AMD and NVIDIA’s respective launch of their latest graphics cards has shaken up the PC gaming market, with the same question being asked: NVIDIA or AMD? Today, we’re comparing two high-end cards from the two teams and deciding the verdict of Radeon RX 7900 XT vs GeForce RTX 4080.
- The AMD RX 7900 XT features a multi-die design, while the GeForce RTX 4080 is based on a monolithic design.
- NVIDIA GeForce RTX 4080 has less, but faster, memory than the Radeon RX 7900 XT.
- NVIDIA’s GeForce RTX 4080 costs at least $300 (33%) more than the Radeon RX 7900 XT.
- The GeForce RTX 4080 is about 11.5% faster than AMD’s RX 7900 XT in 4K gaming.
- The GeForce RTX 4080 runs cooler than the RX 7900 XT in gaming but draws slightly more power.
|GPU||GeForce RTX 4080||Radeon RX 7900 XT|
|Die||AD 103||Navi 31|
|Architecture||Ada Lovelace||RDNA 3|
|Processing Node||TSMC 5nm (4N)||TSMC 5nm (GCD)
TSMC 6nm (GCD)
|Die Size||379 mm²||520 mm²|
|Memory Size (Type)||16 GB (GDDR6X)||20 GB (GDDR6)|
|Bus Width||256 bit||320 bit|
|Base Clock||2205 MHz||1500 MHz|
|Boost Clock||2505 MHz||2394 MHz|
|Ray Tracing Cores||76||84|
|TDP||320 Watts||300 Watts|
|Launch Date||Nov 16th, 2022||Dec 13th, 2022|
|Best Variants||Best RTX 4080 Cards|
There are some significant differences between the cards, which will be discussed below.
Also Read: RTX 4080 Vs RTX 3080
The first significant difference is that the AMD Radeon RX 7900 XT uses a dual-die design, with a separate Graphics Compute Die and a Memory Cache Die. NVIDIA’s Lovelace-based RTX 4080 still uses the same monolithic design with some improvements.
The Radeon RX 7900 XT uses a 5nm node for the Graphics Compute Die and a 6nm node for the Memory Cache. The GeForce RTX 4080 uses a specialized 5nm node called 4N from TSMC.
AMD’s RX 7900 XT uses dual-issue stream multiprocessors. One SIMD lane can now execute two instructions per cycle, or they can be executed in parallel. The GeForce RTX 4080 does not have this dual-issue capability and can only execute a single instruction in one cycle.
AMD’s Radeon RX 7900 XT houses the new DisplayPort 2.1, which enables bandwidth of up to 54Gbps and resolutions of up to 4K at 480 Hz. NVIDIA’s GeForce RTX 4080 houses DisplayPort 1.4a, which limits the bandwidth to 32.4Gbps.
NVIDIA DLSS 3.0 Vs AMD FSR 3.0
AMD and NVIDIA’s upscaling technologies have been very similar since AMD moved to temporal upscaling in FSR 2.0. Both companies announced the third major iteration of their respective upscaling technologies along with the launch of their cards.
NVIDIA’s DLSS 3.0 will use new optical flow accelerators to generate entirely new frames instead of just pixels. AMD is going down a similar path, employing its fluid motion frame technology to generate new frames for smooth gameplay. AMD, however, has not revealed any other detail about this new iteration of FSR.
DLSS 3.0 will rely on NVIDIA’s fourth-generation Tensor Cores and their improved optical flow accelerators, which means that DLSS 3.0 support will be limited to the RTX 4000 series cards. On the other hand, AMD FSR 3.0 support will likely be added to older AMD as well as NVIDIA cards.
Read More: NVIDIA DLSS Vs AMD FSR
AMD Radeon RX 7000 Series
AMD’s RX 7000 series launched almost two months after NVIDIA launched their flagship cards, the GeForce RTX 4090 and the RTX 4080. The arrival of these RDNA 3-based cards is in due time to compete with their NVIDIA counterparts.
AMD has, thus far, launched two graphics cards from this lineup: The Radeon RX 7900 XT and the 7900 XTX. Both of these feature the same core architecture but a difference in specifications.
First, let’s discuss this new architecture.
Also Read: RX 7900 XTX Vs Radeon RX 6950 XT
As mentioned, these new cards are based on AMD’s much-awaited RDNA 3 architecture. These cards feature a chiplet design similar to AMD’s Ryzen 3000 CPUs. This makes these cards the first to feature a chiplet design.
Here, AMD has separated the Graphics Compute Die (which contains the shaders) and the Memory Cache Die (which has AMD’s second-generation infinity cache and the 64-bit GDDR6 memory controllers).
