At this year's CES 2024, NVIDIA has further refined the RTX 40 series graphics card assortment with GeForce RTX 40 Series Super graphics cards. Today, the ninth graphics card in the GeForce RTX 40 series, the NVIDIA GeForce RTX 4080 Super, officially arrived.
The new GeForce RTX 4080 Super GPU delivers revolutionary AI performance for gaming, creation, and work, and is enhanced by a rich software acceleration library built on top of RTX GPUs.
This time, we will bring you the performance test of the GIGABYTE RTX 4080 Super Gaming OC Magic Eagle 16G graphics card, let's see how this RTX 4080 Super graphics card performs in terms of appearance, workmanship, heat dissipation, etc., and actually test how the RTX 4080 Super performs in AI graphics, 3D rendering, etc., and how far ahead the game performance is compared with the RTX 4080 and the previous generation RTX 3090 Ti, and how far ahead of RTX How much performance gap is there for the 4090.
GIGABYTE RTX 4080 Super Gaming OC 16G Specs GIGABYTE RTX 4080 Super Gaming OC 16G graphics card has 10240 stream processors, 112 raster processing units, 320 texture units, 64MB L2 cache, graphics card core clock of 2295-2595MHz, and power input interface is 16PIN (12+4). It uses 16GB GDDR6X video memory, with a memory bit width of 256-bit and a memory bandwidth of 7363Gbs at 23Gbps equivalent.
In terms of specifications, compared to the RTX 4080 graphics card, the number of CUDA cores of the RTX 4080 Super graphics card has increased from 9728 to 10240 at full specification, an increase of 5. The base and acceleration frequencies have been increased by 90MHz and 50MHz, respectively, while the TGP remains unchanged.
The AD103-400 GPRTX 4080 Super uses a complete AD103 GPU core using the TSMC 4N process, with a core area of 379mm2 and 45.9 billion transistors. There are a total of 7 GPCs, of which 6 GPCs contain 6 TPCs and 1 GPC contains 4 TPCs, each TPC has two sets of SMs, for a total of 80 sets of SMs, that is, 10240 CUDAs.
A brief analysis of the advantages of RTX 40 series super graphics cards
If there's one of the most amazing technologies about the NVIDIA GeForce RTX 40 Series GPUs, it's DLSS 3.
Powered by the new Tensor Cores and Optical Flow Accelerator on GeForce RTX 40 Series GPUs, DLSS 3 Frame Generation adds AI-powered frame generation to DLSS 2 Super Resolution technology to create more frames and improve image quality, making games more realistic and dramatically increasing frame rates.
NVIDIA officially released DLSS 3 at the end of August 20235. The technology adds a new Ray Reconstruction (RR), whose main function is to replace the artificially designed denoiser effect with an AI network trained on an NVIDIA supercomputer, resulting in higher quality pixels between sampled rays, which significantly improves the ray-traced image quality of all GeForce RTX GPUs.
In real-world games, lighting with light reconstruction turned on is more accurate, the game is sharper, more detailed, with less ghosting, and more responsive when lighting conditions change dramatically.
GeForce RTX 40 Series GPUs are built on the ADA architecture that unleashes the power of ray tracing, simulating real-world light behavior. With the power of GeForce RTX 40 Series GPUs and 3rd Gen RT Cores, gamers can experience incredible details in virtual worlds like never before.
To date, more than 500 games and apps support NVIDIA RTX technology, and NVIDIA is officially kicking off the RTX 500 event to celebrate this milestone.
In addition to gaming, GeForce RTX 40 Series GPUs also have a very strong performance in AI, with dedicated AI Tensor cores, Studio Clipping, which is very popular with creators, NVIDIA Canvas that turns simple lines into realistic landscapes, NVIDIA Broadcast, Stable Diffusion, and D5, a ray-traced 3D renderer Render and other software can provide users with powerful performance and innovative features, making it easy to improve their creative ability and improve work efficiency.
GIGABYTE RTX 4080 SUPER Magic Eagle Graphics Card Appearance (1).
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card was used in this test, which is GIGABYTE's mid-to-high-end graphics card series, which achieves a good balance between performance, heat dissipation, appearance and price.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card accessory is basically a must-have 16pin to 3*8pin adapter cable, as well as a custom graphics card bracket. The bracket is tailor-made for this graphics card, which allows the graphics card to be more stable and safe, while its style and color scheme also match the appearance of the graphics card, and can be easily integrated with the graphics card.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card is mainly black, with a matte shell on the front and various diagonal lines. The size of the graphics card is 342*150*75mm, which can be compatible with most cases.
