Most consumers still have certain questions about the RTX 4080 graphics card. The reason is that the price of the GeForce RTX 4080 series graphics card has been greatly improved compared to the previous generation. As the second leader in the RTX 40 series graphics card, how does GeForce RTX 4080 16GB perform? Let us tell you the answer through this article.
Ada Lovelace architecture parsing
AD103 and AD102 are not very different in architecture design. One computing unit includes FP32 computing unit, computing unit shared by FP32 and INT32, and the fourth-generation Tensor Core, and also includes the third-generation RT computing unit, making deep learning and ray tracing more efficient.
AD103 has 45.9 billion transistors, 378.6 square millimeters of chip area, and at the same time, it has 80 SM cells, with up to 10,240 transistors, and is equipped with 320 Tencor Cores, 80 RT Cores and 112 ROPs, while the L2 cache is 64MB. There is no doubt that in terms of specifications, compared with AD102, the specifications of AD103 are still greatly reduced, which to a certain extent widens the performance gap between GeForce RTX 4080 16GB and GeForce RTX 4090 graphics cards. At the same time, from the perspective of specifications, it seems that GeForce RTX 4080 16GB and GeForce RTX 4090 graphics cards. At the same time, from the perspective of specifications, it seems that GeForce RTX 4080 16GB and GeForce RTX RTX The 3080 Ti cannot widen the gap. Will this affect the gaming performance gap between the two?
Customized TSMC 4N process
If you want to stuff more transistors with almost the same core area, there is no doubt that it needs to be greatly updated in the process. This time, Nvidia adopted the TSMC 4N NVIDIA customization process, which not only can insert more transistors, but also the core frequency of GPU has been greatly improved, which actually improves the graphics performance of the graphics card.
The second generation of hardware optical flow accelerator
In recent years, ray tracing has become the standard for many 3A masterpieces. As manufacturers become more and more aware of ray tracing, in addition to some 3A masterpieces that are realistic in the picture, some small-level games have also begun to add ray tracing technology one after another. Domestic game manufacturers are loyal users of these new picture technologies. Many games also add ray tracing effects. Of course, the addition of ray tracing technology will undoubtedly greatly increase the computing power of graphics cards, which is even more so for high-resolution gaming. With the help of the second generation of hardware optical flow accelerator, NVIDIA has brought consumers the DLSS 3 technology. The principle of
hardware optical flow accelerator is to extract the image optical flow information in it, and at the same time refer to the motion vector information in the game engine to generate new intermediate frames with the help of AI. In addition, with the help of interpolation and super resolution, the RTX 40 series graphics card can achieve almost the same picture output as long as the traditional 1/8 of the rendering volume, thus allowing the GPU's traditional image rendering engine to release the performance, and the game frame rate has been greatly improved. However, from the review of RTX 4090, it can be seen that DLSS 3 has increased the game's frame rate, but also increased the game's delay. At this time, NVIDIA Reflex is about to take action. With the help of NVIDIA Reflex, the game delay performance based on DLSS 3 can be improved.
specification parameters
NVIDIA GeForce RTX 4080 graphics card is based on the AD103 core and has 9728 stream processors , 304 texture units and 112 raster units. The pixel filling rate reaches 28 billion per second, while the texture filling rate is 761.5 billion. In addition, it is equipped with 16GB of GDDR6X video memory, with a memory bit width of 256Bit and a bandwidth of 716.8GB/s. In terms of other parameters, the core frequency of the NVIDIA GeForce RTX 4080 graphics card is 2205MHz, the Boost frequency is 2505MHz, and the video memory frequency is 1400MHz.
NVIDIA GeForce RTX 4080 graphics card seems to be not much different from NVIDIA GeForce RTX 3080 Ti in terms of specifications, but thanks to the new architecture and higher frequency, the actual performance is naturally better.
