Best Graphics Cards For Gaming. Today’s market is full of many graphics cards to choose from, a situation that can confuse many users when purchasing a new card, so we have prepared this article where we will talk about some of the best options to choose from. by price range. To which we answer the typical questions: Which is the best Graphics Card to Buy in 2022? So we’ve broken down the options available into three ranges: high-end, mid-range, and low-end. Here we go!
The high range corresponds to the most powerful cards which ensure excellent performance by playing at high or maximum graphic detail levels with aggressive filters to improve image quality like MSAA, this will also allow us to play at high resolutions such as 2K or 4K, although in the latter case, it will be necessary to reduce the graphic detail and especially the filters such as antialiasing. Either way, we will get a much better gaming experience than the current generation Xbox One X and PS4 Pro consoles.
Nvidia does a great job with its Turing graphics architecture and has become the undisputed queen of high-end graphics cards. Its new GPUs have shown a level of performance and energy efficiency that is hard to beat for rival AMD, which has focused on the mid-range in its new generation of graphics cards. But AMD has redesigned its graphics architecture and released new models that at least match or closely follow Nvidia’s high-end GPUs.
17 Best Graphics Cards For Top-Notch Gaming Performance 2022
- Gigabyte GeForce RTX 2080 Ti GAMING OC
What about this device designed only for extreme games; It is the most powerful graphics card for today’s games, without any other that eclipses it except the other custom models of this same RTX category. We opted for this model instead of the Asus because the results were a bit better and so was the price. Of course, we look forward to the Asus ROG Matrix which will surely be positioned as the best.
It is an Nvidia Turing graphics core that mounts 4,352 CUDA cores, 544 Tensor cores, and 68 RT with no less than 11 GB of GDDR6 memory running at 14 Gbps. To calm this bug, the brand opted for a WINDFORCE 3X heat sink with many copper heat pipes which are responsible for collecting all the heat and distributing it through a colossal finned exchanger accompanied by three fans. Its length is 290 mm for an approximate weight of 1 Kg.
We have nothing to say about performance in games and benchmarks, if the game is optimized the performance will be available to the maximum and if it is done badly it will already make sure to get the most out of it. The consumption at maximum load is 342 W with 48 W at rest, so recommend a power supply of at least 750 W. For its part, the temperatures supported reached 71 degrees, and 33 in standby mode, so the Gigabyte heater does a great job.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b, 1 USB Type-C
- Measures: 286x114x50 mm
- Occupied pitches: 3 pitches
- Software: Gigabyte
Benefits
- 11 GB GDDR6 dedicated.
- Rear protection plate.
- Windforce 3x cooling system with alternating rotation.
- Gigabyte GeForce RTX 2080 SUPER Gaming OC
Since the NVIDIA RTX 2080 is discontinued, we think the Nvidia RTX 2080 Super is a very competent graphics card for WQHD and 4K resolutions (taking into account that it depends on the title, we will have +60 FPS and in others, we will go a little lower).
This version of Gigabyte features a powerful WINDOFRCE 3X triple-fan heat sink, a black plate that provides rigidity and aesthetics to our graphics card. It only takes up 2.5 slots, and it’s also available in white, which is pretty unusual for a GPU.
As technical characteristics, it has a processor speed of 1770 MHz, 2560 CUDA cores, 8 GB of GDDR6 memory with a 256-bit memory bus, 14 Gbps, and a PCI Express 3.0 interface. Of course, if you find a similar model at a lower price, ask us about it and we will confirm if it is a good choice.
- Video outputs: 3 DisplayPort 1.4 and 1 HDMI 2.0b
- Measures: 286x 114 x 50 cm
- Occupied pitches: 2.5 pitches
- Software: Nvidia Drivers
Benefits
- Core clock at 1830 MHz.
- Integrated with 8GB 256-bit GDDR6 memory interface.
- Windforce 3x cooling system with reciprocating rotary fans.
- RGB Fusion 2.0 – 16.7m customizable color lighting.
