RX 7900 XT Vs RTX 3080 [Gaming Benchmarks 2023]

Every one and a half years, enthusiasts see AMD and NVIDIA release the latest series of graphics cards. As such, AMD’s latest Radeon RX 7000 series graphics cards are the talk of the town. However, how do they fare against NVIDIA’s previous-gen GPUs? Therefore, we’re testing the RX 7900 XT Vs RTX 3080 GPUs to find out.

Key Takeaways

The NVIDIA GeForce RTX 3080 is one of the high-end GPUs from NVIDIA’s previous 3000 series lineup. On the other hand, the AMD Radeon RX 7900 XT is one of the high-end GPUs from AMD’s latest lineup.
The AMD Radeon RX 7900 XT completely plummets the NVIDIA GeForce RTX 3080 in gaming performance at 1440P and 4K resolutions.
The AMD Radeon RX 7900 XT is ~9C cooler than the NVIDIA GeForce RTX 3080.
The AMD Radeon RX 7900 XT is ~28.6% more expensive in terms of MSRP.
While the AMD Radeon RX 7900 XT has more RT cores, the NVIDIA GeForce RTX 3080 might have better raytracing capabilities.

Architectural Differences

Because both graphics cards we’re testing are from different companies, several architectural differences exist. Moreover, these differences cause a rift in performance, which is why they’re important for consumers to know when making a purchase. For example, if you’re interested in raytracing, AMD graphics cards might not be suited to your requirements.

Also Read: RTX 4080 Vs RX 6900 XT

RDNA 3 — AMD Radeon RX 7000 Series Architecture

Since their failed “Vega” graphics cards, AMD’s RDNA technology has gained traction with their 6000 series GPUs. With added benefits and an introduction to raytracing and upscaling technologies, RDNA 2 competed on par with NVIDIA’s GeForce RTX 3000 series GPUs. AMD has launched its latest RDNA 3 graphics cards, which offer greater performance.

With the latest multi-chiplet design, AMD has produced better graphics cards without exponentially increasing costs.

Memory Cache Die

multi-chiplet design introduction — An introduction to the multi-chiplet design in AMD’s Radeon RX 7000 series GPUs. Source: AMD.

AMD has taken inspiration from its Ryzen processors by enveloping a new chiplet design for its Radeon 7000 series GPUs. To explain, RDNA 3 offers multiple dies, named “Memory Cache Dies” (MCD). Each MCD contains 16MB of Infinity Cache. That equals 80MB of L3 Cache in the Radeon RX 7900 XT.

Furthermore, the RDNA 3 has a single die that contains all the GPU cores, called the “Graphics Compute Die” (GCD). However, the GCD is based on TSMC’s 5nm process node. On the other hand, the MCD is based on TSMC’s 6nm process node. Because the latter is older technology, it’s cheaper for AMD to mass-produce MCDs, dropping overall costs.

Also, as the GCD houses all of the GPU cores, higher transistor density due to the 5nm process node should help improve performance.

Also Read: RTX 3080 Vs RX 6800 XT

AI Accelerators

Another new feature with AMD’s Radeon RX 7000 series GPUs is the AI accelerators. Every Compute Unit (CU) has two raytracing accelerators and one AI accelerator. While these AI accelerators’ exact motive is unclear, we can assume they will improve AMD’s FSR 3.0 technology.

Another use for these AI accelerators can be centered around GPU-based machine learning. Because NVIDIA and Intel have extremely powerful upscaling technologies such as DLSS and XeSS, AMD might use these accelerators to improve its FSR technology.

Until RDNA 2’s Radeon RX 6000 series GPUs, AMD was infamous for relying on its spatial FSR filter. Due to that, FSR could never match DLSS’ level. However, we might see FSR overthrow DLSS with new updates.

Nonetheless, with AMD’s AI accelerators, we may see that for the usage of FSR, users must have GPUs that contain these accelerators, similar to NVIDIA’s tensor cores.

Fidelity FX Super Resolution 3.0

FSR 3.0 for RX 7000 series — Introducing FSR 3.0 at AMD’s RDNA 3 keynote announcement. Source: AMD.

In recent years gamers have seen various software-oriented technologies released to help improve the gaming experience. Raytracing and upscaling technologies have almost completely changed how enthusiasts used to choose graphics cards. However, in AMD’s case, raytracing is a lost hope, but its upscaling technology has the potential to rival its competitors.

AMD released its Fidelity FX Super Resolution (FSR) technology with the Radeon RX 6000 series GPUs. While FSR held potential, because AMD did not have dedicated technologies to upscale frames, they relied on their spatial FSR filter. However, we’re seeing AI accelerators and Fluid Motion Frames with the latest FSR 3.0 for RX 7000 series GPUs.

While almost no information on Fluid Motion Frames has been released, AMD has stated the technology will help generate unique frames on its own. AMD has also shared that Fluid Motion Frames utilizes frame generation similar to NVIDIA’s recently released DLSS 3.0.

