Nvidia's RTX 40 series could be arriving sooner than expected

• Let's hope the miners won't be interested.

As 2021 draws to a close, we begin to look ahead to what’s brewing in 2022. Arguably the most anticipated PC hardware launch of 2022 is that of Nvidia’s next generation GeForce RTX 40 series graphics cards. We’re still many months away from launch, but new little tidbits of information are dripping out, indicating that development is well advanced. Factories are gearing up for production indicating that RTX 40 series cards could be on shelves sooner than anticipated. A report by Digitimes (paywalled) noted by RetiredEngineer via wccftech indicates that preparations for RTX 40 production are underway. Whether this means that the launch could come sooner than expected, or whether it’s a tactic to shore up supply of vital components is as yet unknown. Whatever the case, we hope that RTX 40 cards will be available in volume. Perhaps the release will coincide with Ethereum’s shift away from proof of work to proof of stake. Lower mining demand combined with ample supply should hopefully lead to more cards on shelves at lower prices, with more cards in the hands of gamers. Well, that’s the theory anyway. RTX 40 cards – if that is what they end up being called, will utilise the Ada Lovelace series of GPUs, replacing the Ampere based RTX 30 series. They’re expected to be built using the TSMC 5nm node. Ampere generation cards are built using Samsung’s 8nm process, so the shift to TSMC is expected to further add to the competition for TSMCs advanced nodes. No doubt Nvidia had to write a rather large cheque, with Apple, AMD, Qualcomm and others all relying on TSMC for their high end production. Ada Lovelace GPUs are expected to feature a monolithic design. Well known leaker Greymon55 claims that the high end AD102 GPU will feature up to 18432 CUDA cores, 24GB of 21 Gbps GDDR6X memory and a 2.3 to 2.5 GHz clock. This compares to the RTX 3090 with 10496 CUDA cores, 24GB of 19.5 Gbps GDDR6X and 1.4 to 1.7 GHz boost clock. This points towards the hypothetically named RTX 4090 being a very fast card indeed, though its likely to have a very steep power requirement too. The PCIe 5.0 power connector can deliver up to 600W. The days of 300W flagship cards appear to be well behind us.  

Let me add more: N31 =GFX11 =5nm+6nm TSMC =120WGP 15360SP ≈256bit 32G GDDR6 18Gbps? ≈256/512mb Infinity Cache? =3D Infinity Cache ≈2.4~2.5GHz? ≈75T FP32?

— Greymon55 (@greymon55) November 9, 2021

The Ada Lovelace generation takes its name from Augusta Ada King, the Countess of Lovelace, who was an English mathematician in the early 19th century. She worked on Charles Babbage’s Analytical Machine and is considered by many to be the first ever computer programmer. Nvidia’s RTX 40 series cards will go head to head with AMD’s RDNA 3 cards that could launch around the same time. They’ll be joined by Intel’s Alchemist series of GPUs which are launching much sooner in Q1 2022, though they are expected to compete against current generation cards. Intel’s second generation Battlemage GPUs are shaping up to compete with RTX 40 and RDNA 3. 2022 is looking like a very exciting year for graphics card buyers. (Source: PcGamer.com)