At CES 2022, NVIDIA unveiled the new GeForce RTX 3090 Ti. It has 10,752 stream processors with a base frequency of 1560 MHz and an acceleration frequency of 1860 MHz, equipped with 24GB of GDDR6X video memory at a rate of 21 Gbps, plus a 384-bit memory bit width, and the theoretical memory bandwidth exceeds 1 TB/s, which will be the first time in the history of GeForce graphics cards to cross 1 TB/s of bandwidth.
NVIDIA said that the GeForce RTX 3090 Ti single-precision floating-point performance is 40T FLOPS, reaching the highest level of graphics cards currently in the Ampere architecture. Its TDP was increased to 450W and it was the first graphics card to feature the new PCIe 5.0 external power interface. However, there are still some interesting things about the GeForce RTX 3090 Ti, so let's take a look at it.
The complete ga102 core appears
The complete GA102 core has finally come this time, it still follows the GPC-TPC-SM hierarchy architecture set from the NVIDIA Pascal architecture, the complete GA102 core has a total of 7 complete GPC, each group of GPC contains 6 groups of TPC, each group of TPC contains 2 sets of SM, so the complete GA102 core has 84 SM units.
And in the Ampere architecture, the number of CUDA cores per SM unit is now 128, which makes the GeForce RTX 3090 Ti have a total of 10752 stream processors of superscale. But compared to the previous GeForce RTX 3090, is the GeForce RTX 3090 Ti primarily just a few more stream processors? This is not the case.
Let's first take a look at the GeForce RTX 3090 Ti compared to the TITAN RTX, GeForce RTX 3090 and GeForce RTX 3080 Ti specifications:
In terms of positioning, GeForce RTX 3090 Ti and GeForce RTX 3090 are the successors of the previous generation OF TITAN RTX, the positioning of the previous generation of TITAN RTX is a solution for AI and data science, but from the current generation, titan this series seems to be cut, so GeForce RTX 3090 Ti and GeForce RTX 3090 were entered into the GeForce lineup, But their target group is clear - for 8K gamers and heavy creators, and among the two, they are actually more inclined to heavy creator users.
24GB GDDR6X video memory with error correction code (ECC).
If geForce RTX 3090 positioning is a distinctive feature of productivity tools is 24GB of large-capacity GDDR6X video memory, then GeForce RTX 3090 Ti is a step further in this regard, that is, its memory configuration is not only large capacity, but also supports ECC error correction codes, which is one of the two places where I personally think geForce RTX 3090 Ti is the most different from GeForce RTX 3090. It also makes the GeForce RTX 3090 Ti look a lot like the RTX A6000, a specialty card in the field of design and visualization solutions.
But the GeForce RTX 3090 Ti and the RTX A6000 have their own strengths, the typical feature of the RTX A6000 is that it can be equipped with up to 48GB of video memory, but the GeForce RTX 3090 Ti features faster memory than the RTX A6000 - GDDR6X. Relatively speaking, the design and visualization production process is more focused on the demand for large capacity, and the game requires low latency, so there are also requirements for the speed of the memory, the faster the better, of course, the resolution of the game and the quality of the texture also affect the demand for the memory capacity.
So here we can see a GeForce RTX 3090 Ti an unsurpassed point - that is, its memory features combine the "fast" of GDDR6X, the "large" capacity of 24GB and the "stability" of supporting ECC error correction codes, which can be said to be a generalist without shortcomings.
There is an ECC switch in the control panel of the driver, the default is off, for the creation software to open the ECC will increase stability, and for game users, it does not need to be opened, due to one more step of error correction code calculation, will slightly affect the performance, the default is not moving.
The frequency increase is huge
The above mentioned video memory support for ECC error correction codes is one of the two places where the author personally believes that the GeForce RTX 3090 Ti is the most different from the GeForce RTX 3090, and the second place is not more stream processors, but frequency.
