NVIDIA GeForce RTX 40 Series graphics cards featuring the Ada Lovelace GPU architecture are expected to deliver the same generational performance jump that we saw moving from the 9-series Maxwell GPUs to 10-series Pascal GPUs. The rumor comes from Ulysses, who has been talking about NVIDIA's next-gen parts for a while now.
NVIDIA GeForce RTX 40 GPUs With Ada Lovelace Architecture Expected To Deliver Same Generational Jump As We Saw With Maxwell To Pascal
The NVIDIA GeForce GTX 10 series graphics cards based on the Pascal GPU architecture were a huge performance jump compared to its Maxwell-based GeForce 9 series precessors. The 16nm chips delivered a major improvement to performance, efficiency, and overall value, also marking one of the biggest leaps in 'Ti' graphics performance. The GeForce GTX 1080 Ti is still regarded as the best 'Ti' graphics card ever made, a performance jump that NVIDIA has been unable to match with its Turing and Ampere flagships.
rtx30series—rtx40series like Maxwell—Pascal
— Red_Like_White_Roses (@Red_WhiteRoses) August 1, 2021
Now the NVIDIA GeForce RTX 40 series is expected to deliver the same generational performance jump over the GeForce RTX 30 series. Based on the Ada Lovelace GPU architecture, the GeForce 40 series graphics cards are expected to utilize TSMC's 5nm process node and while they will be very power-hungry, their efficiency numbers will go up tremendously thanks to the huge performance jump.
There are also some other details mentioned regarding clock speeds and launch timeframe. We know that the GeForce RTX 40 series is a long way from now and the rumor is that we won't expect these cards to launch until late Q4 2022. This is also due to the fact that NVIDIA will reportedly be offering an intermediate SUPER refresh of its GeForce RTX 30 series lineup in 2022. So if that lineup was to come before RTX 40 series, then we can expect the launch to slip further in Q1 2023. This means that AMD might just have its RDNA 3 lineup out by the time NVIDIA launches its new GPU family.
nope
2022 12or2023Q1 pic.twitter.com/CsTVP5YzFb— Red_Like_White_Roses (@Red_WhiteRoses) August 1, 2021
In terms of clock speeds, the Ada Lovelace-powered NVIDIA GeForce RTX 40 Series GPUs are said to offer clock speeds between 2.2 to 2.5 GHz (boost). This is a nice improvement to 1.7-1.9 GHz clocks that the Ampere architecture churns out currently on average. The Pascal architecture also clocked impressively and was the first GPU architecture to breach the 2.0 GHz clock speed limit however, it is AMD who has taken the clock speed throne with its RDNA 2 GPU architecture which can hit clocks beyond 2.5 GHz with ease.
Based on these numbers, if the RTX 4090 or whatever the AD102 GPU is featured inside features 18,432 CUDA cores, then we are getting up to 80 TFLOPs of FP32 compute performance at 2.2 GHz which is insane and over 2x the single-precision floating-point jump over RTX 3090. These numbers do align with the rumors that we can expect up to a 2.5x performance jump with the Ada Lovelace-GPUs-based NVIDIA GeForce RTX 40 graphics cards.
NVIDIA CUDA GPU (RUMORED) Preliminary:
| GPU | TU102 | GA102 | AD102 |
|---|---|---|---|
| Flagship SKU | RTX 2080 Ti | RTX 3090 Ti | RTX 4090? |
| Architecture | Turing | Ampere | Ada Lovelace |
| Process | TSMC 12nm NFF | Samsung 8nm | TSMC 4N? |
| Die Size | 754mm2 | 628mm2 | ~600mm2 |
| Graphics Processing Clusters (GPC) | 6 | 7 | 12 |
| Texture Processing Clusters (TPC) | 36 | 42 | 72 |
| Streaming Multiprocessors (SM) | 72 | 84 | 144 |
| CUDA Cores | 4608 | 10752 | 18432 |
| L2 Cache | 6 MB | 6 MB | 96 MB |
| Theoretical TFLOPs | 16 TFLOPs | 40 TFLOPs | ~90 TFLOPs? |
| Memory Type | GDDR6 | GDDR6X | GDDR6X |
| Memory Capacity | 11 GB (2080 Ti) | 24 GB (3090 Ti) | 24 GB (4090?) |
| Memory Speed | 14 Gbps | 21 Gbps | 24 Gbps? |
| Memory Bandwidth | 616 GB/s | 1.008 GB/s | 1152 GB/s? |
| Memory Bus | 384-bit | 384-bit | 384-bit |
| PCIe Interface | PCIe Gen 3.0 | PCIe Gen 4.0 | PCIe Gen 4.0 |
| TGP | 250W | 350W | 600W? |
| Release | Sep. 2018 | Sept. 20 | 2H 2022 (TBC) |
But the biggest question remains, will the NVIDIA GeForce RTX 40 Series graphics cards with Ada Lovelace GPUs bring back the same pricing as Pascal? Not only was the GTX 1080 Ti the best card for its performance, but it also offered the best value of any 'Ti' graphics card which is why we all loved the Pascal family.