- MSI RTX 2070 Super Gaming X Trio
The Nvidia RTX of the Super range is a reality, of which we already know the RTX 2060 Supermodels and this, the RTX 2070 Super which essentially replaces the RTX 2080, which will be discontinued, as you hear it. Only the RTX 2080 Super would be left for later. These cards are essentially a brand update to increase the perks a bit more and jump up a notch in each model.
We have already tested the MSI model on our benchmark, a variant with a custom TRI FROZR three-fan heatsink with TORX FAN 3.0 and ZERO FROZR technology that works wonders. The VRM has also been increased to 8 + 2 phases, compared to the benchmark 7 + 2 models to allow for better overclocking.
This card mounts a variant of the TU104 core, made up of 2560 CUDA, 320 Tensor, and 40 RT cores, for a total of 64 ROPs and 184 TMUs, in short, performance almost identical to the previous RTX 2080. All running at 215W TDP only. The results we got in our review put it above the benchmark model and the RTX 2080, with frame rates above 100 FPS at 1080p and 2K and nearly 60 FPS at 4K.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 327x140x55.6 mm
- Occupied pitches: 3 pitches
- Software: Dragon Center
Benefits
- TORX FAN 3.0 – Award-winning fan design combining two types of blades for cooling and quiet.
- TRI-FROZR THERMAL DESIGN – Using three of the award-winning TORX FAN 3.0 fans, Tri-Frozr is the ultimate in air cooling.
- ZERO FROZR – Eliminates fan noise by turning them off in low-load situations so you can focus on your game.
- RGB MYSTIC LIGHT – Customize LED colors and effects with MSI software and synchronize appearance with other components.
- AERODYNAMIC MASTERY – Using aerodynamic techniques, the radiator is optimized for efficient heat dissipation.
- Asus ROG Strix RTX 2060 Super OC
The second GPU that we find in this renewed range is the RTX 2060 Super, more precisely the Asus version gives us very good feelings both for its aggressive overclocking and for its large triple-fan ROG Strix radiator with AURA RGB lighting.
Asus gave this RTX 2060 Super a good grip over the benchmark model of no less than 180 MHz to put it at 1830 MHz in OC mode. In fact, on its PCB we have a button to switch between basic mode and OC mode quickly and easily. Very important is having an additional 6 pin power connector, in addition to the 8 which has the reference configuration.
Inside we have a variant of the TU106 chipset from the RTX 2070 which will make it virtually identical to this one, despite having a few fewer cores inside, specifically 2,176 CUDA, 272 Tensor, and 34 RT cores. It emerges as a perfect option to play in 1080p and 2K with graphics in ultra and high, with FPS rates above 70 and 80 in 2K, almost nothing.
- Video outputs: 2 DisplayPort 1.4, 2 HDMI 2.0b, 1 USB Type-C
- Measures: 301x131x49 mm
- Occupied pitches: 2.5 pitches
- Software: Asus AURA
Benefits
- RT Hearts: Ray tracing hardware produces a more realistic real-time representation of objects and environments, with shadows, reflections, refractions, and more accurate and natural overall lighting.
- Simultaneous Processing of Floating Numbers and Integers: Turing GPUs handle today’s heavy gaming workloads more efficiently.
- High-Speed GDDR6: Enjoy fast-paced, high-resolution action games with a memory bandwidth of up to 496 GB / s.
- Axial-Tech fans feature longer blades and a locking ring that increases downward air pressure.
- 0dB technology lets you enjoy light gaming with virtually no noise.
- MSI Radeon RX 5700 XT Evoke OC
The AMD Radeon RX 5700 XTs are basically designed to compete with the Nvidia RTX 2060 Super, and therefore with the previous RTX 2070. In them, AMD implemented a new architecture called RDNA with which it improved the speed per Watt consumed by 50 % and the CPI of its cores made in 7nm by 25%.
In addition, MSI offers us a GPU with a factory overclock that reaches up to 1945 MHz in turbo mode, being one of the highest for this card. Along with that, we have a custom dual-fan heatsink that will perform infinitely better than the fan on the benchmark models.
