GPU Hierarchy 2023: Graphics Card Tier List

GPU Hierarchy - Graphics Card Tier List

Technology has come a long way in the realm of graphics cards, from VGA days in the old days to several upgrades two decades later. Each year since then brings forth awesome new milestones – with ever more powerful and efficient GPUs that deliver top-notch visuals!

To help you make informed decisions, we’ve developed an innovative GPU hierarchy that gives a comprehensive overview of the various chips on the market. With this tool at your fingertips, it’s never been easier to monitor technological advancements and stay ahead in terms of graphics card capabilities!

When searching for the best graphics card, performance is key. That’s why we compare GPUs by running high-stakes tests to gauge their capabilities – from calculating average frames per seconds in popular AAA games and evaluating G-Flops/T-Flops to ultimate rending trials. It may seem like a complex process but it ensures you get the most bang out of your buck!

To figure out which video card is the best for gamers, we need to go beyond a basic analysis of its release date and value. Instead, our focus should be on raw performance – that’s where trusted benchmarks come in! Thankfully this will give us reliable results so everyone can choose their perfect graphics power.

Assessing GPUs: The Importance of Generations and Architectures

Assessing GPUs accurately requires taking into account the generation and architecture of the card. Simply having a higher bit-rate or more V-Ram does not necessarily make a GPU superior.

Both AMD and Nvidia have produced multiple generations of GPUs, each with its own architectural differences (such as Nvidia’s Ampere architecture).

While new technologies can enhance overall performance, it is still possible for older generation cards to compete with newer, more advanced ones. For example, the 1080 versus the 1660 Super and the RTX 2080 Ti versus the RTX 3060 Ti, where the older GPU outperforms the newer one by a significant margin.

GPU Performance Is Measured Using Benchmarks As Our Metric.

Performance benchmarks offer a glimpse into the capabilities of a GPU and its comparison with similar devices.

The GPU hierarchy table typically includes only newer GPUs, while older models such as Nvidia’s GTX 1000 or GTX 900 series may be excluded, with some exceptions.

It’s crucial to acknowledge that benchmark outcomes may fluctuate across various origins due to discrepancies in benchmarking approaches. For instance, Geekbench’s OpenCL benchmarks incorporate Sobel, Canny, Particle Physics, and Depth of Field, whereas PassMark (G3D Mark) conducts simulations and general-purpose computing assessments.

Most benchmarking websites gather user performance data to increase accuracy, resulting in millions of results that can be aggregated to generate the final score for the GPU.

It’s crucial to be aware that while benchmarks can provide valuable insights, they may not be an accurate representation of a GPU’s definitive performance capabilities. Therefore, it’s important to take benchmark scores with a grain of salt.

Top Keys for Measuring GPU Hierarchy

Clock Speed

The clock speed, measured in MHz, defines the velocity of the GPU core within a graphics card. However, it alone doesn’t determine performance. Overclocking the core clock speed can enhance performance.

VRAM

VRAM is the dedicated graphics memory, or RAM, that affects the resolution and texture quality of graphics. The quantity of VRAM is not the sole factor to consider, despite it being measured in GB. Different types of VRAM are available, including the commonly used GDDR5, the high-end GDDR5X used by Nvidia, and the fastest HBM2, which is also the most expensive.

Memory Bus

The memory bus is the communication link between the core and VRAM. A wider memory bus facilitates faster communication between the core and VRAM. HBM2 cards take advantage of this to increase speed across the board.

Tier One 

High-End Graphics Card ($1000+)

GPU Clock Speed VRAM Memory Bus Width
GeForce RTX 4090 2235 MHz 24 GB GDDR6X 384-bit $1,500
GeForce RTX 3090 1395 MHz 24 GB GDDR6X 384-bit $1,499
GeForce RTX 3080 Ti 1365 MHz 12 GB GDDR6X 384-bit $1,199
Radeon RX 6900 XT 2015 MHz 16 GB GDDR6 256-bit $999
Titan V 1200 – 1455 MHz 12GB HBM2 3072-bit
RTX Titan 1350 – 1770 MHz 24 GB GDDR6 384-bit

Champion: RTX 4090

Performance in This Category

The RTX 3090 stands out as a remarkable graphics card that bridges the gap between older Titans and gaming-focused GPUs. With over twice the number of CUDA Cores, 24GB of memory, and a memory bandwidth of 936 GB/sec, it surpasses its forerunners in performance. While Nvidia has yet to release a new Titan level card, the RTX 3090 reigns as an absolute monster in the field.

The RTX 3090’s gaming performance is unrivaled, but its steep price raises questions about its intended audience. Gamers who demand the best and have money to spare will find the RTX 3090’s performance to be unmatched in a high-end system. However, with its high cost and technological advancements, the RTX 3090 is gaining popularity among creatives who require ample VRAM for 3D modeling applications, although it may not be as comprehensive as Quadro.

