Since the 1995-2005 era of VGA cards, and the release of Nvidia’s GeForce 256 GPU in 1999, there have been a lot of technological advancements in the world of Graphics Cards. New generations of cards are released every year, with improved efficiency and record-breaking performance.
To study the progression of GPUs, we create a GPU hierarchy that gives us a solid comparison different chips.
The hierarchy of Graphics Cards is created with respect to their performance, as compared to their peers.
The performance of the GPU is evaluated by running multiple tests, that simulate the real-world usage of the card. This includes the average Framerates of the GPU in the most highly demanding AAA titles, G-Flops and T-Flops testing, bandwidth test, and rendering test.
Some factors like the value of the card, release date, and TDP aren’t of much importance as we are evaluating the cards for raw performance. Benchmarks from trustable sources will be studied and used for comparison.
Generations and Architectures of GPU Matters in Assessment
The most important factor in the assessment of these GPU is to keep track of the generations, and the technologies used in the card. Higher bit-rate or more V-Ram does not automatically translate to ‘better GPU’.
Now there are multiple generations of GPUs, both from AMD and Nvidia. There are architectural differences between these generations (like Ampere from Nvidia).
This is true that new technologies improve the overall performance, but this definitely does not mean that a card from the older generation will not be able to compete with the newer cards.
Some examples of this are the 1080 vs 1660 Super and the RTX 2080 Ti vs RTX 3060 Ti, where the older GPU out-performs the new one by a respectable margin.
Benchmarks Are Our Metric for Gauging GPU Performance
Performance benchmarks provide us with the insight as to what the GPU is capable of, and how it weighs against its counterparts.
In this GPU hierarchy only the more recent GPUs are included in tabular form, while older cards (like the old GTX 1000 or GTX 900 series from Nvidia) are left out with some exceptions.
One thing to note is that there can be conflicting benchmark results from different sources.
The reason for this is the difference in benchmarking methodologies. Geekbench’s OpenCL benchmarks the GPU with Sobel, Canny, Particle Physics, and Depth of field etc.
While PassMark (G3D Mark) runs simulations on the card as well as general purpose computing tests. To improve the accuracy, almost all benchmarking websites collect performance data from their users. This turns in millions of results, which can then be rounded off to reach the final score for the GPU.
Note, the benchmark scores should be taken with a pinch of salt. while they can give insights, they may or not represent the definitive capability of a GPU.
GPU Hierarchy 2022 – Graphics Card Benchmark Performance Ranked
We can divide the performance category of the GPU into several tiers.
Platinum Class GPU: Top of the Line – Worst Performance Per Dollar Value
These are top of the line GPUs. While performance / dollar value is not entirely in favor of these GPU, they are intended for enthusiasts who cannot settle for anything less than the best.
They represent the epitome of performance in the commercial GPU Hierarchy.
At an MSRP of $1,999 and G3DMark score score of 30,105, the RTX 3090Ti has a value of 15 points / USD. Hence not the best value indeed.
NVIDIA RTX 3080 has an MSRP of $699 with G3DMark score/USD = 36 points / USD making it twice more cost effective as compared to an RTX 3090Ti.
|1||NVIDIA GeForce RTX 3090 Ti||24 GB |
|2||AMD Radeon RX 6950 XT||16 GB |
|3||NVIDIA GeForce RTX 3080 Ti||12 GB |
|4||NVIDIA GeForce RTX 3090||24 GB |
|5||AMD Radeon RX 6900 XT||16 GB |
|6||NVIDIA GeForce RTX 3080||10 GB |
|7||AMD Radeon RX 6800 XT||16 GB |
Gold Class GPUs: High End – Best Performance Per Dollar Value
These GPUs are intended for most high performance gaming builds. An average professional gamer would aim for these graphics cards.
These graphics card often represent the best value for money in the GPU Hierarchy.
The NVIDIA RTX 3070Ti G3DMark score/USD value of 39 points / USD (taking the MSRP of $600).
With an MSRP of $399, the RTX 3060Ti has G3DMark score/USD value of cool 50 points / USD!
You can see that the Gold Class GPUs offer a better value as compared to the Platinum Class GPUs.
|8||NVIDIA GeForce RTX 3070 Ti||8 GB |
|9||NVIDIA GeForce RTX 3070||8 GB |
|10||NVIDIA GeForce RTX 2080 Ti||11 GB |
|11||AMD Radeon RX 6800||16 GB |
|12||NVIDIA GeForce RTX 3060 Ti||8 GB |
|13||NVIDIA Titan RTX||24 GB|
|14||NVIDIA GeForce RTX 2080 SUPER||8 GB |
|15||AMD Radeon RX 6700 XT||12 GB |
|16||NVIDIA GeForce RTX 2080||8 GB |
|17||NVIDIA GeForce RTX 2070 SUPER||8 GB |
Silver Class GPUs: Mid Range – Average Performance Per Dollar Value
Silver Class GPUs are mostly from the older generation of graphics cards which are still highly relevant for the GPU Hierarchy and gaming today.