In terms of performance and goals, well you can’t say it did wrong, because at least it’s on par with the RTX 2060 Super, although still a bit far in some IPs from the RTX 2070 Super. Also, the price of this card is a bit cheaper than its rivals, but it has two drawbacks, the first is that they don’t have RT or DLSS, and the second is that it is hot enough if the ventilator RPM is not manually increased from adrenaline. With it, we can play in 1080p and 2K resolutions with high FPS rates without a problem, but the overclocking is quite limited.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 275x98x40 mm
- Occupied pitches: 2.5 pitches
- Software: AMD Adrenalin
- MSI Radeon RX 5700 XT Evoke OC
Benefits
- Torx 3.0 Fan: Award-winning fan design combining two types of blades for cooling and quiet.
- Oc performance: MSI graphics cards come with higher clock speeds for increased performance.
- rDNA Architecture: Designed from the ground up with superior performance and power efficiency, rdna is the architecture that powers AMD’s 7nm gaming GPU, delivering 1.25-per-clock performance compared to previous 14nm processors.
- Msi Afterburner: the overclocking software with advanced control and real-time monitoring.
- Solid backplate: increase the hardness of the board to prevent damage from bending, while complementing the design.
- Gigabyte AMD Radeon RX 5700 XT Gaming OC
One of the always cheaper options is Gigabyte GPUs, a clear example is this 5700 XT. We also offer the version with the WINDFORCE 3X heatsink as one of the ones that offer the best features for this excellent card.
Gigabyte offers us a factory overclocking in turbo mode of 1905 MHz, a little lower than the previous one, but with a good improvement over temperatures. these are two of the most attractive models for this AMD card, which is why they are on the list.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 280x114x50 mm
- Occupied pitches: 2.5 pitches
- Software: AMD Adrenalin
Benefits
- Integrated with 8 GB GDDR6 256 bit memory interface.
- Windforce 3x cooling system with rotating reciprocating fans.
- RGB Fusion 2.0: synchronization with other Aorus devices.
- Metal backplate.
- EVGA GTX 1080 Ti FTW3 Gaming Card
We have also placed this EVGA GTX 1080Ti FTW3 Gaming because of its excellent price, it is one of the very few high-performance GTX 1080 Ti under 1000 euros. The previous generation’s best gaming card still has a lot to say compared to the very expensive RTX 2080, as it’s currently very much on par in terms of performance and FPS and at a slightly lower cost.
The heat sink of this Evga is built in aluminum and copper heat pipes with iCX technology with three fans that collect heat from both the GPU and all power phases to get a true temperature-controlled setting. In addition, it has a temperature monitoring system for each phase and a safety fuse to protect it against overvoltages.
In terms of performance, what interests us is on par with the RTX 2080, so we will have TOP performance at a price below 950 euros. Its 3,584 CUDA cores and 11 GB of 11 Gbps GDDR5X memory, supports resolutions up to 8K with a memory bandwidth of 484 GB / s only exceeded by 2080 Ti.
- Measures: 300x143x50 mm
- Occupied pitches: 2.5 pitches
- Software: EVGA
- EVGA GTX 1080 Ti FTW3 Gaming
Benefits
- Real base clock: 1569 MHz / Real boost clock: 1683 MHz; Memory detail: 11264 MB GDDR5X.
- EVGA iCX Technology – 9 additional temperature sensors to monitor memory and VRM.
- GPU / Memory / RGB PWM LED thermal status indicator.
- New design of ventilated radiator fins and pin fins for optimized airflow.
- Safety fuse to protect components from damage due to improper installation or another component failure.
Mid-Range/Cheap Graphics Cards 2022
We go down a notch in prices and features to find a mid-range that will allow us to play with an excellent level of detail and moderate and high filters to improve the quality and the gaming experience. In this range, we will already get a much better experience than that offered by the new generation consoles.