For large cores such as the GA102-300 of the GeForce RTX 3090 and the GA102-350 of the GeForce RTX 3090 Ti, it is difficult to have high frequencies at the same time, because even if it can be high frequencies, it will also face the problem of power consumption and heat generation, after all, with the increase of frequency, power consumption and heat generation are increased in a ratio of beyond linearity, which is why we see that the frequency of each generation of flagship cards is often not the highest in the same architecture.
The GeForce RTX 3090 Ti is really surprising, its Boost frequency reached 1860MHz, compared to the GeForce RTX 3090's 1695MHz increase of about 10%, compared to the stream processor unit is only less than 3%.
Video memory bandwidth finally crossed the 1 TB/s mark
On the GeForce RTX 3090, the memory bandwidth has reached 936 GB/s, which is only one foot away from the 1 TB/s mark, who will achieve this last foot? Of course, it was NVIDIA itself, so this time the GeForce RTX 3090 Ti, which went further at the end of the beast, finally achieved more than 1 TB/s of video memory bandwidth, reaching 1008 GB/s, which is the first time in the history of GeForce graphics cards to cross 1 TB/s of bandwidth.
And in addition to this first, the GeForce RTX 3090 Ti is also the first graphics card to use the new PCIe 5.0 external power interface.
Similarly, the GeForce RTX 3090 Ti graphics card based on the new generation of Ampere architecture also has some common features of this generation of Ampere graphics cards, and is a great improvement over the previous generation of Turing architecture.
Introduction to the NVIDIA Ampere architecture
The biggest change between the SM unit on the GeForce RTX 30 Series graphics card and the GeForce RTX 20 Series is the doubling of the FP32 unit for traditional computing, the introduction of the second generation RT Core, and the third generation Tensor Core.
As we all know, in the Turing architecture, NVIDIA integer (INT32) and single-precision floating-point (FP32) two different data types are handed over to two different ALU for computation. However, the most common in modern gaming applications is FP32, so in order to improve computing efficiency, NVIDIA has introduced a new ALU that can support both INT32 or FP32 data types on the NVIDIA Ampere architecture. That is, there are now two different data paths, one that can handle integers or single-precision floating-point at the same time, and the other that simply handles single-precision floating-point calculations.
The dedicated hardware unit RT Core, which performs real-time ray tracing operations, has also been updated to the second generation on the NVIDIA Ampere architecture, most notably with the addition of dynamic blur accelerated computing support.
NVIDIA's new interpolation algorithm can improve the efficiency of real-time ray tracing while ensuring the accuracy of motion blur, and the official said that it can reach up to 8 times the speed of the previous generation. In addition, in the basic BVH calculation, the second generation of RT Core can also be 2 times faster than the first generation.
Tensor Core, the hardware unit responsible for running AI computing, has also been upgraded to the third generation on the NVIDIA Ampere architecture. In fact, the previously released A100 computing card has used the new third-generation Tensor Core, which can provide 4 times more performance than the second-generation Tensor Core, but the Tensor Core on the game card has been streamlined to a certain extent, and the throughput of its FP16 FMA calculation is only half of the Tensor Core in the GA100 core.
And what is the use of the more powerful AI operations brought by the third generation Tensor Core? The answer is DLSS. Along with the Release of the GeForce RTX 30 Series graphics cards, there is also an updated version of DLSS - DLSS 8K. As the name suggests, DLSS 8K is a new version of DLSS that stretches the picture resolution to 8K through deep learning technology, specifically stretching the 1440P screen to 4320P, and the number of pixels spans a full 9 times.
Then there's HDMI 2.1, the highly anticipated new output port. Under HDMI 2.1, graphics cards can achieve an output of 8K60Hz or 4K120Hz with a single cable. It's a good boon for players who want to play games on a large-size TV.