Appearance
Let's take a look at the NVIDIA GeForce RTX 4080 public graphics card. Like the NVIDIA GeForce RTX 4090 public graphics card, the NVIDIA GeForce RTX 4080 public graphics card will also be available in the domestic market. At that time, you can purchase this graphics card on the e-commerce platform on the evening of November 16.From the appearance point of view, the NVIDIA GeForce RTX 4080 public graphics card is not much different from the NVIDIA GeForce RTX 4090 public graphics card, and it gives people a big impression. Although the overall size is not much different from the previous generation, the fan is larger than the previous generation public graphics card, giving people a stronger visual effect. The fan size of the
NVIDIA GeForce RTX 4080 public graphics card has increased compared to the previous generation. The official said that it can provide up to about 20% of the air volume and has better heat dissipation performance. Of course, with the help of metal wind, the NVIDIA GeForce RTX 4080 public graphics card still gives people a sense of exquisiteness. The back of the graphics card of
NVIDIA GeForce RTX 4080 still adopts a magnetic patch design, with built-in screw holes, which are designed to install graphics card fixing brackets in some cases, while most ordinary players who do not need a bracket can attach a patch to make the graphics card more beautiful overall.
On the power supply interface, the NVIDIA GeForce RTX 4080 public graphics card adopts a 12+4Pin power supply design, of which 12Pin is responsible for power supply, and can provide 450W power output under default conditions, and 4Pin is responsible for signal transmission, so that the power supply of this 12Pin can reach up to 600W. However, there is no graphics card that can provide 600W power supply yet. Perhaps for future flagship graphics cards, 600W is not impossible. This 12+4Pin power supply can fully meet the use of the NVIDIA GeForce RTX 4080 public graphics card. In addition, the official continues to be equipped with a 3X8Pin adapter cable. Of course, it is best to support the ATX 3.0 standard power supply, which can minimize the clutter of the cable. In terms of
interface, the NVIDIA GeForce RTX 4090 public graphics card continues to use 3 DP 1.4 interfaces and an HDMI interface. There is nothing to say about this. It is a bit regrettable not to use the DP 2.0 interface. After all, there are more and more high-refresh + high-score screen displays. For one thousandth of the harsh players, this wish must be met.
platform configuration
We also chose the enthusiast configuration in recent years in terms of hardware configuration, including the Intel Core i9-12900K processor. In addition, the motherboard uses ROG's Z690 Hero motherboard, and the memory is the DDR5 memory of Yingruida , with a total capacity of 64GB. In addition, we also chose two PCIe 4.0 SSDs, WD_BLACK SN850 and WD_BLACK SN770, to minimize the bottleneck of the system.
In addition, Microsoft Windows 11 is becoming more and more mature, so we have also chosen the Win11 22H2 version system, and the driver is 526.72 version. This is a test driver, and there is still a difference between it and the actual driver. In the future, the performance of graphics cards will be improved with the improvement of the driver.
Performance performance
Theoretical performance
As the first choice for graphical theory testing, 3DMark is naturally the most popular testing software. Through FireStrike, TimeSpy and Royal Port, we tested the performance of graphics cards in DX11, DX12 and ray tracing respectively. Naturally, we also used 3DMark to test the theoretical results of NVIDIA GeForce RTX 4080 16GB.
In the 3DMark test, with higher frequency and stronger architecture, the NVIDIA GeForce RTX 4080 16GB graphics card leads the GeForce RTX 3080 Ti in the theoretical test of 3DMark, and the leading margin is the lowest in FireStrike Extreme, because the CPU performance limits the performance of the NVIDIA GeForce RTX 4080 16GB.
rasterized game
After reading the theoretical test of 3DMark, I believe everyone is also very interested in the actual game test of GeForce RTX 4080 16GB. In actual games, due to the differences in optimization, the actual game performance gap will be larger or smaller. Nowadays, there are more and more ray tracing games, so this time we directly divide the game into traditional rasterized games and ray tracing games for comparison and testing. Considering the positioning of GeForce RTX 4080 16GB, we all turn on all special effects, while ray tracing games turn on the maximum ray tracing special effects, and turn off DLSS to see the performance of the graphics card under heavy pressure.