We recommend cards with 4GB of GDDR5 video memory and of course, the new RTX 2060, which, while still a bit pricey, outperforms a GTX 1070 Ti. We also offered an 8 GB model for its excellent price, although today with Full HD resolutions (1920 x 1080) it is more than sufficient with 2 GB of RAM.
If the high-end is dominated by Nvidia, the situation in the mid-range changes greatly to AMD’s relief. The new AMD Polaris architecture has proven to be able to offer very competitive graphics cards in terms of price and performance, although once again it is Nvidia that takes the performance crown in this range. AMD offers us quite a few alternatives in the segment between 200 and 300 euros for an exquisite performance in 1080p resolution and very capable of handling games at 1440p.
- Asus GeForce GTX 1660 SUPER OC Dual
With the arrival of the Super versions in the mid-range, Nvidia wishes to recover and maintain the benchmark position to which it corresponds. Thanks to the exceptional Radeon RX, Nvidia sees the throne in peril and has put more fuel on the fire with this modified 1660 which increases its performance to be on par with the 1660 Ti and also at a better price.
What really changes is not its GPU, which will remain the 1408-core CUDA TU116, but its VRAM memory. In this case, the level has been upped with 6 GB of GDDR6 type memory running at 14 Gbps, like the rest of the higher GPUs. Plus, the overclocking capabilities it demonstrated in our benchmark puts it very close to the RTX 2060, an impressive job from Nvidia.
- Video outputs: 1 DisplayPort 1.4, 1 HDMI 2.0b, and 1 DVI
- Measures: 240 x 120 x 53 mm
- Occupied pitches: 3 pitches
- Software: Nvidia Drivers
Benefits
- OC Edition: Boost frequency 1860 MHz (OC mode) / 1830 MHz (game mode).
- Axial-tech fans have longer blades and a locking ring that increases air pressure.
- 0dB technology lets you enjoy light gaming with virtually no noise.
- Auto-Extreme technology improves reliability by automating the manufacturing process.
- The 2.7 slot design increases the dissipation area to get the most out of Axial-tech fans.
- Gigabyte GeForce GTX 1660 SUPER OC
And if we are looking for the best option at the best price, we believe that once again Gigabyte gives it to us. That’s essentially 1 FPS of the Asus’ performance, but with slightly higher overclock from the factory and really good temperatures thanks to the triple fan radiator, we can make up for that little bit.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 280 x 116 x 40 mm
- Occupied pitches: 2 pitches
- Software: Gigabyte
- Gigabyte GeForce GTX 1660 SUPER OC
Benefits
- Processor clock frequency: 1830 MHz.
- Graphics card memory type: GDDR6.
- Maximum resolution: 7680 x 4320 pixels.
- Gigabyte GTX 1660 Ti GAMING OC
The new GTX 1660 Ti graphics cards are here, and one of the best in terms of performance and price is this custom card from Gigabyte. One of the most recently tested on our side of the tests and with the impressive three-fan WindForce heatsink it stuck close to quite an RTX 2060.
This card is part of the Turing series, as it includes a GPU of this architecture, although the RT and Tensor cores have been removed, and therefore make it suitable for users who want to enjoy the most demanding games in 2k and 1080p resolutions at good FPS rates without spending a fortune in euros.
It has 6 GB of 12 Gbps GDDR6 memory, 1565 CUDA cores, and 195 bits of memory bus width. This way it can be at the same level as an older generation GTX 1070 and with its excellent overclocking ability to the level of an RTX 2060. Although in its basic form it is slower, of course.
- Video outputs: 2 DisplayPort 1.4, 2 HDMI 2.0b
- Measures: 301 x 132 x 50 mm
- Occupied pitches: 2.5 pitches
- Software: Gigabyte
- Gigabyte GTX 1660 Ti GAMING OC
Benefits
- Breathable leather lining, classic look, and design.
- Sapphire RX 5600 XT Pulse
The first AMD Radeon RX 5600 XTs have already crossed our benchmark. While we don’t like the Sapphire brand very much because buying one of their graphics cards is a lottery, for those who don’t know their most basic models have base heatsinks and a benchmark PCB. The price is very attractive, but at the same time, it can give you dislike.