Asus TIF RTX 3090 Ti O24G GAMING graphics card exterior
The appearance of the ASUS TIF RTX 3090 Ti O24G GAMING graphics card is basically the same as the design of this generation of TIF graphics cards, and it also adopts a metal fairing with a low-key dark pattern design, and is equipped with a metal backplane to protect the PCB while also assisting in heat dissipation. In addition, the surface of the fairing is processed by two processes, namely the frosting process and the drawing process.
The graphics card radiator is equipped with three Axial-Tech axial fans of about 10 cm diameter, which use double ball bearings with fan rings at the end of the fan blade, which can form a higher and more directing air volume and wind pressure than ordinary fan blades, and has a low temperature automatic stop function. In addition, it is worth mentioning that the fan in the middle is opposite to the other two in the direction of rotation, which can effectively reduce the airflow interference between the fans, which is also conducive to improving the heat dissipation effect.
The metal backplane surface of the ASUS TUF RTX 3090 Ti O24G GAMING graphics card is treated by a brushing process, the thickness is about 2mm, which can protect the graphics card PCB from bending due to gravity, and the tail end has a window that allows the cooling wind flow to pass better to enhance the heat dissipation effect of the radiator and the backplane. The overall length of the graphics card is 32.4cm, the height is 12.7cm, and the thickness is 5.9cm, which is equivalent to a 3-slot thickness.
The graphics card does not use the traditional PCIE power supply interface, but the new PCIe 5.0 external power supply interface design, that is, the 16Pin power supply interface. This is a very unique aspect of this card compared to the graphics cards of other GPU models, and in order to ensure compatibility with older power supplies, the ASUS TUF RTX 3090 Ti O24G GAMING graphics card also provides a 12Pin to 3 8Pin adapter cable.
Since the graphics card is a 16Pin power supply interface, and the adapter cable is 12Pin, the user may be worried about whether the 3 8Pin to 12Pin power supply lines connected to the 16Pin power supply interface of the graphics card will not have an impact, which can actually be assured that the signal interface of the 4 Pin is currently not affected.
The development of PCIe 5.0 external power supply interface has gone through a process, NVIDIA launched a 12Pin interface when the GeForce RTX 30 series graphics card was released, which can provide up to 600W of power supply, and not long ago Intel released ATX 3.0 and ATX12VO 2.0 specifications to take a step forward, setting the interface as 16Pin, of which 12Pin is responsible for power supply, and another 4Pin is responsible for signal transmission for power supply and PCIe 5.0 Graphics Card Direct Communication determines the power supply limit.
However, before the ATX 3.0 and ATX12VO 2.0 specifications said that "the new wire connector will mark the corresponding rated power, there are 150W, 300W, 450W and 600W four gears" This time has not seen, the wire in hand is not marked with power information.
After looking at the power supply interface, we look at the display interface, the graphics card provides 3 DP 1.4a interfaces and 2 HDMI 2.1 interfaces, which provides 1 MORE HDMI 2.1 interface than most graphics cards, which has better compatibility for users to perform multi-screen output.
There is a switch next to the power supply interface of the graphics card, which can be switched in the dual BIOS of the graphics card, the two BIOS correspond to the performance mode and the silent mode, the default state of the author to get the card is the performance mode, and the later test will also be tested in this mode to reflect the actual use of most users.
Finally, another point that is very unique in appearance compared to other GPU model graphics cards is the NV Link interface.
ASUS TIF RTX 3090 Ti O24G GAMING graphics card disassembly
Asus TIF RTX 3090 Ti O24G GAMING graphics card the main body heat dissipation using 6 6mm diameter heat pipes, the use of reflow soldering process and the base and heat dissipation fin combination; fins can be divided into two large areas, in addition to the GPU itself also take care of the VRM power supply area before and after the graphics card, the overall heat dissipation area is very large; the base is pure copper nickel plating design, the surface is mirror polished, can make the radiator surface and GPU chip better contact, in order to improve thermal conductivity efficiency.