Ampere — NVIDIA RTX 3000 Series Architecture

As part of the tradition, NVIDIA developed their RTX 3000 series GPUs codenamed “Ampere” almost two years ago. While the GPUs launched in 2020, they have started to gain attraction now due to AMD’s latest releases. However, pricing and thermal efficiency have confused consumers about which side to choose.

In any case, NVIDIA’s RTX 3000 series GPUs come with certain technologies, such as DLSS 2.0, with a neverending fluctuating price.

DLSS 2.0

NVIDIA first released their next-gen Deep Learning Super Sampling (DLSS) technology with their RTX 2000 series GPUs. Due to serious frame drops from activating raytracing, DLSS helps negate that side effect. However, the upscaled frames have a lower quality, which can be problematic for a few gamers.

To counter the low-quality upscaling, NVIDIA released the DLSS 2.0 with their RTX 3000 series GPUs. While quality has improved drastically, enabling raytracing is still a bit difficult. For that reason, NVIDIA has gone one step ahead.

Through iterations and updates, NVIDIA has finally unveiled its latest DLSS 3.0 technology along with the Ada Lovelace GPUs. Unlike traditional upscaling, DLSS 3.0 generates entirely new frames to help increase the FPS. The concept is similar to what AMD might be planning with its AI accelerators for FSR 3.0.

But, along with the RX 7900 XT Vs RTX 3080 war, we’re also looking at a DLSS 2.0 Vs FSR 3.0 war. Although the DLSS 3.0 concept is amazing, the output can have less detailed frames, which are visible. While it might not be an issue for some consumers, it can hinder others. On the other hand, DLSS 2.0 is more stable and provides great quality upscaled frames.

Also Read: NVIDIA DLSS Vs AMD FSR

Process Node

One of the biggest anticipations with next-gen hardware is the improvements in the process node. While sometimes we don’t see many improvements, NVIDIA has shortened the gap with their advancement to Samsung’s 8nm process node. To explain, the NVIDIA GeForce RTX 2000 series GPUs utilized TSMC’s 12nm process node.

The smaller the process node, the greater the number of transistors placed on it. Therefore, we see 28.3 billion transistors on the NVIDIA GeForce RTX 3080 graphics card. While it doesn’t hold a candle against AMD’s Radeon RX 7900 XT, housing 57.7 billion transistors, the RTX 3080 has proven its performance time and time again.

Resizable BAR

With NVIDIA’s GeForce RTX 3000 series graphics cards, we have seen access to the latest NVIDIA Resizable BAR technology. In simple words, Resizable BAR allows transfers of shaders, textures, and geometry through the VRAM to the CPU without any limits. If you think the concept sounds similar, it’s because AMD’s Smart Access Memory (SAM) does nearly the same thing.

Although the technology is amazing, in practicality enabling Resizable BAR will only improve performance by 2-4%. Even though the performance increment isn’t amazing, free performance is always appreciated. However, it would be best to remember that Resizable BAR is only compatible with Ryzen 5000 series processors and above and anything after Intel’s 10th gen processors.

RX 7900 XT Vs RTX 3080 Specifications

After examining the architectural differences between both GPUs, let’s discuss their specifications. While it’s important to test products in practical environments, knowing their specifications gives consumers an idea of which GPU is better. Also, it becomes easier to find both GPUs’ performance per dollar ratio.

**RX 7900 XT Vs RTX 3080**
GPU Name	NVIDIA GeForce RTX 3080	AMD Radeon RX 7900 XT
Node	Samsung 8nm	TSMC 5nm (GCD), TSMC 6nm (MCD)
Architecture	Ampere	RDNA 3
CUDA Cores/ Stream Processors	8,704	5,376
Core Clock	1440MHz	1500MHz
Boost Clock	1710MHz	2394MHz
Memory	10GB GDDR6X	20GB GDDR6
Memory Clock	19 Gbps	20 Gbps
Memory Bandwidth	760.3GB/s	800GB/s
RT Cores	68	84
Tensor Cores	272	N/A
Transistors	28.3 billion	57.7 billion
TDP	320W	300W
Power Connector	1 x 12 pin	2 x 8 pin
Launch Date	September 17, 2020	December 13, 2022
MSRP	$699	$899
Current Price	$975	$900

Regarding the above chart, we see several interesting differences between the AMD and NVIDIA graphics cards. For example, because the AMD Radeon RX 7900 XT is more recent, it utilizes TSMC’s 5nm process. On the other hand, the NVIDIA GeForce RTX 3080 utilizes Samsung’s 8nm process.

However, if we notice the CUDA cores, the NVIDIA GeForce RTX 3080 has a substantially higher number of CUDA cores than the AMD Radeon RX 7900 XT. Similarly, the Radeon RX 7900 XT has a boost clock of 2394MHz, while the NVIDIA RTX 3080 has a boost clock of 1710MHz.

Furthermore, there is a slight difference in power requirements in favor of the AMD GPU. Lastly, the NVIDIA GeForce RTX 3080’s current price is ~8.3% more than the AMD Radeon RX 7900 XT. So, if the former manages to get over 8.3% faster performance than its competitor, it will have a better performance per dollar ratio.