The graphics card adopts a three-fan cooling design, with three 110mm diversion fans. Deflector fans divert airflow from the edge of the fan and smoothly guide airflow through a 3D fringe curve on the blade surface. The fan is made of nano-graphene lubricating oil, which can effectively extend the service life of the fan while making the fan rotate more quietly.
The fan is designed to be forward and reversed, and adjacent fans rotate in different directions, which can effectively reduce turbulence and increase air pressure. The graphics card fan supports the auto-stop function, when the temperature is low, the fan will stop working, and when the temperature rises after the game starts, the fan will automatically resume spinning. The intelligent start-stop not only effectively reduces noise, but also increases the service life of the fan, providing gamers with a better experience.
GIGABYTE RTX 4080 SUPER Magic Eagle Graphics Card Appearance (2).
The GIGABYTE RTX 4080 Super Gaming OC 16G features a Gigabyte RGB Faith Light on the side and a GeForce RTX logo in the middle.
The GIGABYTE RTX 4080 Super Gaming OC 16G features a 12+4pin power connector with a power supply indicator that flashes frequently to alert gamers when there is an abnormal power supply.
The output interface of the graphics card is a standard 3*dp 14a + 1*hdmi 2.1a combination.
The GIGABYTE RTX 4080 Super Gaming OC 16G features a reinforced metal backplate that enhances the structural strength and ESD protection of the graphics card's PCB.
An extra-large windshield at the rear of the backplate allows the fan's airflow to penetrate the fins and backplate with lower resistance for more efficient heat dissipation.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card features a dual BIOS design with both OC (Performance) and Slient (Silent) modes, which can be adjusted on the back of the graphics card, and the default is OC mode.
GIGABYTE RTX 4080 Super Magic Eagle Lighting Effects
The graphics card uses RGB Optical Wheel, which supports 16.8 million color adjustments and multiple lighting effect switching, and can be customized through GIGABYTE Smart Manager.
GIGABYTE RTX 4080 Super Gaming OC 16G graphics card lights up the real shot.
GIGABYTE RTX 4080 Super Gaming OC 16G graphics card lights up the real shot.
GIGABYTE RTX 4080 Super Gaming OC 16G graphics card lights up the real shot.
GIGABYTE RTX 4080 Super Gaming OC 16G graphics card lights up the real shot. GIGABYTE RTX 4080 Super Magic Eagle Graphics Card Teardown
The GIGABYTE RTX 4080 Super Gaming OC 16G PCB board is shorter than the heat dissipation module, and with the perforated backplate and triple fan design, the graphics card can quickly dissipate internal heat and get the most out of the graphics card.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card is covered with a thermal pad on the front of the PCB, which can quickly transfer the temperature of the memory and power supply components to the heat dissipation module, accelerating the heat generated inside the graphics card.
The AD103-400-A1 core used by the RTX 4080 Super.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card is designed with 16+3 phases of power supply, and each MOSFET of the graphics card is designed with over-temperature protection to ensure that the graphics card can continue to work stably and ensure the performance of the graphics card.
In the memory part, 8 GDDR6X memory chips from Micron are arranged around the GPU core, each with a capacity of 2GB, for a total of 16GB of video memory. The memory bit width is 256-bit, and the memory bandwidth is 7363gb/s。
The GIGABYTE RTX 4080 Super Gaming OC 16G features a wind-powered cooling system with three 110mm deflector fans, a pure copper base that touches the GPU, nine composite heat pipes, and an air intake grille, allowing the graphics card to quickly transfer heat from the GPU's cores, allowing the graphics card to have faster heat transfer, lower noise, and more powerful heat dissipation.
Through the teardown, it can be found that the GIGABYTE RTX 4080 Super Gaming OC Magic Eagle 16G graphics card has a lot of material in terms of power supply and heat dissipation, which lays a good foundation for the full performance of the graphics card.
Introduction to the test platform and test methods
Test bench at a glance.
List of test itemsGIGABYTE Smart Manager (GCC).
Gigabyte Control Center (GCC) HomeGIGABYTE Control Center (GCC) is a unified software for all Gigabyte products with an intuitive interface and a rich experience. The software allows you to adjust the frequency, voltage, fan mode, RGB lighting effects, and power consumption of the graphics card in real time, allowing gamers to easily control the status of the graphics card.
Gigabyte Control Center (GCC) RGB lighting adjustment interface.