This Sapphire RX 5600 XT Pulse mounts a dual-fan heat sink which gives optimum results. This is the cheapest version we can find on the market, along with the PowerColor. The latter chose to remove a heat pipe and with the modifications of the BIOS decided to remanufacture the original heatsink. A mess of the first order.
The specs are quite good and its performance is similar to that of the RTX 2060. This graph has a maximum frequency of 1750 MHz. Likewise, we have 6GB GDDR6 memory, three DisplayPort outputs, and one HDMI.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 230 x 98 x 40 mm
- Occupied pitches: 2 pitches
- Software: Sapphire’s own
- Sapphire RX 5600 XT Pulse
- PowerColor Red Devil Radeon RX 5700
The second card launched by AMD with the RDNA architecture is this RX 5700, as the younger sister of the XT version, although with many similarities. For example, we have the same amount of 8GB GDDR6 memory, abandoned for the time being configurations with HBM2. This card could compete directly with the RTX 2070 and RTX 2060 Super, as it was pretty clear in our review that it outperformed the RTX 2060 in FPS rates in the games we tested.
In this case, the chip it uses reaches 1750 MHz in turbo mode, starting from a base of 1465 MHz. The number of transmission processors also decreases compared to the XT version at 2304. The PowerColor version is one of the cheapest that can be found. And it also includes an impressive three-fan game and an aggressive heatsink like few others.
- Video outputs: 3 DisplayPort 1.4, 1 HDMI 2.0b
- Measures: 300 x 132 x 53 mm
- Occupied pitches: 3 pitches
- Software: AMD Catalyst
Benefits
- OC Edition: 8 GB GDDR6.
- 1725 MHz gaming clock.
- Price 4.0.
- Output: HDMI / Displayport x3.
- MSI RTX 2060 Gaming Z Graphic Card
The Nvidia RTX is the best value RTX card, and we have already seen its advantages in our benchmark. This GPU with Turing architecture with only 160 W TDP has no less than 1920 CUDA, 240 Tensor, and 30 RT cores, and implements the new very fast memory of 6 GB GDDR6 of 14 Gbps.
The version offered by MSI against a custom Twin Frozr heatsink with two 14-fin MSI TORX 3.0 fans will only activate when the board goes past 60 degrees. The heat sink is complemented by a sturdy aluminum backplate to completely protect the board. This version is overclocked to give a maximum frequency of 1830 MHz.
As we have seen in our numerous tests and reviews, this is the card that gives us the best performance at a lower price point of this generation, so it’s the best in this mid-range. It ranks on par with the previous GTX 1070 Ti, easily outperforming the Vega 56 and surprisingly close to the RTX 2070, so performance in Ray Tracing, VR, and high resolutions is assured.
- Video outputs: 3 DisplayPort 1.4 and 1 HDMI 2.0b
- Measures: 247 x 129 x 52 mm
- Occupied pitches: 2.5 pitches
- Software: MSI
Benefits
- Award-winning Toex Fan 3.0 fan design combining two different blade types for cool and quiet gaming.
- Aerodynamic control thanks to innovative aerodynamic techniques, the heatsinks are optimized for greater heat dissipation efficiency.
- Zero Frozen eliminates unnecessary noise by turning off fans at low load times, so you can focus on gaming.
- An optimized PCB design with an improved power supply providing a solid foundation for high gaming performance.
- A solid backplate increases the hardness of the board to prevent damage from bending while complementing the design.
- Gigabyte RTX 2060 Gaming OC Pro
For Gigabyte fans, we have the customized version of the RTX 2060. The performance is similar to that of the MSI, and also the price drops a bit, we have the brand’s star radiator, the Gigabyte Windforce with a triple fan. In terms of GPU frequency, we are faced with the same overclocking as the MSI, reaching up to 1830 MHz.