And it also has a heat pipe to put the video card PCB left of the power supply MosFET heat transfer area of the largest set of fins, and also through the thermal paste and the main radiator interconnection, can transfer heat to the main radiator, while the bottom of the integrated design, can achieve a certain degree of heat balance.
In addition to the role of reinforcing and protecting the PCB, the back panel of the graphics card is also pasted with thermal conductive adhesive, which can guide part of the heat of the video memory and power supply circuit to the back panel and play a better auxiliary heat dissipation effect.
The graphics card uses a GA102-350 core with the full specifications of the GA102 core, a total of 84 sm cells, a total of 10752 stream processors.
There are 12 Micron GDDR6X memory around, the model is Micron's D8BZC, a single 2GB capacity, 32-bit wide, 12 pieces a total of 24GB 384-bit memory specifications.
The change in video memory this time is a very significant change compared to the GeForce RTX 3090, the new 2GB particles to replace the GeForce RTX 3090 1GB particles can make the GeForce RTX 3090 Ti equipped with the number of memory particles limited to 12, so they can all be placed on the front of the PCB, through the heat sink efficient heat dissipation, which simplifies the design of the graphics card, and reduces the heat dissipation needs on the back of the graphics card.
The author has so far remembered the high heat on the back of the previous GeForce RTX 3090, GeForce RTX 3090 due to the relatively large amount of heat generated by the video memory on the back, the overall heat dissipation design of the graphics card puts forward higher requirements, many users are also troubled by this, and the GeForce RTX 3090 Ti 2GB particles can reduce the heat on the back of the graphics card. And compared to the GeForce RTX 3090, the new memory particle data rate has also increased, from 19.5Gbps to 21Gbps.
In addition to the GPU and video memory on the PCB, the remaining space is basically used to arrange the power supply components, and there are power supply circuits on the left and right sides of the GPU, a total of 21 phases, of which 18 phases power the GPU core and 3 phases power the video memory. The mosFETs in the power supply part are mp86957 of the core source system (MPS), and the power supply control is mainly the MP2891 of the 1 core source system (MPS).
Performance testing
As can be seen through GPU-Z, the base frequency of the ASUS TUF RTX 3090 Ti O24G GAMING graphics card is 1560MHz, and the Boost frequency is 1920MHz, which is 60MHz compared to the 1860MHz frequency of the public version, and the video memory equivalent data rate is 21Gbps.
The FAN speed in standby can also be seen through GPU-Z, due to the previously mentioned ASUS TUF RTX 3090 Ti O24G GAMING with a fan intelligent start-stop function.
In addition, the default temperature cap of the ASUS TIF RTX 3090 Ti O24G GAMING is 83 degrees Celsius, players can choose to unlock up to 90 degrees Celsius, the power consumption limit is 450W by default, and players can choose to unlock up to a maximum of 516W.
Test platform with description
Test platform as shown in the table above, the motherboard uses ROG STRIX Z690-A GAMING WIFI D4 snow blowing motherboard, memory with two Zhiqi Royal Halberd DDR4-3600 16GB to form a dual channel, of which BIOS settings, in addition to turning on XMP to let the memory run at 3600MHz frequency, while opening the Supportable BAR in the motherboard BIOS, others use default settings.
The graphics cards representing each GPU model are: ASUS TUF RTX 3090 Ti O24G GAMING, Boost frequency 1920MHz, NVIDIA GeForce RTX RTX 3090 Founder Edition, Boost frequency 1695MHz, Xunjing XFX Radeon RX 6900 XT Overseas Edition, Boost frequency 2365MHz.
This time the power supply should be specially introduced, because it comes with a 16Pin native cable, used to adapt to the new GeForce RTX 3090 Ti graphics card, for users who have previously purchased Raytheon 2 generation 1000W and above and Raytheon 1 generation 1200W and above, you can apply online like the LGA 1700 buckle of Z690 before, and ASUS will reissue 16Pin cables to old ROG Raytheon users.