In terms of gaming performance, the RTX 3090 is capable of running games easily in 4K resolutions and VR, as well as 1440P at 144HZ and 240Hz, although some settings may need to be adjusted for more demanding titles. The RTX 30 series GPUs’ main selling point is their ability to handle 8K gaming.

We advise against buying the RTX 3090, despite its significant performance lead of about 20-30% over the 2080 Ti. The 3080 or RX 6800XT, which can be found for $700-800, offer similar performance. The RTX Titan, on the other hand, is much more expensive.

Waiting for Nvidia to release a new Titan could be the optimal choice if you seek the most exceptional professional card available. Nonetheless, upgrading to the RTX 3090 might be a prudent decision if you possess a low-end Quadro, Tesla, or an older RTX Titan.

Tier Two

Diamond Level Graphics Cards ($600 – $999)

GPU Clock Speed VRAM Memory Bus Width
Radeon RX 6800XT 2015 – 2250 MHz 16 GB GDDR6 256-bit
Radeon RX 6900XT 1274 – 1546 MHZ 8GB HBM2 2048-bit
GTX 1080 Ti 1480 – 1582 MHZ 11GB GDDR5X 352-bit
RTX 2080 Ti 1350 – 1545 MHz 11 GB GDDR6 352-bit
RTX 3080 1440 – 1710 MHZ 10 GB GDDR6X 320-bit

 

Top Pick: RX 6800XT

Performance in this category is exceptional, and for those seeking an upgrade, the RTX 3080 may seem like the obvious choice with its widely-proven benchmarks. However, the highly-anticipated AMD Radeon RX 6800XT offers tremendous value, and based on its launch, there seems to be little-to-no compromise in performance. Additionally, with AMD’s latest Smart Access Memory, pairing the RX 6800XT with a Ryzen 500 series processor could result in even better benchmark results.

Both of these GPUs should make 4K gaming easily attainable, and handling VR and lower resolutions should be effortless. Although the RTX 3080 and RX 6800XT perform comparably, their performance may vary depending on the game. Consequently, we suggest selecting the best value GPU in this category, unless you have particular requirements.

Rest assured, these graphics cards should perform as expected for at least the next four years.

Tier Three

Gold-Level Graphics Cards ($300 – $500)

GPU Clock Speed VRAM Memory Bus Width
Radeon RX 5700XT 1605 – 1905 MHZ 8 GB GDDR6 256-bit
Radeon RX 6800 1815 – 2105 MHz 16 GB GDDR6 256-bit
GTX 1080 1607 – 1733 MHz 8GB GDDR5X 256-bit
RTX 2070 1410 – 1620 MHz 8 GB GDDR6 256-bit
RTX 2070 Super 1410 – 1710(OC) 8 GB GDDR6 256-bit
RTX 2080 1515 – 1710 MHz 8 GB GDDR6 256-bit
RTX 2080 Super 1650 – 1815 MHz 8 GB GDDR6 256-bit
RTX 3070 1500 – 1730 MHz 8 GB GDDR6 256-bit

 

Top Pick: RTX 3070

Performance in this category is essentially entry-level 4K, with strong 1440p and VR capabilities. Thanks to recent launches from AMD and Nvidia, some of the best 4K results we’ve seen at this price point are now available. As you climb higher tiers, the value you get for your money becomes increasingly questionable due to diminishing returns.

At the time of the RTX 3070’s launch, Nvidia dominated this category and remained uncontested. However, with the recent announcement of the RX 6800, performance looks similar to the 3070, although this is relatively unknown until third-party benchmarks are available.

Investing in an RTX 3080 or RX 6800XT may be highly beneficial if you can afford the additional expense. Nonetheless, for budget-conscious individuals seeking advanced technology, the RTX 3070 is unequivocally the most cost-effective option in this category.

These graphics cards are anticipated to perform adequately for at least the following four years, but certain sacrifices might be required to run 4K titles.

Tier Four

Silver-Level Value Graphics Cards ($200 – $300)

GPU Clock Speed VRAM Memory Bus Width
RX 580 1257 – 1340 MHz 8GB GDDR5 256-bit
Radeon RX 5600XT 1375 – 1560 MHZ 6GB GDDR5 192-bit
GTX 1660 Ti 1500 – 1770 MHZ 6GB GDDR5 192-bit
GTX 1660 Super 1530 – 1785 MHz 6GB GDDR5 192-bit
RTX 2060 1365 – 1680 MHZ 6 GB GDDR6 192-bit
RTX 2060 Super 1470 – 1650 MHz 8 GB GDDR6 256-bit

 

Top Pick: RTX 2060 Super

Performance in this category is highly contested due to the number of cards available. Expect excellent 1080p performance, with strong 1440p and VR capabilities. However, 4K gaming is not very feasible in this price range, and attempting high-refresh-rate gaming above 1080p may not go well.