is a rasterized game first. Compared with the RTX 3080 16GB, the performance improvement of GeForce RTX 4080 16GB is about 30%, among which "Red Dead Redemption 2" has the highest improvement, with an increase of 39%, which is a regular upgrade. Compared with RTX 4090's RTX 4090, the performance improvement of GeForce RTX 4080 16GB at 2K resolution is still about 33%, which is not much different from the improvement range at 4K resolution. It seems that CPU has not yet completed it to form a bottleneck, which is still a certain difference from the severe drag of the RTX 4090 by the CPU at that time.
ray tracing game
. In ray tracing games, GeForce RTX 4080 16GB uses more advanced ray tracing computing units, so its performance improvement is higher than that of traditional ray tracing games, reaching 35%.
At 2K resolution, the performance improvement of GeForce RTX 4080 16GB can also reach about 33%. It’s just that in the test of Ubisoft ’s “Far Cry 6”, the performance of GeForce RTX 4080 16GB is not the GeForce RTX 3080 Ti, which should have a lot to do with Ubisoft’s optimization.
From the tests of both, we can see that the game performance frame rate of GeForce RTX 4080 16GB can be increased by 30-35% compared to the GeForce RTX 3080 Ti, which is a normal upgrade. After all, we are testing the GeForce RTX 3080 Ti. If you are a GeForce RTX 3080 graphics card, the performance improvement will continue to increase.
DLSS 3 Test
The above tests are actually conventional game tests, after all, the NVIDIA DLSS function is not used, and DLSS 3 is a feature highlighted by the RTX 40 series graphics card this time. This technology uses the built-in optical stream processor of the RTX 40 series graphics card to realize the simulation and interpolation of game frames, greatly improving the game frame rate, and at 4K resolution, there is no loss in the actual game screen. So how is the gaming performance of NVIDIA DLSS 3 for the GeForce RTX 4080 16GB? Does it play a very positive role in improving game frame rate like the GeForce RTX 4090?
In games that support DLSS 3, such as "Cyberpunk 2077", we can clearly find that it is divided into multiple options in DLSS, including DLSS Frame Generation and DLSS Suer Resolution. The former is an important feature of DLSS 3, interpolated frame rendering, while the latter is DLSS 2, that is, super resolution, including automatic, performance, quality, super performance and other options. At the same time, considering the problem of rendering queues in the game, all DLSS 3 games are currently bound to NVIDIA Reflex. As long as DLSS 3 is opened, NVIDIA Reflex will be opened by default.
In fact, DLSS 3 adds frame generation technology to improve game frame rate. Therefore, in terms of picture quality performance, DLSS 3 is similar to DLSS 2. Now for 3A masterpieces, DLSS has been used quite well, so there is not much difference between DLSS 3 and native 4K screens, especially in close shots, but in the distance, the DLSS 3 screens have blurred behavior.
The biggest need for players to choose DLSS 3 is to improve the game frame rate, especially for players who use high resolution and high refresh rate displays. Therefore, when comparing with the native 4K resolution, we found that DLSS 3 has greatly improved the frame rate of the game, with an average increase of about 3 times, which is quite obvious.
In fact, in addition to DLSS 3, current games also provide DLSS 2 options. You can turn on DLSS 2 mode by turning off the DLSS Frame Generation option. In the comparison between DLSS 2 and DLSS 3, the improvement of DLSS 3 is about 25-35%, among which "Cyberpunk 2077", which has the most complex ray tracing, has the largest improvement. However, in the game "Shadow Torch City", DLSS 2 and DLSS 3 have the same frame rate under 4K picture quality, while DLSS 3 has significantly improved under 2K picture quality. It is suspected that DLSS Frame Generation does not play a switch role.
Of course, while DLSS 3 increases the game frame rate, the side effect of the game is to increase the game rendering delay. Therefore, NVIDIA uses Reflex to reduce the game delay. Through the test, when turning on Reflex Boost Plus mode, DLSS 3 and DLSS 2 do not turn on Reflex. Therefore, when using DLSS 3 special effects, it is recommended to turn on the maximum Reflex Boost Plus.