Benchmarks
We used 3DMark as a graphics card benchmark, including Fire Strike, Fire Strike Extreme, Fire Strike Ultra, Time Spy, Time Spy Extreme, and Port Royal. Among them, Fire Strike, Fire Strike Extreme, and Fire Strike Ultra test the performance index of the graphics card at 1080p resolution, 2K resolution and 4K resolution in DX11 games, respectively, and the two projects of Time Spy and Time Spy Extreme are the performance index of graphics cards in DX12 games at 2K resolution and 4K resolution, Port Royal is the performance index of real-time ray tracing of the tested graphics card, the specific results are shown in the following table, and the scores listed in the table are the scores of the 3DMark graphics card individual items.
In terms of benchmarking, the GeForce RTX 3090 Ti is an average 10% improvement over the GeForce RTX 3090, while the comparison with the Radeon RX 6900 XT still shows different results with different rendering pressures, in the FireStrike project with the lowest rendering pressure, the Radeon RX 6900 XT has the highest running score, higher than the GeForce RTX 3090 Ti, as the rendering pressure increases, GeForce RTX 3090 Ti has a growing advantage, and by the Time Spy Extreme project, geForce RTX 3090 Ti has shown a significant lead, and the Port Royal project with light chase has widened the gap. Overall, the average result for the 6 projects was a 7% improvement in the GeForce RTX 3090 Ti compared to the Radeon RX 6900 XT.
4K game testing
4K rasterized game testing
The traditional rasterized game screen test part will set the game quality to the highest quality setting in the preset ("Metro: Gone" select Ultra image quality, "Cyberpunk 2077" is the default preset that does not turn on ray tracing), the default is not full-screen manual change to full screen, the default opens the vertical synchronization manually turn off vertical synchronization, in addition to the other options are default settings, all use the game's own Benchmark output results.
In traditional rasterized game screen tests, the GeForce RTX 3090 Ti improved by an average of 8.5% over the GeForce RTX 3090, and the gap was more than 10% in Cyberpunk 2077, Far Cry 6 and Gears of War 5, with frame rates 12.5%, 10.8% and 12% higher, respectively.
The GeForce RTX 3090 Ti is an average 16% improvement over the Radeon RX 6900 XT, with the biggest gap in Cyberpunk 2077, Total War: Three Kingdoms and Anno 1800, while the GeForce RTX 3090 Ti is 30%, 24% and 25% higher than the Radeon RX 6900 XT, respectively.
Real-time ray tracing game testing
In terms of light chasing game screen test, the ray tracing effect is set to "high", because the algorithm of resolution stretching technology such as DLSS and FSR is not the same, and the final picture effect is also different, in order to maintain the accuracy of the test and ensure the same picture results, these tests do not open DLSS and FSR and other resolution stretching technologies.
In terms of light-chasing game screen testing, the GeForce RTX 3090 Ti improved by an average of 10.6% compared to the GeForce RTX 3090, which was a big gap in Cyberpunk 2077, Bright Memory and Frontier, and the GeForce RTX 3090 Ti had a 13.8%, 12% and 14.4% higher frame rate, respectively.
The GeForce RTX 3090 Ti is an average 40% improvement over the Radeon RX 6900 XT, with the biggest gap in Cyberpunk 2077, Watchdog: Legion and Frontier, with the GeForce RTX 3090 Ti 105%, 65% and 86% higher than the Radeon RX 6900 XT, respectively.
8K light chase game test
This part does not have the results of the Radeon RX 6900 XT, and the Radeon RX 6900 XT is currently not up to the 8K light chasing game even with the FSR. This part sets the game quality to the highest quality setting in the preset ("Metro: Gone" select Ultra quality), turn on the ray tracing effect to "high" (the preset maximum quality of Cyberpunk 2077 has turned on the ray tracing effect, this part is no longer changed), turn on DLSS to super performance mode, the default is not full-screen manual change to full screen, the default is turned on vertical synchronization manually turn off vertical synchronization, in addition to other options are default settings, All use the Benchmark output that comes with the game.