The $699 US pricing made the GTX 1080 Ti an amazing value but not only that, but NVIDIA also dropped pricing of its standard lineup to $499 (GTX 1080) and $349 for the GTX 1070 which we definitely should be going back to. However, the rise in component & logistics costs may not bring back the same level of pricing again.
NVIDIA GeForce GPU Segment/Tier Prices
| Graphics Segment | 2023-2024 | 2022-2023 | 2021-2022 | 2020-2021 | 2019-2020 | 2018-2019 | 2017-2018 | 2016-2017 | 2014-2016 |
|---|---|---|---|---|---|---|---|---|---|
| Titan Tier | GeForce RTX 4090 | GeForce RTX 4090 | GeForce RTX 3090 Ti GeForce RTX 3090 | GeForce RTX 3090 | Titan RTX (Turing) | Titan V (Volta) | Titan Xp (Pascal) | Titan X (Pascal) | Titan X (Maxwell) |
| Price | $1599 US | $1599 US | $1999 US $1499 US | $1499 US | $2499 US | $2999 US | $1199 US | $1199 US | $999 US |
| Ultra Enthusiast Tier | GeForce RTX 4080 SUPER | GeForce RTX 4080 | GeForce RTX 3080 Ti | GeForce RTX 3080 Ti | GeForce RTX 2080 Ti | GeForce RTX 2080 Ti | GeForce GTX 1080 Ti | GeForce GTX 980 Ti | GeForce GTX 980 Ti |
| Price | $999 US | $1199 US | $1199 US | $1199 US | $999 US | $999 US | $699 US | $649 US | $649 US |
| Enthusiast Tier | GeForce RTX 4070 Ti SUPER | GeForce RTX 4070 Ti | GeForce RTX 3080 12 GB | GeForce RTX 3080 10 GB | GeForce RTX 2080 SUPER | GeForce RTX 2080 | GeForce GTX 1080 | GeForce GTX 1080 | GeForce GTX 980 |
| Price | $799 US | $799 US | $799 US | $699 US | $699 US | $699 US | $549 US | $549 US | $549 US |
| High-End Tier | GeForce RTX 4070 SUPER GeForce RTX 4070 | GeForce RTX 4070 GeForce RTX 4060 Ti 16 GB | GeForce RTX 3070 Ti GeForce RTX 3070 | GeForce RTX 3070 Ti GeForce RTX 3070 | GeForce RTX 2070 SUPER | GeForce RTX 2070 | GeForce GTX 1070 | GeForce GTX 1070 | GeForce GTX 970 |
| Price | $599 $549 | $599 US $499 US | $599 $499 | $599 $499 | $499 US | $499 US | $379 US | $379 US | $329 US |
| Mainstream Tier | GeForce RTX 4060 Ti GeForce RTX 4060 | GeForce RTX 4060 Ti GeForce RTX 4060 | GeForce RTX 3060 Ti GeForce RTX 3060 12 GB | GeForce RTX 3060 Ti GeForce RTX 3060 12 GB | GeForce RTX 2060 SUPER GeForce RTX 2060 GeForce GTX 1660 Ti GeForce GTX 1660 SUPER GeForce GTX 1660 | GeForce GTX 1060 | GeForce GTX 1060 | GeForce GTX 1060 | GeForce GTX 960 |
| Price | $449 $299 | $399 US $299 US | $399 US $329 US | $399 US $329 US | $399 US $349 US $279 US $229 US $219 US | $249 US | $249 US | $249 US | $199 US |
| Entry Tier | RTX 3050 8 GB RTX 3050 6 GB | RTX 3050 | RTX 3050 | GTX 1650 SUPER GTX 1650 | GTX 1650 SUPER GTX 1650 | GTX 1050 Ti GTX 1050 | GTX 1050 Ti GTX 1050 | GTX 950 | GTX 750 Ti GTX 750 |
| Price | $229 $179 | $249 US | $249 US | $159 US $149 US | $159 US $149 US | $139 US $109 US | $139 US $109 US | $149 US | $149 US $119 US |
NVIDIA and AMD have really been moving the price bar up for their high-end cards for a while now and unless there's some really heated GPU competition between both with enough supply in the channels, we don't expect 'Ti' variants to fall back down the $1000 US pricing segment.