Gigabyte Control Center (GCC) fan control interface.
Gigabyte Control Center (GCC) performance tuning interface3DMark benchmarks
3DMark Fire Strike test.
3DMark Fire Strike Extreme test.
3DMark Fire Strike Ultra test.
3DMark Time Spy test.
3DMark Time Spy Extreme test3DMark ray tracing, DLSS test
3DMark Port Royal test.
3DMark DirectX Ray Tracing Function Test One of the major changes in NVIDIA GeForce RTX 40 series graphics cards is the addition of DLSS3 technology, DLSS3 is based on the previous generation DLSS, and infers the target image generated by the next frame through optical flow acceleration, which greatly saves traditional CUDA computing power, allowing GPUs to be more comfortable when dealing with high-resolution real-time rendering games, compared with not using DLSS, Theoretically, the performance of the game can be increased by up to 4 times. DLSS3 is theoretically twice as capable of delivering as DLSS2 thanks to Frame Generation, which allows GeForce RTX 40 series graphics cards to run all DLSS3-enabled games at high resolution and high quality with less stress.
3DMark DLSSS3 Functional Testing.
3DMark DLSSS2 Functional TestIn the 3D Mark DLSS test category, the results are very similar at 4K resolution, but with DLSS enabled (Performance), the GIGABYTE RTX 4080 Super Gaming OC 16G graphics card can improve the DLSS3 performance by 235% at 4K, far exceeding the DLSS2 159% improvement.
dlss 3.5 games "2077", "Wake Killer 2" test
Since the most valuable advantage of the RTX 40 series graphics cards is the addition of DLSS3 support, we selected 6 DLSS3 support for the game test, and two of them also support the latest DLSS 35 game masterpieces to conduct game performance comparison tests.
Let's test the game performance of DLSS3 (RTX 40 series) and DLSS2 (RTX 30 series) with DLSS turned on and DLSS turned off respectively to see how much the RTX 4080 Super graphics card can be ahead of the RTX 4080 graphics card, how much different it is from the RTX 4090 graphics card, and how much the RTX 4080 Super with DLSS3 can be ahead of the RTX 3090 Ti with DLSS2 on.
Comparison of the performance of Cyberpunk 2077 in 4K resolution.
Comparison of the performance of Cyberpunk 2077 at 2K resolution.
4K resolution "Alan Wake 2" game performance comparison.
2K resolution "Alan Wake Killer 2" game performance comparison"Dying Light 2", "F1 22" game test
Dying Light 2 game performance comparison in 4K resolution.
Dying Light 2 at 2K resolution game performance comparison.
Comparison of the performance of the 4K resolution F1 22.
Comparison of the performance of the 2K resolution F1 22 game"Bright Memory" and "A Plague Tale" game test
Comparison of the performance of the 4K resolution "Bright Memory: Infinite" game.
Comparison of the performance of the 2K resolution "Bright Memory: Infinite" game.
Comparison of the performance of the 4K resolution A Plague Tale: Requiem.
Comparison of the performance of A Plague Tale: Requiem in 2K resolutionPlaytest Summary:
Through testing, we can find that compared to the RTX 4080, the GIGABYTE RTX 4080 Super Gaming OC 16G graphics card has 224% performance advantage, and the suggested retail price is 1400 yuan lower than RTX 4080, and it does increase the volume and reduce the price.
Compared to the previous generation flagship RTX 3090 Ti graphics card, the GIGABYTE RTX 4080 Super Gaming OC Magic Eagle 16G graphics card has a significant gaming performance advantage, with an overall performance advantage of 43%, of which the actual game frame rate is increased by up to 55% with DLSS3 enabled, and a 25% performance advantage even without DLSS enabled.
Compared with the contemporary flagship RTX 4090, the GIGABYTE RTX 4080 Super Gaming OC 16G graphics card still has a certain gap, and the actual gaming performance is 22% lower.
In terms of actual gaming experience, the GIGABYTE RTX 4080 Super Gaming OC 16G graphics card performs with ease, and is the only high-quality game with full ray tracing support, Cyberpunk 2077 and Alan Wake 2, which are the only high-quality games that support panoramic ray tracing, and can also guarantee an average game rate of 60+ FPS when DLSS3 is enabled, bringing gamers a smooth gaming experience.
AI Mapping – Stable Diffusion Testing
In addition to their gaming performance, GeForce RTX 40 Series GPUs are becoming the go-to tool for creators, with AI being a productivity multiplier.