Gaming Performance At 1440P

As stated above, when comparing two pieces of hardware, we must compare their performance in a practical environment. So, in the case of a GPU comparison, it’s natural to compare their performance in multiple games at different resolutions. Because we’re analyzing high-end GPUs, we shall look at the benchmarks of multiple games at 1440P and 4K resolutions. Benchmark PC Tech has conducted the following 1440P benchmarks.

Testing Rig

Processor: Intel Core i5-13600K
Motherboard: ASUS STRIX Z790-F
RAM: Kingston Fury Renegade RGB DDR5 32GB 6400MHz
Storage: Samsung 980 SSD 1TB
Case: Cooler Master H500P
Graphics Card: AMD Radeon RX 7900 XT & NVIDIA GeForce RTX 3080

Dissecting the specifications of the testing rig, we see the Intel Core i5-13600K serving as the main processor. While the graphics cards being tested are higher-end products, the Intel Core i5-13600K won’t bottleneck either GPU. Also, because a greater 1440P resolution is used, most of the load falls on the GPU. So, a slightly weaker processor shouldn’t change the outcome.

After discussing the processor, we see that the ASUS Strix Z790-F motherboard houses the CPU. Not only is the ASUS Strix Z790-F one of the best motherboards in the market right now, but it also offers amazing compatibility with the Intel Core i5-13600K. Moreover, if necessary, you can easily overclock your Intel 13th gen processors on the ASUS Strix Z790-F.

Now, moving toward the RAM, we have the Kingston Fury Renegade DDR5 kit. Although DDR5 RAM is slightly overkill by current standards, it narrowly improves performance in certain games. So, 32GB of 6400MHz RAM will benefit gaming and productive tasks. Additionally, the Samsung 980 1TB SSD will help decrease game load times.

Lastly, the build is encompassed within the Cooler Master H500P case. Because of the size of the latest GPUs, it is important to buy bigger cases to accommodate these massive blocks of performance. Furthermore, the Cooler Master H500P allows for great air cooling, improving the GPUs’ overall thermal efficiency.

Battlefield V

RX 7900 XT and RTX 3080 test — Battlefield performance at 1440P resolution with ultra settings.

The AMD Radeon RX 7900 XT averages 204 FPS, whereas the NVIDIA GeForce RTX 3080 averages 182 FPS.
The AMD Radeon RX 7900 XT is ~12.08% faster than the NVIDIA GeForce RTX 3080.
In contrast, the NVIDIA GeForce RTX 3080 is ~8.3% more expensive yet loses to the AMD Radeon RX 7900 XT.

Red Dead Redemption 2

RX 7900 XT performance test — Red Dead Redemption 2 performance test.

The AMD Radeon RX 7900 XT averages 129 FPS. On the other hand, the NVIDIA GeForce RTX 3080 averages 93 FPS.
The AMD Radeon RX 7900 XT is ~38.7% faster than the NVIDIA GeForce RTX 3080 in Red Dead Redemption 2.

Far Cry 6

RTX 3080 and RX 7900 XT tests — Far Cry 6 FPS test at 1440P resolution.

The gap widens with the AMD Radeon RX 7900 XT averaging 122 FPS, whereas the NVIDIA GeForce RTX 3080 averages 100 FPS.
The AMD Radeon RX 7900 XT keeps the lead by being ~22% faster than its competitor in Far Cry 6.

God Of War

The AMD Radeon RX 7900 XT thrashes its rival by averaging 146 FPS. Meanwhile, the NVIDIA GeForce RTX 3080 averages 95 FPS.
While the NVIDIA GeForce RTX 3080 costs ~8.3% more, the AMD Radeon RX 7900 XT is ~53.68% faster in God of War.

Assassin’s Creed Valhalla

The AMD Radeon RX 7900 XT averages 127 FPS, whereas the NVIDIA GeForce RTX 3080 averages 80 FPS.
The AMD Radeon RX 7900 XT Vs RTX 3080 war rages on, with the RX 7900 XT being ~58.75% faster in Assassin’s Creed Valhalla.

Microsoft Flight Simulator

The AMD Radeon RX 7900 XT wavers with an average of 53 FPS. Moreover, the NVIDIA GeForce RTX 3080 gains ground by averaging 51 FPS.
The AMD Radeon RX 7900 XT is ~3.9% faster than the NVIDIA GeForce RTX 3080 in Microsoft Flight Simulator.
While the gap was wider, Microsoft Flight Simulator is a more CPU-intensive game, dropping the overall FPS.

Hitman 3

The AMD Radeon RX 7900 XT averages 135 FPS, while the NVIDIA GeForce RTX 3080 averages 122 FPS.
The AMD Radeon RX 7900 XT is ~10.65% faster than its competitor in Hitman 3.

Gaming Performance At 4K

After looking at the performance benchmarks at 1440P resolution, we must evaluate the differences at 4K resolution. While the AMD Radeon RX 7900 XT was dominant in all of the games at 1440P, the situation may reverse when the resolution is changed. It is also important to note that none of these benchmarks utilize raytracing and DLSS. Mark PC has conducted the following tests.