Wednesday, November 20th 2019
NVIDIA Readying GeForce RTX 2080 Ti SUPER After All?
NVIDIA could launch a "GeForce RTX 2080 Ti Super" after all, if a tweet from kopite7kimi, an enthusiast with a fairly high hit-rate with NVIDIA rumors is to be believed. The purported SKU could be faster than the RTX 2080 Ti, and yet be somehow differentiated from the TITAN RTX. For starters, NVIDIA could enable all 4,608 CUDA cores, 576 tensor cores, and 72 RT cores, along with 288 TMUs and 96 ROPs. Compared to the current RTX 2080 Ti, the Super could get faster 16 Gbps GDDR6 memory.
It's possible that NVIDIA won't change the 352-bit memory bus width or 11 GB memory amount, as those would be the only things stopping the card from cannibalizing the TITAN RTX, which has the chip's full 384-bit memory bus width, and 24 GB of memory. Interestingly, at 16 Gbps with a 352-bit memory bus width, the RTX 2080 Ti Super would have 704 GB/s of memory bandwidth, which is higher than the 672 GB/s of the TITAN RTX, with its 14 Gbps memory clock. These design choices would ensure NVIDIA has a sufficiently faster product than the RTX 2080 Ti, without an increase in BOM, provided it has enough perfectly-functional "TU102" inventory to go around. There's no word on availability, although WCCFTech predicts a CES 2020 unveiling.
Sources:
kopite7kimi (Twitter), WCCFTech
It's possible that NVIDIA won't change the 352-bit memory bus width or 11 GB memory amount, as those would be the only things stopping the card from cannibalizing the TITAN RTX, which has the chip's full 384-bit memory bus width, and 24 GB of memory. Interestingly, at 16 Gbps with a 352-bit memory bus width, the RTX 2080 Ti Super would have 704 GB/s of memory bandwidth, which is higher than the 672 GB/s of the TITAN RTX, with its 14 Gbps memory clock. These design choices would ensure NVIDIA has a sufficiently faster product than the RTX 2080 Ti, without an increase in BOM, provided it has enough perfectly-functional "TU102" inventory to go around. There's no word on availability, although WCCFTech predicts a CES 2020 unveiling.
139 Comments on NVIDIA Readying GeForce RTX 2080 Ti SUPER After All?
The game may be modding friendly, but the underlying game scales and performs terribly. The Titan cards are not the bottleneck.
Nvidia's pro dominance has only *increased* with the lack of any convincing answer to Tensor Cores, Ray-Tracing Cores, and OpenCL still being abysmal to leverage for pro & high-end use cases like CUDA can.
That trickles down to the dominant position Nvidia has in the gaming front; more than anything Nvidia gains a lot with the existing RTX series now that console generation is finally being caught up to leveraging deep-learning & ray-tracing in games via the Xbox Series X & PS5.
Your position just sounds either-or to ignore.That doesn't make sense as the previous generations didn't have a Super product in between them and the fact that the latter 2 cards are targeting optimal rendering at 4x more pixels (4K) + ray-tracing + deep-learning.
For most pros & enthuaissts, It was obviously going to be obviously take longer releases between substantially different new GPUs from Nvidia as they're targeting MUCH harder outputs to markedly improve from. And it's going to cost some dough that prosumers can easily afford. Gamers not so much until economies of scale occur that are finally going to happen w/ this new card further + next gen of console gaming finally catching up.A new design doesn't mean it's going to be any close to the competitor's existing & past designs. AMD knows this too well from what Nvidia's Maxwell and up cards (1080TI+) have done to them vs. their newer cards.
It's gotten so bad, we all as GPU enthusiasts must hope Intel as a new alternative does a better job giving options and competition to the market with their new GPUs.
At minimum considering their current release of GPUs was without a doubt a disappointment, AMD has earned skepticism if Navi 2 is even going to actually compete w/ Nvidia RTX 1 flagship cards.
Heck, Nvidia current GPUs, thanks to their emphasis properly on deep-learning & ray-tracing supports all tiers of Variable-rate shading (VRS); AMD's card do not. AMD is touting support for that & ray-tracing with again warranted skepticism that it'll be anything close to Nvidia RTX 1 cards (next-gen console will fortunately at least have support for 'em to not hold back gaming further).
I'm not expecting miracles from RDNA2, but I hope they'll be competitive enough to stop Nvidia from pricing cards at $700+.
Simply shrinking Turing to 7nm is already a tall order, but Nvidia may decide to throw some new stuff into the ring again. Probably not so soon after RTX/DXR and VRS, but who knows?
Wait for Ampere. Buying a 1.5 year old architecture is almost pointless now.
I have a feeling nVidia is really going to dedicate a lot more die space to RTX... essentially making the current RTX and RTX Super cards look laughably slow in ray-tracing.
After playing Metro Exodus, ray-traced global illumination is the real deal and worth the performance hit.