Monday, December 20th 2021

ASUS TUF Gaming GeForce RTX 3090 Ti Box Pictured
It looks like GeForce RTX 3090 Ti is indeed the name of the maxed out "Ampere" GA102 silicon, and NVIDIA did not go with "RTX 3090 SUPER" for its naming. A picture emerged of an ASUS TUF Gaming GeForce RTX 3090 Ti graphics card that confirms the naming. The board design of the card looks similar to the RTX 3090 TUF Gaming, except that the Axial-Tech fans have changed, with more blades on the impellers.
The GeForce RTX 3090 Ti is expected to max out the GA102 silicon, featuring all 10,752 CUDA cores, 84 RT cores, and 336 Tensor cores, physically present on the silicon. The memory size is unchanged over the RTX 3090, with 24 GB of GDDR6X memory. What's new is that NVIDIA is reportedly using faster 21 Gbps-rated memory chips, compared to 19.5 Gbps on the RTX 3090. The typical board power is rated at 450 W, compared to 350 W on the RTX 3090. NVIDIA is expected to announce this card at its January 4 press event along the sidelines of the 2022 International CES.
Sources:
HXL (Twitter), VideoCardz, ITHome
The GeForce RTX 3090 Ti is expected to max out the GA102 silicon, featuring all 10,752 CUDA cores, 84 RT cores, and 336 Tensor cores, physically present on the silicon. The memory size is unchanged over the RTX 3090, with 24 GB of GDDR6X memory. What's new is that NVIDIA is reportedly using faster 21 Gbps-rated memory chips, compared to 19.5 Gbps on the RTX 3090. The typical board power is rated at 450 W, compared to 350 W on the RTX 3090. NVIDIA is expected to announce this card at its January 4 press event along the sidelines of the 2022 International CES.
63 Comments on ASUS TUF Gaming GeForce RTX 3090 Ti Box Pictured
RTX 3080 10 GB, and RTX 3090 24 GB is absolute nonsense.
In the worst case, it should have been:
RTX 3080 12 GB and RTX 3090 16 GB.
or RTX 3080 16 GB, and RTX 3090 20 GB..
The RTX 3080 especially would have been pretty sweet with 15 GB VRAM (10 chips x 1.5 GB per chip).
Halo's menu seems to be like New World, a GPU toaster.
every memory chip is bound to a 32bit bus.
you have 8Gbit chips (1GB)
with a 320 Bit memory bus you can use 10 or 20 memory chips. and nothing in between (without making the card a complete mess and barely functional as soon as the "odd" memory gets used. like with the 970.
It has only 3.5 for another reason.
And yes, Nvidia has already designed other models with decoupled MI bandwidth-VRAM capacity ratios :D
And first - you have to design the MI interface with the memory capacity in mind, not the other way round.
If you want 16 GB, then give the card a 256-bit MI.
RTX 3080 VRAM usage warnings and the issue with VRAM pool sizes: the compromise of 4K gaming | ResetEra
Far Cry 6 needs more VRAM than the Nvidia RTX 3080 has to load HD textures | PCGamesN
To each their own, it was easy to predict this.
So no, Nvidia won't do the hardware workaround for 'sufficient' VRAM anymore, after they got burned on Maxwell they're designing around it and they sell you the argument that all is well with 50% reduced VRAM compared to Pascal in relative core power. Meanwhile, they still do release double VRAM for the entire stack. Early adopter heaven, cash twice on fools with money. Its much better than getting forced to settle with customers for your 3.5GB.
The newest argument to avoid buyers' remorse now is 'muh muh but we have DLSS, so it looks great anyway'. :roll::roll::roll: Okay, enjoy being on Nvidia's DLSS leash for your gaming, what used to be waiting for your SLI profiles now is waiting for your DLSS profiles. But! This time, it will all be different, right? :D
Its a piss poor gen, this one. Even the 450W on this 3090ti underlines it - they can't make a full die to save their life without pulling all the stops on voltage. Samsung 8nm. Fan-tas-tic node.