In terms of price-to-performance, the RTX 2060 Super is exceptional and boasts a considerable advantage in gaming, particularly at 1440p. The 2060’s prices and synthetic benchmarks trend slightly higher, but the GTX 1660 Super also deserves recognition for its superb FPS outputs across multiple games.

Assuming the availability of RTX 2060 Super, it is likely to perform as anticipated for the next four years. However, lower-end graphics cards may encounter difficulties sooner, especially when gaming at 1440p.

Tier 5

Bronze-Level Value Graphics Cards ($100 – $200)

GPU Clock Speed VRAM Memory Bus Width
RX 560 1175 – 1275 MHz 4GB GDDR5 128-bit
RX 570 1168 – 1244 MHz 8 GB GDDR5 256-bit
GTX 1660 1530 – 1785 MHZ 6GB GDDR5 192-bit

 

Top Pick: RX 570

In the bronze value category, the performance is mostly limited to high-quality 1080p gaming, which is considered entry-level for most users. While VR is possible, 1440p gaming may not be recommended and may only work for older titles. However, this should be suitable for most people who still use 1080p gaming monitors.

The RX 570 and GTX 1660 cards have surpassed the previously leading 1050 Ti in terms of performance. It is not recommended to purchase cards below this tier as they are unlikely to provide satisfactory performance and will not be a good value for your money.

These GPUs should serve well for 1080p gaming in the foreseeable future, but the RX 570 and GTX 1660 may require more settings adjustments and compromises within the next two years.

New Improvements and Current Market Trends

The latest GPUs from AMD and Nvidia are now built on their 7nm and 8nm architecture, respectively, resulting in a decrease in TDP when compared to earlier generations. Additionally, every new generation has displayed a notable increase in performance, as evidenced by the RTX 3080, which is 58% faster than its predecessor, the RTX 2080, according to userbenchmarks.com. This boost in performance can be attributed to advancements in architecture, the inclusion of more cores, improved teraflops, and AI integration.

Furthermore, AMD has demonstrated steady progress with each new release. The RX 5700 XT achieved a Geekbench score of 71091, but it was surpassed by the RX 6700 XT, which saw a 36% increase in effective speed and scored 102988 on Geekbench.

The constant progress in the tech industry suggests that there are more advancements to come. Reports suggest that Nvidia is preparing to release its RTX 4000 series, and AMD is getting ready to introduce its RX 7000 series, both of which are expected to feature upgraded technologies and baselines.

The AMD RX 7000 series is expected to be based on their new RDNA 3 technology, combined with the new flagship Navi 33 processor, and is anticipated to have a strong presence in the upper echelon, with the RX 7800, 7800XT, and the top-tier RX 7900 XT.

Meanwhile, Nvidia is planning to launch its RTX 4000 series later this year, following a delay from its original mid-July 2022 release date. The upcoming graphics cards are anticipated to utilize Nvidia’s new Ada Lovelace architecture, which will be built on 5nm technology. This advancement should further reduce TDP and heating issues while simultaneously boosting performance.

Among the expected releases, the Nvidia RTX 4080, 4080 Ti, and the highly anticipated 4090 Ti are predicted to be the standout performers, occupying the top tier of the hierarchy alongside AMD’s 7000 series.

How to Choose a Graphics Card

Determine your needs

Before purchasing a graphics card, determine what you plan to use it for. Do you plan to use it for gaming, video editing, 3D modeling, or other graphic-intensive tasks? 

Selecting a graphics card that aligns with your specific requirements is crucial, as different tasks necessitate varying levels of performance. For instance, a graphics card that excels in gaming might not be suitable for video editing.

Consider the specifications

Graphics cards have several specifications that determine their performance. The most important specifications to consider are:

Graphics Processing Unit (GPU): The GPU is the heart of the graphics card and determines its overall performance.

Clock speed: The clock speed refers to how fast the GPU runs, and a higher clock speed means better performance.

Memory: Graphics cards have their own memory, known as Video RAM (VRAM), which stores data used by the GPU. A higher VRAM means better performance.

Bus width: The bus width refers to how much data the graphics card can transfer at once, and a higher bus width means better performance.

TDP (Thermal Design Power): The TDP refers to the amount of power the graphics card consumes and how much heat it generates.

Check compatibility

Before purchasing a graphics card, ensure that it’s compatible with your computer’s motherboard and power supply unit (PSU). Graphics cards require a specific slot on the motherboard, and some may require additional power connectors from the PSU.

Budget

There is a wide range of graphics cards available at various price points, making it crucial to choose one that aligns with your budget. However, keep in mind that the more expensive the graphics card, the better the performance.