Graphics Compute Die
The GCD is based on TSMC’s 5nm node, while the MCD has a 6nm node, also from TSMC. This enabled AMD to achieve 61 TFLOPs of compute performance and a chiplet interconnect speed of 5.3 TB/s. AMD has advertised a 54% improvement in performance/watt thanks to the new design.
The Graphics Compute Die contains Graphics Compute Units which contain the main component: the shaders, as well as “ray tracing accelerators” and “AI accelerators.” The graphics compute unit has an advertised 165% higher transistor density than RDNA 2.
The new stream processors are “dual-issue,” meaning they each have two instructions through which different integers or floating point operations can move according to the workloads. This leads to much more shader efficiency and, consequently, better performance.
On the other hand, the new, dedicated AI accelerators have new instructions that are “unified” with the compute units (2 per compute unit). These deliver 2.7 times better AI performance.
The ray tracing accelerators have 1.5 times more rays in flight than the previous generation and also hold new instructions. These offer an advertised 50% better performance per compute unit.
Unlike NVIDIA, AMD’s Radeon RX 7900 XT and Radeon RX 7900 XTX come with DisplayPort 2.1 support. With a bandwidth of up to 54Gbps, these support up to 8K at 165 Hz or a whopping 4K at 480 Hz. They also enable 12-bit color depth for up to 68 billion colors.
Monitors with DisplayPort 2.1 will start arriving in early 2023 from major manufacturers such as Samsung, Asus, and Dell.
New Dual-Media Engine
Also present in the GCD is the new dual-media engine with 8K60FPS AV1 encode/decode and simultaneous encode/decode for AVC/HEVC. AI enhances the video encoding capabilities, and there’s also a 1.8 times increase in the engine’s frequency. Overall, this enables up to 7 times faster transcoding with AV1 encoding. These massive improvements in streaming capabilities are colossal news for streamers and AMD fans alike. Also, make sure to check out our comparison of RTX 4080 vs 6950 XT.
One term that has been going around since AMD’s RX 7000 series launch is “decoupled clocks,” which is the separation of front-end frequency from the shader frequency. Engineers at AMD found that the front-end frequency was proving to be a limiting factor in performance in many workloads, thus leading to this change.
The front-end frequency is increased to 2.5 GHz (a 15% improvement and the shader frequency is now at 2.3 GHz (which led to 25% better power efficiency).
Improvement in Compute Performance
All these changes enabled compute performance of up to 61 TFLOPs, a massive improvement over RDNA 2’s figure of 23 TFLOPs.
AMD introduced us to FidelityFX Super Resolution 3.0 during their advance_gaming event in November 2022. While we did not get much detail to work with, we learned that AMD would employ a new “fluid motion frame” technology to be combined with the Super Resolution Technology. AMD claims a two-times increase in 4K gaming performance over FSR 2.0.
FSR 3.0 will be officially released and ready for use in January 2023.
NVIDIA RTX 4000 Series
On to the RTX 4000 series, where NVIDIA has launched two graphics cards: GeForce RTX 4080 16 GB, and the GeForce RTX 4090 24 GB. NVIDIA has not given a whole lot of detail about these cards, but we’ll discuss the details we do have. You may also be interested to learn about our comparison of the GeForce RTX 4090 with the RTX 3080 Ti.
These cards are based on the new “Ada Lovelace” architecture. NVIDIA is using a specialized “4N” node from TSMC for these chips, which is a modified 5nm node and not a 4nm one. This allows NVIDIA to pack in 76 billion transistors which are 70% more than the 8 nm-based RTX 3000 cards. Furthermore, NVIDIA is using the same monolithic die as before. These are the two significant factors distinguishing NVIDIA and AMD’s latest cards.
NVIDIA claims to have improved all three core processors on this chip: Shaders, Ray Tracing Cores, and Tensor Cores.
First up, the Shaders or Streaming Multiprocessors with a compute performance of 90 TFLOPs, more than twice as much as the previous generation’s Ampere. At the same time, they are offering the same performance as the previous generation at half the power.
The big improvement in these new SMs is Shader Execution Reordering, which reschedules GPU threads “on the fly” to similar tasks, reducing execution divergence. SER also allows the decoupling of ray intersection and shading. This fixes the central problem of data divergence and execution divergence in ray tracing caused by poor shader execution.
Also Read: RTX 4080 Vs RTX 3090
Third Generation Ray Tracing Cores
The third generation of Ray Tracing cores clock up to 200 TFLOPs in ray tracing and perform twice faster in calculating ray triangle intersection. These cores also contain two new components to improve performance: an Opacity Micromap (OMM) Engine and a Displayed Micro-mesh (DMM) Engine. The OMM Engine speeds up alpha-test geometry, which helps render complex geometry, by two times. In real-time, the DMM Engine enhances the geometric detail of items like rocks, trees, and creatures at much faster speeds.
