Nvidia’s next-gen GPUs — and, it turns out, AMD — have a problem: pricing, or rather relative pricing. The problem is that with the Radeon RX 6800 XT and more GeForce RTX 3080, our GPUs strike an excellent balance between price and performance. Both have been pared down compared to the flagship models, but each retains most of their gaming prowess. For whatever reason, their recently released counterparts – the RTX 4080 and RX 7900 XT – can’t match the same sweet spot in the “dollar frame”.The new RTX 4070 Ti addresses this somewhat, but still emphasizes the extra performance with Big price cuts are a thing of the past.
That said, the rebranded RTX 4080 12GB does, to some extent, finally solve the RTX 3080’s value challenge. The RTX 3080 has a confirmed MSRP of $799/£799 – $100/£150 – and you’ll get value for your extra cash. The card’s performance is generally in the RTX 3090 to RTX 3090 Ti range. You get an additional 2 GB of framebuffer memory. On top of that, there’s DLSS 3, which hasn’t yet established itself as a game-changer for DLSS 2, but is quickly becoming a very useful-game-changer like The Witcher 3: Complete Edition.
However, the MSRP of $799 (down from $899 for the “unreleased” RTX 4080 12GB) depends on third-party partner designs reaching Nvidia’s suggested price point. There are no Founders Edition cards this time, which is a bit disappointing. Third-party cards seem to be dominated by unnecessarily bulky designs, and this extends to the Gigabyte OC model we sent for review from Nvidia. For a card that delivers 50% to 55% of the performance of the RTX 4090 under extreme RT workloads, we have a card that is much larger than the flagship’s already bulky design. We just wish there was a form factor similar to the previous-gen RTX 3070 Ti, which handles similar power throughput in a more compact design.
| Graphics card 4090 24GB | Graphics card 4080 16GB | RTX 4070 Ti 12GB | |
|---|---|---|---|
| processor | AD102 | AD103 | AD104 |
| transistor | 76.3B | 45.9B | 35.8B |
| Die size | 608 square millimeters | 379 square millimeters | 295 square millimeters |
| CUDA color | 16384 | 9728 | 7680 |
| boost clock | 2.52GHz | 2.51GHz | 2.61GHz |
| memory interface | 384 bits | 256 bits | 192 bits |
| memory bandwidth | 1018GB/sec | 742GB/sec | 557GB/sec |
| Three Gorges Project | 450W | 320W | 285W |
| power supply recommendation | 850W | 750W | 700W |
| power cable | 4 x 8-pin | 3 x 8-pin | 2 x 8-pin |
| basic price | $1499/£1649 | $1199/£1199 | $799/£799 |
| Release date | October 12, 2022 | November 16, 2022 | January 5, 2023 |
The table above shows how the RTX 4070 Ti fits into the lineup, and you may have noticed that we’re seeing a very different processor than the more expensive ones. The RTX 3080 shares the same chip as the RTX 3080 Ti, 3090, and 3090 Ti, but this time we’re looking at a more diverse chip lineup. The RTX 4070 Ti uses AD104, which offers about 47% of the compute silicon of the RTX 4090 (albeit at slightly faster clock speeds), half the memory bandwidth, and half the framebuffer memory. The AD103 processor used in the RTX 4080 is like a halfway house, plugged in accordingly in terms of specs and performance.
These specs raise some concerns, as mapping equivalent silicon comparisons to the RTX 3000 lineup could put your product more in line with an RTX 3060 Ti than a 70-class GPU. While this is a valid point from this perspective, it doesn’t take into account the architectural improvements offered by the Ada Lovelace architecture – a design necessity given the much higher cost of TSMC’s latest process node. As it stands, the RTX 3060 Ti effectively replaces the 2080 Super, while the RTX 3070 matches the 2080 Ti. This time, the 4070 Ti is more in line with the RTX 3090 and 3090 Ti.
You might be wondering what the “OC” element of the Gigabyte model adds to the equation, but it turns out it’s only +30MHz of boost clock – insignificant for a card that typically runs around 2.8GHz. I’d prefer something that offers a reference clock, a reference TGP, and a form factor similar to the RTX 3070 Ti or RTX 3080 Founders Edition models, so hopefully we’ll see those in the launch lineup as well.
In fact, the card offers few surprises other than its massive form factor. Three DisplayPort 1.4a outputs are lined up with HDMI 2.1 as standard – a perfect lineup despite the exclusion of DisplayPort 2.1 – while power is provided by the same 12VHPWR socket used on the RTX 4080 and RTX 4090. The difference here is that the power requirements are much lower, requiring only two eight-pin PCIe cables connected to the GPU via a two-socket dongle. While realizing that at some point a new power connector will be needed, I’d prefer a standard PCIe power input for a card of this class.
While there’s been a lot of controversy surrounding Nvidia’s latest line of GPUs, one of the design elements that’s getting a lot of praise is the Ada Lovelace architecture and its power efficiency. To test this with the RTX 4070 Ti, I repeated the gaming benchmarks I used in my RX 7900 XTX and RX 7900 XT reviews.