See Also: GPU Artifacting: Common causes and how to fix GPU artifacting?

Refresh Rate and Monitor Resolution

When choosing your next graphics card, it’s important to consider your monitor and its refresh rate. If you already have a monitor, make sure you know its resolution, and if you don’t, decide what resolution and refresh rate you want for your next one. These factors will impact your gaming experience, especially if you get a high-resolution and refresh rate monitor but pair it with a budget-tier graphics card instead of a mid-range or high-end one. In this instance, the graphics card will be unable to match the capabilities of your monitor.

For example, if your target resolution for both competitive and visually stunning games is 2K at 60Hz or 1080p at 144Hz, consider a graphics card like the RTX 3080 or RX 6800 XT. If you’re aiming for 4K resolution, opt for an RTX 4080, RTX 4090, or RX 7900 XT. For the popular resolution of 1440p with a 144Hz refresh rate, consider a graphics card like the RTX 4070 Ti or RX 7900 XT.

Case Compatibility and PSU

Graphics cards have grown more powerful, but that has come with an increase in power usage as well. Today’s flagship and mid-range graphics cards alone consume around 250-500W of power at stock settings. As a result, it’s highly recommended to have a sufficient power supply to run everything smoothly without the risk of damage from burnout or electrical fluctuations. Ensure that you get a high-quality PSU with all the necessary certifications and recommended power output.

Additionally, graphics cards have become bulkier and larger, necessitating a larger chassis with ample space for ventilation. Many mATX chassis lack the required width to accommodate beefy graphics cards like the RTX 4090 or RTX 4080. These graphics cards are long and chunky, occupying up to 3-4 slots in the case, making it difficult to fit them in any mATX chassis. Therefore, such GPUs are best suited for ATX chassis. You must check the measurements of the graphics card you plan to purchase and choose a compatible chassis accordingly to avoid any inconvenience.

Brand reputation

Lastly, it’s important to take into account the reputation of the graphics card manufacturer. Some well-known brands include Nvidia, AMD, and Intel. It’s essential to choose a reputable brand that offers good customer support and a warranty.

Nvidia versus AMD

At present, Nvidia and AMD dominate the GPU market, with Intel also joining the competition but with less impact. For those seeking top-notch performance and can afford pricey GPUs, Nvidia is the way to go, as they consistently offer the best-performing cards for each generation, including the RTX 4090, RTX 3090, and RTX 2080 ti. Meanwhile, AMD prioritizes price-to-performance products that may fall slightly behind their Nvidia counterparts.

Despite this, AMD has introduced high-performance GPUs over the years, intensifying the rivalry between the two companies. The competition has driven both companies to release better graphics cards, ultimately benefiting gamers by enabling them to play even the most demanding games on ultra settings.

See Also: What Does Ti Mean In A GPU?

Frequently Asked Questions

What is the classification of graphics cards?

Graphics cards can be categorized based on their intended use, level of performance, and physical form factor. Based on purpose, they can be gaming, workstation, or integrated graphics cards. Based on performance, they can be budget, mid-range, or high-end graphics cards. Based on form factor, they can be full-sized, half-sized, or low-profile graphics cards.

What are the tiers of RTX 3000?

The tiers of the RTX 3000 series graphics cards from Nvidia are as follows (in order of performance and price): RTX 3090, RTX 3080 Ti, RTX 3080, RTX 3070 Ti, RTX 3070, and RTX 3060 Ti.

Is RTX 3060 entry level?

The RTX 3060 is not generally considered an entry-level graphics card, but rather a mid-range option. It sits at the lower end of the RTX 3000 series lineup and is designed to provide good performance and features for gamers who want to play modern games at 1080p or 1440p resolutions. 

However, there are other options in Nvidia’s lineup, such as the GTX 1650 or the GTX 1660, that are typically considered entry-level graphics cards due to their lower performance and price point.

Also Read: CPU Hierarchy – PC Processors Tier List

Conclusion

The advancements made in the field of graphics cards over the past few decades have been staggering. The introduction of Nvidia’s GeForce 256 GPU in 1999 sparked a period of rapid innovation, with each subsequent generation of cards boasting improved performance and efficiency. 

To keep track of this evolution, the GPU hierarchy was devised, providing a comprehensive comparison of various graphics chips based on their relative performance. This hierarchy takes into account real-world usage, including testing for frame rates, G-Flops, T-Flops, bandwidth, and rendering. By relying on trusted benchmarks, we can ensure accuracy and reliability in our comparisons. 

In general, the future of graphics cards appears promising, and we can anticipate ongoing advancements and innovation in this domain.

Author:
I am a computer engineer holding a bachelor's degree in Computer Science, complemented by a Master's in Business Administration from University of Strathclyde, Scotland. I currently work as a Senior IT Consultant in Melbourne, Australia. With over 15 years of...