Finally, NVIDIA introduced new fourth-generation Tensor Cores to accelerate deep learning processes, employing a new FP8 Transformer Engine. The Tensor Cores increase AI inference performance by 5 times over Ampere.
Nvidia also announced DLSS 3.0 during their GTC September 2022 Keynote event. Powered by their new Tensor Cores, DLSS 3.0 produces completely new frames. NVIDIA introduced a new “Optical Flow Accelerator,” which feeds information from the current frame and the frame before that to the neural network. The neural network then generates an entirely new, high-quality frame. This process does not rely on the game, and NVIDIA claims a performance boost of over four times over the previous iteration of DLSS.
RX 7900 XT Vs RTX 4080 Gaming Benchmarks
Now, let’s compare the gaming performance of the two graphics cards, thanks to data provided by Testing Games. A look at the test bench:
- CPU: Intel Core i9-13900K
- Motherboard: ASUS ROG Strix Z790-E Gaming
- Memory: G.SKILL Trident Z5 RGB 32GB (2×16) DDR5 6000MHz CL38
- CPU Cooler: be quiet! Dark Rock Pro 4
- Storage: Samsung 970 EVO M.2 2280 1TB
- Power Supply: Corsair RM850i 850W
- Operating System: Windows 10 Pro
The two graphics cards will be compared across two games, in average FPS and the %1 low. All the games are tested at 4K resolution at ultra or equivalent settings. So without any further due, let’s dive into Radeon RX 7900 XT vs GeForce RTX 4080 gaming benchmarks.
- The GeForce RTX 4080 comes out on top, averaging 61 FPS, while the Radeon RX 7900 XT trails with 54 FPS.
- In terms of %1 lows, the GeForce RTX 4080 stands at 54 frames, and the Radeon 7900 XT falls behind with 47 frames.
Forza Horizon 5
- The GeForce RTX 4080 takes a big lead of 21% in this title, with 147 FPS on average. The Radeon RX 7900 XT scores 121 FPS in reply.
- Both the graphics cards have the same gap between average FPS and %1 lows, so the Radeon 7900 XT also falls behind in this regard.
God of War
- The GeForce RTX 4080 manages 96 FPS on average, and the Radeon RX 7900 XT trails by nearly 13% at 85 FPS.
- In terms of %1 lows, the Radeon RX 7900 XT has 72 FPS, while the RTX 4080 has 82 FPS.
- The Radeon 7900 XT takes its first victory in Hitman 3, beating the GeForce RTX 4080 by a measly 1 FPS.
- AMD Radeon 7900 XT takes a slightly bigger win in %1 low with 121 FPS, compared to GeForce RTX 4080’s 115.
Horizon Zero Dawn
- The GeForce RTX 4080 grabs another big win with 118 FPS on average. The Radeon 7900 XT only manages 99 FPS, losing by 19%.
- The Geforce RTX 4080 recorded 89 FPS in terms of the %1 low, and the Radeon RX 7900 XT is not far behind, at 80 FPS. The GeForce RTX 4080 shows a much wider gap between the average FPS and %1 low compared to the Radeon RX 7900 XT.
Similar Comparisons: NVIDIA RTX 4080 Vs RTX 3080 Ti
Need For Speed Unbound
- The GeForce RTX 4080 is, again, victorious in this Radeon RX 7900 XT vs GeForce RTX 4080 battle, with 125 average FPS in this title. The Radeon RX 7900 XT trails at 113 FPS, an 11% difference.
- Radeon RX 7900 XT trails the GeForce RTX 4080 in %1 lows, too, managing 82 FPS. The RTX 4080 scores 91 FPS.
- Both graphics cards show a big gap between average FPS and %1 low, indicating poor game optimization.
- In Red Dead Redemption 2, NVIDIA’s GeForce RTX 4080 averages exactly 100 FPS, and the Radeon RX 7900 XT follows at 91 FPS.
- The GeForce RTX 4080 manages 80 FPS as far as the %1 low is concerned. The Radeon RX 7900 XT is at 69 FPS. This is a difference of nearly 16%.
- The GeForce RTX 4080, again, shows a wider gap between the average FPS and %1 low.
- In our final title, the GeForce RTX 4080 emerges as the winner with an average FPS of 93. The Radeon RX 7900 XT falls behind with 85 FPS, a difference of 9.4%.
- The RX 7900 XT scores 73 FPS in the %1 low, and the GeForce RTX 4080 scores 80 FPS.
Average FPS Across Tested Games
To summarize the average FPS difference between the two GPUs, we derived a mean of the average FPS across the 8 tested games.
- The GeForce RTX 4080 stands tall with 111.3 FPS on average across the 8 tested games. On the other hand, AMD’s Radeon RX 7900 XT averages just under 100 FPS across the 8 titles.