In Control and Dying Light 2 running in RT mode, and using Forza Horizon 5 and Hitman 3 for rasterization tests, we measured performance and power consumption. By dividing the watts by the frame rate, we can calculate the number of joules per frame – the lower the number, the better. There are many ways to measure efficiency. Some prefer to measure total system power from the wall (which is perfectly valid), while others measure power consumption by a set frame rate limit. We prefer our technique because it ensures the GPU is running at full capacity – something that frame rate throttling can’t do.
In the table below, you’ll see how the RTX 4070 Ti stacks up against the RTX 3080 it effectively replaces, while you’ll also see how it compares to AMD’s RX 7900 XT and the more expensive, higher-performing RTX 4080. Generally, the lower you go down the GPU stack, the more efficient you’ll see. However, this is not the case here. While we’re still seeing big power improvements over the RTX 3080 and even the RX 7900 XT, all of our measurements point to the more expensive RTX 4080 being more efficient — likely because the newer card has a higher boost. Press the clock.
| Radeon RX 7900 XT | GeForce RTX 3080 | GeForce RTX 4070 Ti | GeForce RTX 4080 | |
|---|---|---|---|---|
| Control, 4K, High RT | 305W/30fps – 10.2 joules per frame | 315W/31fps – 10.2 joules per frame | 259W/34fps – 7.62 joules per frame | 302W/44fps – 6.86 joules per frame |
| Dying Light 2, 4K, Ultra RT | 306W/30fps – 9.0 joules per frame | 311W/33fps – 9.42 joules per frame | 261W/39fps – 6.69 joules per frame | 301W/50fps – 6.02 joules per frame |
| Forza Horizon 5, 4K, Extreme, RT Off | 302W/100fps – 3.02 joules per frame | 296W/79fps – 3.74 joules per frame | 203W/100fps – 2.03 joules per frame | 224W/121fps – 1.85 joules per frame |
| Hitman 3, 4K, Max, Realtime Off | 309W/175fps – 1.77 joules per frame | 314W/109fps – 2.88 joules per frame | 269W/135fps – 1.99 joules per frame | 301W/181fps – 1.66 joules per frame |
Control is an interesting case because it’s a bit of a “banana peel” game with scalability issues on the Ada Lovelace architecture. The faster the card, the better performance – relatively speaking – you’ll get. Even so, the new offering still outperforms the RTX 3080 at 25% lower power consumption per frame – but interestingly, the RTX 4080 has a 10% advantage over the RTX 4070 Ti judging by the same metrics.
The RTX 4080 was always more efficient in all our tests. In Dying Light 2, the RTX 4070 Ti consumes 29% less power per frame than the RTX 3080. But again, the 4080 has about a 10% per-frame power advantage over its smaller sibling. On both RT tests, the RX 7900 XT was more comparable to Nvidia’s previous-generation architecture—perhaps to be expected, bearing in mind how AMD effectively delivered RT performance in line with older GeForce cards.
Looking at our rasterization tests, Forza Horizon 5 continues to impress, as it uses far less power than the Ada Lovelace card’s TGP limit, producing amazing efficiency. Yes, the RTX 4080 still leads the RTX 4070 Ti by about 10 percentage points, but the new card cuts the power requirement per frame by 46% compared to the RTX 3080. The 4070 Ti needed only 67% more power to generate the comparison to the RX 7900 XT – which might be a bit odd, as Forza Horizon 5 didn’t perform particularly well on the RDNA 3 architecture.
That wasn’t the case with Hitman 3, RDNA 3 was very fast and was the only win over the RTX 4070 Ti. Nvidia’s latest offerings require 12% more power per frame, but neither of these latest offerings touches the RTX 4080. Even so, the RTX 4070 Ti still uses 31% less power per frame compared to the RTX 3080.
As you probably already know from these seven pages of reviews, we have a lot of benchmarks to share, but before we get started, here’s a brief overview of the new PC we built to help test the new generation of GPUs. At the heart of our system is Intel Core i9 12900K, running at 5.2GHz P core and 3.9GHz E core. The CPU is augmented with 6000MT/s Trident Z5 RGB DDR5 memory from G-Skill. The Noctua NH-D15 is our go-to cooler for powerful processors while all games are run from the PCIe Gen 3 NVMe SSD. All components are mounted on an ASUS ROG Maximus Hero Z690 motherboard, while the power supply is powered by a Corsair RM1000i power supply.
Nvidia GeForce RTX 4070 Ti Analysis
- Introduction, Hardware and Power Analysis [This Page]
- RT Benchmarks: Dying Light 2, Cyberpunk 2077, Control, F1 22
- RT Benchmarks: Hitman 3, Metro Exodus Enhanced, Marvel’s Spider-Man Remastered
- RT/DLSS vs. FSR2 Benchmarks: Cyberpunk 2077, Dying Light 2, Marvel’s Spider-Man Remastered
- Gaming Benchmarks: Control, Cyberpunk 2077, F1 22
- Gaming Benchmarks: Hitman 3, Forza Horizon, Red Dead Redemption 2, Shadow of the Tomb Raider
- Nvidia GeForce RTX 4070 Ti: Digital Foundry Verdict