In the 8K Light Chasing Game Screen Test, the GeForce RTX 3090 Ti improved by an average of 12% compared to the GeForce RTX 3090, and the gap was larger in Cyberpunk 2077, Metro: Gone (Enhanced Edition), Light Memory and Tomb Raider: Shadow, with frame rates 12.4%, 13.2%, 12.8% and 12.5% higher, respectively.
Author app experiences
GeForce RTX 3090 Ti is suitable for heavy creators, mainly because of its large-capacity video memory configuration, so the graphics card compared here is not the same capacity of GeForce RTX 3090, but GeForce RTX 3080 Ti, and because the test of heavy authoring project files is large, it requires a large amount of memory capacity to fully load correctly, so two 16GB of memory are added to form a total capacity of 64GB. Some of blender is compared under Windows 10 due to bugs under Windows 11, and other projects are compared under Windows 11.
Compared with GeForce RTX 3090 Ti and GeForce RTX 3080 Ti, the difference in the experience of creation is not mainly a difference between a faster and a slower one, but when facing some large projects, one direct one can work normally, and the other will be because of insufficient memory capacity, the project cannot continue, or the direct program error situation, which has a great impact on the creative process.
Blender
In Blender, after the material loading of this test, due to insufficient video memory, the GeForce RTX 3080 Ti directly reported an error when performing rendering, indicating that the video memory was not enough.
The GeForce RTX 3090 Ti completes the workflow normally.
Davinci
In Davinci, this test is an 8K RED RAW material, after adding some effects to the material, GeForce RTX 3080 Ti has not been able to play correctly, the memory consumption slowly increased until it overflowed and then emptied, and then the software will indicate insufficient video memory and report an error.
The GeForce RTX 3090 Ti can complete the playback preview normally with the same effect settings.
OctaneRender
The situation in OctaneRender is similar to the previous one, with the GeForce RTX 3080 Ti appearing with an orange "OCC" reminder during rendering, that is, the graphics card has insufficient video memory, and the rendering calls memory as temporary storage. You can see that the GeForce RTX 3080 Ti takes up 8.7GB of video memory and borrows 6.9GB of memory.
The GeForce RTX 3090 Ti is easy, with about 18.5GB of project data all in the video memory, which makes it render particularly fast.
It can be seen that due to insufficient video memory, the rendering time is very different, and the GeForce RTX 3080 Ti took a total of 479 seconds, almost 8 minutes to render. The GeForce RTX 3090 Ti takes only 15 seconds to complete.
Power consumption testing
Through the PCAT kit in our hands, we can accurately measure the number of graphics card PCI-E and external power interface watts, the maximum power consumption of the graphics card is obtained in the 3DMark Time Spy stress test, and the standby power consumption is recorded for 1 minute after entering the system to take the average.
The results of the statistical power test calculate that the total standby power consumption of this ASUS TIF RTX 3090 Ti O24G GAMING is 32.4W, the average power consumption of full load is 448.2W, and the peak power consumption is 472W. For such a level of graphics card power consumption, if the power consumption of other components of the platform is not high, then it can be driven by 850W power supply, if the power consumption of other components is also high, it is recommended to start with a power supply of at least 1000W.
Temperature testing
Our GPU thermal tests are all tested in bare metal (if installed in the chassis, the GPU temperature will be about 5°C higher), and the test ambient temperature is about 21.1°C. The standby temperature is recorded for 10 minutes after booting, and the full load temperature is recorded after completing the 3DMark Time Spy stress test, and the data is recorded through the LOG to File function of GPU-Z, and the following is the temperature test curve.