Taking the MSRP of these GPUs, they do not always have the best value.
Take for instance the MSRP of RTX 2060 Super. It is the same as that of the Gold Class 3060Ti i.e $399.
Hence, unless you can find a super cool deal on the Silver Class GPUs, buying the older gen graphics cards at their MSRP value would not be the best course of action.
The only current gen GPU in this list is the AMD Radeon RX 6600. With the MSRP of $330 and a G3DMark score of 14,009, this has a value of 42 Points / USD – which is fairly decent but still not as great as Gold Class GPUs.
|18||NVIDIA GeForce RTX 2060 SUPER||8 GB |
|19||NVIDIA GeForce GTX 1080||8 GB |
|20||NVIDIA GeForce GTX 1070Ti||8 GB |
|21||AMD Radeon RX 6600||8 GB |
|22||NVIDIA GeForce RTX 2060||6 GB |
|23||AMD Radeon RX 5600XT||6 GB |
Bronze Class GPUs: Low-End
These are the bear minimum graphics card that any gamer worth their salt would choose to go for.
While certainly not the best in the GPU Hierarchy, they have an excellent performance / dollar value which is almost as good, if not better, than the Gold Class CPUs.
Take the NVIDIA GeForce RTX 3050, with an MSRP of $250, this GPU has a G3DMark / USD value of 51 points per USD.
Similarly take the AMD RX 6500XT. With an MSRP of $200 and a G3DMark score if 9324, this GPU has a G3DMark/USD value of a cool 46.62 points / USD.
But again, they are not intended for QHD, let alone 4K gaming or gaming at higher than 60FPS a FHD.
|24||NVIDIA GeForce GTX 1070||8 GB|
|25||NVIDIA GeForce RTX 3050||8 GB|
|26||NVIDIA GeForce GTX 1660 Super||6 GB|
|27||NVIDIA GeForce GTX 1660 Ti||6 GB|
|28||NVIDIA GeForce GTX 1660||6 GB|
|29||NVIDIA GeForce GTX 1650 Super||4 GB |
|30||AMD Radeon RX 6500 XT||4 GB |
|31||AMD Radeon RX 5500 XT||4 GB |
There are further variants of some GPUs (Max-Q design chips etc.) that have not been included in this list. This is because these variants do not necessarily represent the GPU and have significantly lower performance. Basically, all the mobile variants as well as the GPUs made for industrial use, have been excluded.
There are a few GPUs, such as the 1080 Ti, that have a decent score with one benchmark tool (in this case PassMark: 18349) but not with the other tools.
This comes down to the testing methodologies of the benchmarking service. Anyways, the GPUs with the most balanced score over all websites have been included in the list.
This list also does not represent the in-game FPS that these cards will produce. Some game settings will benefit more from an RTX enabled GPU, while others will benefit more from a graphics card with a higher VRAM (TITAN RTX for instance).
Hence, the exact performance of the GPU differs from game to game.
The Current Trend and Improvements of the GPU Market
The most recent GPUs are now being designed on 7nm and 8nm architecture by AMD and Nvidia respectively. This has reduced the TDP of these cards as compared to the older generations.
We have also seen a significant jump in the performance (benchmark scores) every single generation. For instance, the RTX 3080 is 58% faster than its predecessor, RTX 2080. (userbenchmarks.com).
This improvement can be credited to better architectures, additional cores, enhanced teraflops and AI integration etc.
AMD is has also shown gradual improvement with every release. The RX 5700 XT which had a 71091 score on Geekbench, was superseded by the RX 6700 XT which got a 36% effective speed improvement and got a 102988 score on Geekbench.
All in all, with innovations happening in the tech industry every day, we can safely assume there is more to come. According to rumors, Nvidia is set to launch its RTX 4000 series, while AMD is preparing its RX 7000 series.
Both of these new generations of cards will have upgraded technologies and baselines. AMD’s RX 7000 series is speculated to be based upon their new RDNA 3 technology, teamed with the new flagship Navi 33 processor.
The AMD RX 7800, 7800XT and the top tier RX 7900 XT can be expected to hold solid positions in the top hierarchy.
Nvidia on the other hand is set to release its RTX 4000 series later this year. The time of release was set to be around mid-July of 2022, but has been pushed back.
This series is expected to be based upon Nvidia’s new Ada Lovelace architecture. The most interesting thing about these cards, is that they are expected to be based on 5nm architecture which will further reduce the TDP and heating issues, as well as provide enhanced performance.
The RTX 4080, 4080 Ti, and the most anticipated 4090 Ti will be the showstoppers for Nvidia, and will surely land firmly at the top of the hierarchy alongside AMD’s 7000 series.