Stable Diffusion is a popular AI painting generation tool that allows you to easily generate the image you want by typing in a sentence or a specific keyword. Based on AI calculations, the software can be widely used in plane and 3D design, and by generating a large number of ** with the help of diffusion models, the drawing time is greatly reduced and the efficiency is significantly improved. The software is accelerated by NVIDIA RTX GPUs, and the higher the memory capacity, the higher the resolution of the images it can support, and the higher the GPU level, the faster the rendering speed.
Stable Diffusion, an AI drawing tool, supports TensorRT, a high-performance deep learning inference SDK from NVIDIA that optimizes inference performance, accelerates a wide range of workloads, and supports large language model inference.
After referencing the relevant plugins, you can add the "Tensorrt" tag to Stable Diffusion, and then go through a series of settings to speed up AI painting through Tensorrt.
The graphics card used in this test is:
GIGABYTE RTX 4080 Super Gaming OC 16G
nvidia geforce rtx 3090 ti
The test model was SD XL Base 10_0.9vae.safetensors
The sampling method chosen in the test is EULAR A, the number of steps sampled is 50, the width and height of ** is 1024 1024, the correlation coefficient of the prompt word is 7, the total number of batches is 10, the number of single batches is 1, and the random number seed is 13 (make sure that the same image can be generated).
Stable Diffusion regular drawing (RTX 3090 Ti) in a total of 1733 seconds, an average of 346 images.
Stable Diffusion regular drawing (RTX 4080 Super) in a total of 1368 seconds, an average of 4 per minute4 images.
TensorRT accelerated Stable Diffusion drawing (RTX 4080 Super) in a total of 635 seconds, an average of 9A comparison of the 5 images shows that the RTX 4080 Super graphics card outperforms the RTX 3090 Ti by 27% when using the same SD XL for Stable Diffusion regular drawing. After adding TensorRT acceleration that supports RTX 4080 Super, the AI drawing efficiency of the RTX 4080 Super graphics card has been greatly improved, which is 175% higher than that of RTX 3090 Ti, and even compared with the RTX 4080 Super graphics card itself without TensorRT acceleration enabled, it also has a 116% efficiency increase after turning on TensorRT acceleration, which can give users a faster AI generation experience.
3D Ray Tracing Renderer - D5 Render test
D5 Render is a free, real-time ray-traced renderer for 3D designers and professionals working on large-scale architectural or landscape projects. With advanced real-time ray tracing, a D5 Gi global illumination solution, and a host of ready-to-use assets, D5 Render can complete projects in less time with high realism.
In the latest generation of D5 Render, NVIDIA DLSS 3., developed for realistic ray-traced visuals, has been added5. The integration includes DLSS super-resolution, frame generation, and ray reconstruction powered by AI neural networks.
Among them, ray reconstruction is a new neural rendering AI model, which can further enhance the visual quality of ray tracing by providing a fast and intelligent denoising solution for various contents. And NVIDIA DLSS frame generation enhances ray tracing performance and increases live viewport frame rates for a smoother editing experience for intuitive, interactive 3D creation.
This time, we tested it with a GIGABYTE RTX 4080 Super Gaming OC 16G and an NVIDIA GeForce RTX 3090 Ti graphics card to take a look at the latest addition of DLSS 35 What an increase in D5 render rendering.
D5 render test (RTX 4080 Super) with DLSS3Super resolution sampling + ray reconstruction + real-time high frame rate in 5 with an average frame rate of 97fps
D5 render test (RTX 4080 Super), DLSS3Super resolution sampling + ray reconstruction + real-time high frame rate in 5 with an average frame rate of 34fps
D5 render test (RTX 3090 Ti) with DLSS3Super resolution sampling + ray reconstruction in 5 with an average frame rate of 44fps
D5 render test (RTX 3090 Ti) with DLSS3Super resolution sampling + ray reconstruction + real-time high frame rate in 5, with an average frame rate of 20FPS, you can see that DLSS 4080 is turned on with an RTX 3 Super graphics cardAfter all the options in 5, rendering a scene with D5 Render can get up to 2x faster performance. Compared to the RTX 3090 Ti graphics card, the RTX 4080 Super has DLSS 3 turned on5 has a 120% performance advantage when it can be turned on. In the same close DLSS 3Under the premise of 5 options, the RTX 4080 Super also has a 70% performance advantage over the RTX 3090 Ti, which can give users a smoother creative experience.