In addition, everyone is also concerned about how many games that support DLSS 3 can be played when GeForce RTX 4080 is launched? It is reported that on November 15, a total of ten games including "Plague Legend: Requiem", "Light Memory: Infinite", "Marvel Spider-Man: Remake", " Microsoft Flying Simulation " support DLSS 3. In the future, the DLSS 3 game library will continue to be added with the development of game manufacturers. NVIDIA is still very fast in game access, and basically every month, new DLSS 3 games will be added to the game library.
Power consumption and temperature
The biggest impression that this generation of RTX 40 series graphics cards gives people is its huge size, which naturally brings a better cooling environment. Even for flagship graphics cards like GeForce RTX 4090, the heat dissipation of public graphics cards is only about 68 degrees. So the official TGP of the RTX 4080 graphics card that uses a new process has dropped to 320W, which makes people wonder how its actual gaming power consumption performance is? We used the 3DMark TimeSpy Extreme stress test, and selected the opening scenes of "Cyberpunk 2077" BenchMark and "Plague Legend: Requiem", and used FrameView to record the actual power consumption of the graphics card.
In fact, compared with the graphics cards in the past, the RTX 40 series graphics cards give people an impression that they cannot run through the TGP power consumption wall. For example, in the 3DMark TimeSpy Extreme test, the peak power of GeForce RTX 4080 16GB is also 300W nodded. In actual gaming test, the power consumption of GeForce RTX 4080 16GB is about 280W in the future, which is far lower than the TGP threshold of 320W. The RTX 30 series graphics cards will basically touch the upper limit of TGP. In addition, GeForce RTX 4080 The maximum temperature of 16GB is only 66 degrees, which is not much different from non-water-cooled non-public graphics cards. It can be said that this year's public graphics cards are very worth buying in terms of performance and temperature.
overclocking
Currently, the performance of graphics cards is almost at its peak after leaving the factory, and is called "factory ashes" by players. However, the overclocking ability of graphics cards is still a little bit. We also use overclocking software to simply overclock the GeForce RTX 4080 16GB.
Thanks to the TSMC 4N process, the frequency of the GeForce RTX 4080 16GB graphics card is increased to around 250MHz. Although the Boost frequency is 2505MHz, when it is actually running, the frame rate of the graphics card is 2775MHz. After overclocking, the frequency of the graphics card reaches 3015MHz, but the performance improvement is about 3.6%, which is better than nothing.
summary
If you only judge it from the performance and power consumption temperature, then the GeForce RTX 4080 16GB graphics card can be regarded as a remarkable product, and the performance improvement is also in line with expectations. The 35% increase compared to the RTX 3080 Ti is also expected. At the same time, thanks to the excessively hardened radiator, the temperature of the graphics card is only around 65 degrees, and the performance improvement of DLSS 3 is also extremely obvious. It seems that the GeForce RTX 4080 16GB graphics card is indeed worth buying.
Only after the price factor was added, the GeForce RTX 4080 16GB graphics card will be re-evaluated. After all, players regard it as a replacement for GeForce RTX 3080. The GeForce RTX 3080 is priced at 5,499 yuan, while the GeForce RTX 4080 16GB price has reached 9,499 yuan. This price increase has far exceeded the performance improvement. For consumers with a budget of around 7,000 yuan, they have to consider whether this graphics card is worth buying at a higher price.
It is not a bad idea to regard the GeForce RTX 4080 16GB graphics card as a newer and more product of the GeForce RTX 3080 Ti. After all, the performance improvement of this graphics card is there. As for whether you are willing to pay for the DLSS 3 special effects, the decision lies with the consumers. In addition, Nvidia's old rival AMD has also recently released the Radeon RX 7900 series graphics cards, including Radeon RX7900XTX and other graphics cards. At this node, Nvidia launched the GeForce RTX 4080 16GB graphics card, so it is naturally necessary to choose to confront it head-on.The GeForce RTX 4080 16GB graphics card with higher energy efficiency ratio and DLSS 3 bonus is also full of confidence in the upcoming opponents and waiting for their opponents to face their positive challenges.