Thanks to this spectacular cooling system, we have a performance that exceeds the MSI configuration of several FPS, so if you want one of the more optimized versions of the RTX 2060, this is the option you can go for. It also has another version with a slightly cheaper dual fan.
- Video outputs: 3 DisplayPort 1.4 and 1 HDMI 2.0b
- Measures: 280x164x40 mm
- Occupied pitches: 2 pitches
- Software: Gigabyte
Benefits
- Powered by GeForce RTX 2060.
- Nvidia Ansel, Gsync, Highlights / DP + HDMI – 14 Gbps memory (1920 cubes).
- Sapphire RX 590 Nitro+ Special Edition
Another of the best midrange GPUs is this RT 590 Nitro, which although it doesn’t have custom versions from other manufacturers, has a nice design with a blue aluminum heatsink with transparent fans with LED lighting.
The price is also one of the big demands since we will get it for less than 280 euros. And wait and see its specs and performance, as we are up against a GPU with 2304 stream processors and a 12nm architecture capable of running at a frequency of 1560MHz. It equips 8 GB of GDDR5 memory at 8400 Mbps and 256-bit bus width. All this moves with a TDP of 250 W.
In terms of performance, we place this RX 590 above the GTX 1060 and very close to the GTX 980 Ti in both 1080p and 2K and 4K, although it is true that where it travels best, it’s 1080p and acceptable. in 2K.
- Video outputs: 2 DisplayPort 1.4, 2 HDMI 2.0b, and 1 DVI-D
- Measures: 260x135x43 mm
- Occupied pitches: location 2.2
- Software: AMD Catalyst
Benefits
- High efficiency.
- 8 GB.
- Asus ROG RX 580 Strix OC Gaming Card
The customized and overclocked RX 580 by Asus is a little more expensive than the RX 590 model, but the efficiency increases and equals the latter. Additionally, we have the well-known Strix heatsink which has done the brand so much good to keep this GPU’s temperatures at bay even when we overclock it.
This RX 580 mounts a Polaris 20 core made up of a total of 33 Computing Units (CUs) that add up to no less than 2304 stream processors, 144 TMUs, and 32 ROPs at a maximum onboard frequency of 1340 MHz. This all moves to a maximum TDP of 222W after our testing.
This model is very close to the RX 590 and is geared towards Full HD resolutions, where it will go very well with all the latest generation games. In 2K resolutions we also had good results, never dropping below 60 FPS, so this is a very good option for narrow pockets.
- Video outputs: 2 DisplayPort 1.4, 2 HDMI 2.0b, and 1 DVI-D
- Measures: 232 x 121 x 36 mm
- Occupied pitches: 2 pitches
- Software: Asus
Benefits
- MaxContact technology doubles the area in contact with the GPU for more efficient temperature transmission.
- Wing-Blade IP5X certified fans ventilate better and last longer.
- ASUS FanConnect II incorporates hybrid connectors to optimize PC cooling.
- Auto-Extreme Technology with Super Alloy Power II offers the best quality and reliability.
- ASUS Aura Sync allows you to customize the RGB LED lighting in your system.
- Asus ROG RX 570 Strix OC Gaming Card
We continue with the one that is the cheapest of this mid-range, and therefore the one that will give us the least advantages. We are talking about less than $250 for a custom model with a Strix radiator with two fans.
Also in the specifications, it is lower than the previous two, although, with its 4 GB of GDDR5 at 7 Gbps, we will have more than enough to play the latest generation titles. This Polaris 20 core is made up of a total of 33 Computing Units (CU) which total no less than 2048 stream processors, 128 TMUs, and 32 ROPs. Consumption is also quite contained at this range because, at maximum factory performance, we get 216W and 49W at rest. Although it is a mid-range, it can also be overclocked, like all the previous ones.
In our results, we had a good full HD and 2K gaming experience, with high-quality games, so it’s not bad at this price point. Of course, in terms of performance, it ranks just below the RX 580 and on par with the RX 480 Strix also from Asus.
