fan control, GSODs, Physx enabled games instability, DX12 games crashing like crazy just to name a few.. My close friend has RX 590 and plays Shadow of the Tomb Raider in DX12 no problems, my 2080ti on the other hand crashes after like 30 sec of play, DX11 work 100% stable tho, so it's not my PC like some NVidia fanboys were saying, besides Steam forums are flooded with people having the same problems with NVidia cards in many DX12 games.

So I can't wait for ''Big Navi'', I think we might have similar situation like when HD4000/5000 or HD 7000 series GPU's were released, if RX 6900 XT (or whatever it's name will be) will be around RTX 3080 performance with more VRAM and less than 350W TDP and hardware ray tracing it's a win I say. It's just the waiting that's killing me..
But at least it will come before Cyberpunk 2077 hahaha

P.S. Why nobody complains about huge hunger for power of RTX 3080/3090?
If that would be AMD doing it everybody would lose their sh** xD

 

^^ What can i say (Vega XTX here) -> Just make UV and then Play

Note:
Same goes for any GPU, 3080 When UV behaves a lot better.
For Me it doesn't matter the TDP, i want Performance then i will Tweak as always

 

AMD Dimgrey Cavefish in Mesa 20.03-devel

AMD today added a new GPU codename to its Linux OpenGL driver stack (Mesa 20.03-devel). The GPU called ‘Dimgrey Cavefish’ has the same GPU architecture level as Navi 21 and Navi 22.
The new Dimgrey is more than likely the codename of Navi 23.

The codename can be easily associated with Navi 23 because AMD has already provided the same hexadecimal ranges within other files (0x3C to 0x46).




-> https://videocardz.com/newz/amd-adds-dimgrey-cavefish-navi-23-to-linux-opengl-driver

 

^^ yeah.... now we know that we will have three main "branches" of RDNA2.

GFX1030 NAVI 21 Sienna Cichlid RDNA 2.0 RX 6900 TBC
GFX1030 NAVI 22 Navy Flounder RDNA 2.0 RX 6800/6700 TBC
GFX1030 NAVI 23 Dimgrey Cavefish RDNA 2.0 RX 6600/6500 TBC

I really hope AMD will release the RX 6900XTX with like 80 CUs and at least similar mem bandwidth as 3080 (doesn't matter what kind of special sauce they will apply like e.g. infinity cache). I really like to think they will be able to compete with NV this time and not only with price, but also power consumption and - most important for me - high end level performance (not necessarily the highest top tier, but at least second best).
Oh and btw 3080 10GB already obsolete - 20GB cards confirmed not only from Gigabyte but also Galax....
or to all the defenders of the 10GB -- you stay defending 3080 10GB in 1080p and mabby 1440p with textures set to "high" instead of ultra in new AAA titles (as it would make any sense to buy it for those resolutions)

 

i know teraflops arent the tell all for gaming performance but I crunched some numbers and at 5700xt and series x clocks a 80cu card will be theoretical 18tf and at ps5 clocks will be roughly around 23tf , definitely a beast of a card.......and with zen3 rumors saying intel should be scared and hardly any leaks on the gpu side I’m really excited for next month....

 

23TFLOPs of FP32 on RDNA1 with 80 CUs would be at 2.246GHz. For Big RDNA2 (80 CUs), performance will be Utilization game. If scheduling and extra cache enables CUs to be utilized more often than RDNA1 was, then it will be big Win for us.

 

how did you come up with those numbers? I’m not doubting I’m just trying to figure it out....

I took 5700xt 9.75 TFLOPs and divided it by the compute units then multiplied it by 80 for the 80 compute units for rdna1 performance

then I did the same for the ps5, I took 10.28 TFLOPs and divided by 36 cu’s then by 80 to get rdna2 clock rates and ipc increases....

it might not be right but it makes sense in my small mind...

 

Old good: SP_count * 2 * clock.

 

https://wccftech.com/nvidia-geforce-rtx-3080-users-report-crashes-black-screens-during-gaming/


But but.....it just works!

 

The cards have only been out 2 minutes. It could be anything from early drivers, under powered PSU, overclocked systems...

 

They mention that only custom designs seems to have such problems, my guess is that it could mean not stable enough power from PSU cuz those custom models use 2x 8pin while founders edition uses NVidia's custom solution (12pin), just my speculation..