- Overall, the GeForce RTX 4080 is 11.5% faster than the Radeon RX 7900 XT in gaming performance.
Also Read: Radeon RX 7900 XT vs GeForce RTX 3080 Ti
Let’s look at the average power drawn by Radeon RX 7900 XT vs GeForce RTX 4080 during the 8 benchmarks. Remember that the GeForce RTX 4080’s TDP rating is 320 watts, while that of AMD’s Radeon RX 7900 XT is 300 watts.
The Radeon RX 7900 XT draws more power than the GeForce RTX 4080 in 4 games. On the other hand, the GeForce RTX 4080 draws more power in the remaining 4 titles. The graphics cards show similar power consumption in Cyberpunk 2077 and Need for Speed Unbound. The remaining titles show significant contrast in the average power drawn.
Related read: Do You Need a New Power Supply For RTX 4080?
The most significant differences in power draw were seen in Hitman 3 (7.02%) and Spider-Man: Miles Morales (5.8%).
On average, the GeForce RTX 4080 drew 286.73 watts of power, while AMD Radeon RX 7900 XT drew 282.63 watts. So, the average power consumption difference between the two is a mere 1.7%.
Having observed the average power draws, we can move on to the thermals. Let’s look at the average temperature of the two graphics cards during the 8 benchmarking tests.
While both the graphics cards remained under reasonable temperatures during 4K gaming, the Radeon RX 7900 XT gave us a shocker, running significantly hotter than NVIDIA GeForce RTX 4080 during every test except Forza Horizon 5.
Some of the biggest temperature differences can be seen in God of War (8 degrees), Hitman 3 (7 degrees), and Need for Speed Unbound (7 degrees).
The GeForce RTX 4080 kept its temperature under 60 degrees Celsius in every test, a very impressive thermal performance! AMD’s RX 7900 XT ran somewhat hotter, hitting 66 degrees Celsius at maximum while running Hitman 3.
On average, the NVIDIA GeForce RTX 4080 ran at 57.6°C, while the AMD Radeon RX 7900 XT ran hotter, at 62.5°C. You can read about safe idle graphics cards temperature here.
Pricing And Availability
Having looked at the performance of Radeon RX 7900 XT vs GeForce RTX 4080, let’s discuss their prices and price-to-performance ratios.
The GeForce RTX 4080 launched with an MSRP of $1199 on Nov 16th, 2022. On the other hand, the AMD Radeon RX 7900 XT launched with a price tag of $899. That’s $300 cheaper than the RTX 4080. To make matters worse for NVIDIA, hardly any variants of the card are available for the base MSRP.
We saw in the gaming performance tests that the GeForce RTX performs better in almost every game, with an overall lead of 11.5% in gaming. At the same time, it costs 33.3% more than the AMD RX 7900 XT. Also, it draws slightly more power than its AMD competitor. Thus, the RX 7900 XT is easily the better graphics card as far as the price-to-performance ratio is concerned.
Regarding availability, we know that both these cards were launched just a few months ago. Despite that, NVIDIA is facing issues of shortages. Most GeForce RTX 4080 variants, including the Founder’s Edition, are out of stock as of 15th January 2023. A lot of the cards are being scalped and sold at higher prices. So, you may struggle to find the card at its base MSRP.
On the other hand, the Radeon RX 7900 XT is readily available in the markets, including Amazon and Newegg, and can easily be purchased for its original $899 price tag.
Which One Should You Get?
Finally, we can conclude this topic since we have discussed all the factors you should keep in mind. So which card should you get?
The GeForce RTX 4080 provides better performance in 4K gaming than the Radeon RX 7900 XT, but as we discussed, the GeForce RTX 4080 costs $300, or more, extra than its AMD counterpart. The Radeon RX 7900 XT has a significant price-to-performance advantage, and it delivers outstanding performance for a much more economical price tag than NVIDIA.
In terms of thermals, the RTX 4080 runs cooler than its AMD competitor, but the 7900 XT doesn’t run alarmingly hot by any means. The 7900 XT also draws slightly less power than the RTX 4080, but the difference here is also insignificant. In any case, we hope that our Radeon RX 7900 XT vs GeForce RTX 4080 guide will help you make the right decision.
Frequently Asked Questions
The significant difference between the two cards is in their core design. The GeForce RTX 4080 has a monolithic die, while the Radeon RX 7900 XT features two dies: A Graphics Compute Die and a Memory Cache Die.
Yes, as far as sheer gaming is concerned, the RTX 4080 is definitely superior to the RX 7900 XT, with an average FPS lead of about 11.5%.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔
Ali Tasueef is an experienced writer and computer enthusiast with a passion for graphics cards and processors. With a wealth of knowledge in the field of computer technology, he regularly publishes blogs and in-depth comparisons on Tech4Gamers.