Web**User Benefits – RTX VSR technology
I believe that when you watch**, you will often encounter situations where it is difficult to get a satisfactory viewing experience due to the low definition of the picture. Previously, users who encountered this situation were helpless. However, with the advent of RTX VSR, these issues have been solved – users with GeForce RTX 40 and RTX 30 series graphics cards can use RTX VSR technology to enhance the viewing experience in real time.
The full name of RTX VSR is RTX Video Super Resolution, which uses the Tensor cores in the RTX GPU to do AI calculations to remove blocky compression artifacts from the ** compression algorithm in each browser, so as to achieve the purpose of improving resolution and clarity. This feature currently supports mainstream browsers such as Google Chrome and Microsoft Edge (to upgrade to the latest version), in addition, the Gigabyte RTX 4080 Super Gaming OC Magic Eagle 16G has dual codecs, and also adds a hardware encoding function for **1, whether you are on Station B, Betta or what other platform*** it makes **quality worth it for you to match4K display.
The VSR's feature switch is integrated into the NVIDIA control panel, and it can be seen in the interface of the "Drop-down bar to adjust" Image Settings - RTX enhancements. Tick this option and click on Apply, you can turn on RTX VSR, and it also supports the setting of the optimization level, which is 1 by default and 4 at the maximum.
By increasing the complexity of the VSR algorithm step by step, it is possible to optimize up to 360p to 1440p visuals, and the GPU usage will also increase, which is directly proportional to the quality level of the VSR. Level 1 and Level 2 have a smaller GPU usage, which has significantly improved the quality, and it is also very friendly to lower-end GPUs.
*The left side of the dividing line is the original ** screen, and the right side is the picture with RTX VSR turned on, which shows that the clarity is significantly enhanced, almost close to the effect of rendering CG.
Most of what we see on the web is re-encoded and compressed, reducing capacity and losing some quality. Whereas previously sharpened processes tended to be distorted while improving sharpness, RTX VSR intelligently sharpens and sharpens features and edges while eliminating annoying compression artifacts by combining NVIDIA Tensor Cores with state-of-the-art AL image processing to deliver image quality comparable to the native resolution labeled.
Thermal test and conclusion
The GIGABYTE RTX 4080 Super Gaming OC 16G Fan supports the Idle Idle Idle feature, which can greatly reduce the noise of the graphics card during low load operation and increase the life of the fan.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card delivers excellent cooling performance at room temperature 26 and GPU temperature of 40 after 20 minutes of standby5. The fan is in a stationary state.
GIGABYTE RTX 4080 Super Gaming OC 16G Graphics Card Thermal TestWe used a Furmark oven to test the heat dissipation of the graphics card by observing the final temperature of the graphics card after the temperature of the graphics card has stabilized for a period of time.
The baking time is about 20 minutes, and the core temperature of the graphics card is 636. The memory temperature is 52, the GPU usage is 100%, and the current power consumption is 3184w。The fan load is 53% and the speed is 1883rpm. The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card is very good in terms of heat dissipation.
Conclusion
As the third graphics card of the NVIDIA GeForce RTX 40 Super series, the overall performance of the RTX 4080 Super graphics card is good, on the basis of the price point lower than the previous RTX 4080 1400 yuan, CUDA cores increased by 5%, and the core frequency, video memory bandwidth and equivalent frequency have been improved to a certain extent, and the amount has indeed been increased and the price has been reduced.
In terms of gaming performance, the RTX 4080 Super graphics card can slightly surpass the RTX 4080 in gaming performance at a lower price, and can reach an average game frame rate of 60+ FPS in the highest 4K resolution games, which can bring gamers a smooth 4K gaming experience.
In terms of AI, NVIDIA's dedicated AI Tensor core, combined with software such as Stable Diffusion and D5 Render, can bring users higher performance and efficiency, and give users a smoother creative experience.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card in this review has a low-key and restrained appearance without losing its personality.
The graphics card is made of solid materials, 16+3 high-specification power supply design with an enlarged area of heat dissipation modules, coupled with a powerful wind power cooling system, so that the GPU can be kept in the best working condition for a long time, providing an excellent space for the continuous high-performance output of the graphics card. With a 4-year warranty, the graphics card is a graphics card product that combines appearance, performance, and heat dissipation.
The GIGABYTE RTX 4080 Super Gaming OC 16G graphics card has been launched on major e-commerce platforms at a price of 8799 yuan, which is suitable for gamers who like RGB lighting surrounding, have high requirements for heat dissipation and silence of the graphics card, and focus on 4K high-quality game masterpieces or content creators who have high requirements for AI performance.