This ASUS TIF RTX 3090 Ti O24G GAMING has a fan standby shutdown function, so it relies on passive heat dissipation in standby. After ten minutes of standby testing, from the beginning of 37.6 degrees Celsius to about 42.4 degrees Celsius, the hottest temperature in the 10-minute standby process is the lowest at the beginning of 46.3 degrees Celsius, and the back section is 50.8 degrees Celsius.
In the full load state, the maximum temperature appears in the moment before the fan is just started, reaching 72.5 degrees Celsius, with the active heat dissipation began, the temperature is controlled, after stabilization, the maximum temperature of the card is only 65.3 ° C, pay attention to the hottest temperature, after stabilization, the highest is only 75.7 degrees Celsius, which is a very good full load temperature performance.
Overclocking test
After unlocking the power cap and the maximum temperature limit to the maximum, after several manual overclocking attempts, the GeForce RTX 3090 Ti finally overclocked at a core frequency of +150MHz and a GDDR6X video memory data rate of +2Gbps. In this state, this ASUS TIF RTX 3090 Ti O24G GAMING can successfully pass 3DMark's Time Spy Extreme project test and score the highest score, at this time the card has a Boost frequency of 2070MHz and a memory equivalent frequency of 23Gbps. This is a pair of GPU scores for the ASUS TIF RTX 3090 Ti O24G GAMING running 3DMark Time Spy Extreme projects in an overclocked state and a GPU score in silent mode as shown in the table below.
Under this amplitude, Asus's GeForce RTX 3090 Ti O24G GAMING graphics card score in the Time Spy Extreme is 11994 points, which is about 8.1% higher than the silent frequency, which is terrifying for the big core of the card king level.
summary
Overall, the GeForce RTX 3090 Ti is generally positioned to be similar to the previous GeForce RTX 3090, for 8K gamers and heavy creators, but the GeForce RTX 3090 Ti surpasses the GeForce RTX 3090 in all directions, which is equivalent to pushing the current top consumer graphics card one step further, giving current consumers a new choice for the top GPU.
For the 4K and below resolution of the game environment, in fact, the most recommended is geForce RTX 3080 Ti, GeForce RTX 3080 Ti can already meet the top 4K game experience, and in terms of price, it is also quite appropriate, GeForce RTX 3080 Ti The design of all aspects of the entire card is also geared towards game design. For the GeForce RTX 3090 Ti, it is still relatively recommended for creative purposes, but it also provides good game performance, but for creative purposes, its value is greater, and the characteristics of the video memory error correction code (ECC) can also be played.
Graphics mini ladder list
The GeForce RTX 3090 Ti's 450W power consumption is also a point that cannot be ignored, but this power consumption is also the result of a significant increase in frequency, which also brings a powerful performance improvement. Overall, for a graphics card that starts at almost $15,000, it's destined to be in the pockets of a few people, and in terms of marginal utility, it's not as good as the GeForce RTX 3080 Ti. Overall, this card looks very revealing.
Combined with the recent Intel's Sharp A series mobile terminal unique display and AMD may soon launch the Radeon RX 6x50 XT series graphics, and the new graphics release behind NVIDIA is estimated to be waiting until the next generation of GeForce RTX 40 series to meet again, the focus of the graphics card market may fall on Intel and AMD for a long time, so the author believes that this GeForce RTX 3090 The launch of Ti is a message that NVIDIA wants to convey, that is, in the graphics card field, even if I temporarily "hide the market", you must not forget that there is such an insurmountable mountain.
Finally, it's interesting to note that, given the current environment, the GeForce RTX 3090 Ti is perhaps the most likely graphics card that users will be able to buy at the suggested retail price around launch date and time in the last two years. At present, the news I get from ASUS is that the GeForce RTX 3090 Ti graphics card of its TIF series is 14999 yuan if it is a public frequency, then the price is 14999 yuan at the suggested retail price, if it is an OC version graphics card of the TUV series, the price is slightly higher at 15499 yuan, and the higher-end STRIX series is 16499 yuan.