- Video outputs: 1 DisplayPort 1.4, 1 HDMI 2.0b, and 2 DVI-D
- Measures: 242 x 129 x 39 mm
- Occupied pitches: 2 pitches
- Software: Asus
Benefits
- MaxContact technology doubles the area in contact with the GPU for more efficient temperature transmission.
- Wing-Blade IP5X certified fans ventilate better and last longer.
- 1310 MHz frequency in OC mode for great performance and gaming experience.
- DirectCU II with Wing-Blade 0dB fans. A patented design that ventilates better and makes it 3 times quieter.
- IP5X dust resistance certification allows fans to last longer.
- ASUS FanConnect II incorporates hybrid connectors to optimize PC cooling.
- Auto-Extreme Technology with Super Alloy Power II offers the best quality and reliability.
Graphics Card Buying Guide 2022
If you are new to a PC graphics card, you will probably be wrong on your first purchase… But with the criteria listed below, you will easily be able to find the best one for PC!
What are a GPU and an APU?
We often talk about “GPU” and “APU” when we talk about graphics cards. But what do these terms really mean and what do they refer to?
A GPU or graphics processing unit is basically a graphics processor. Just as a processor is the central processing unit of our computer, in this case we are dealing with the graphics processing unit of the computer.
A GPU is NOT a graphics card, but rather the chip that is responsible for performing the graphics processing, floating-point, and 3D calculations that form the greatest weight in a game or graphics rendering program.
The GPU can be located on an expansion card, which we acquire independently from our own team, or it can be integrated into the motherboard. In any case, this processing is specially designed for graphics processing, thus freeing our main processor from such complex and heavy calculations.
We will now see the meaning of an APU or accelerated processor. This term was coined by AMD to define processors with a GPU integrated inside the package. This means that in a normal processor or CPU, we will also have in a specific place, another circuit that will be in charge of the 3D processing of the graphics of our computer.
Many current processors have this type of cores embedded in the same silicon, even with an external graphics card. Of course, the graphics processing capacity of an APU is much less than that of a dedicated graphics card.
What must be deduced from this is that having a processor with integrated graphics does not mean that we cannot have a dedicated graphics card, in fact, it is the most normal today, and surely you have an APU yourself. on your PC. Game consoles are the living example of an APU, they contain a processor that will act as both CPU and GPU.
In these paragraphs, the term dedicated card has been mentioned on several occasions. Let’s see what it is.
Difference Between a Dedicated card and an Internal card
Integrated Graphics Card
We start with integrated graphics cards. We have already seen that a processor can have a graphics processor in itself to form an APU. Well, that precisely means having an internal graphics card. In this type of computer, there won’t be a card connected to the PCI slots, but we will have a DisplayPort or HDMI connector coming straight out of our motherboard.
This is very common on laptops, where space is very limited and manufacturers have to achieve maximum integration into the components to make everything fit. We will quickly notice that our laptops have an integrated graphics card if we don’t see an Nvidia or Radeon sticker anywhere, or we go into the device manager and find in the graphics section something similar to “Intel HDxxxx Graphics” or “AMD Embedded”.
Currently, the processors have powerful integrated GPUs, with which we can reproduce the content in 4K, and even play many games, but they will NEVER reach the level of a dedicated graphics card. Also, an integrated graphics card will take some of the RAM for its own use, so we will have less available for normal PC use.
Dedicated Graphics Card
These are the ones that interest us, the ones that are purchased independently and connect to a PCI Express slot. We will notice our computers have one when we see an Nvidia sticker on the outside or a Radeon sticker in some cases. These types of cards have their own high-performance GPUs and are exclusively designed to handle 3D graphics and floating-point operations. In addition, they install their own RAM, called VRAM or GDDR RAM, and it is also much faster than normal RAM.
Most powerful ultrabooks or gaming laptops will almost all have a dedicated graphics card. This does not mean that we can extract it and modify it, because it, even dedicated, will be installed via a chip on the same motherboard as the processor. We will notice it because it will have its own radiator.
What are the pros and cons of having a dedicated graphics card?
Well, for a player it will be practically mandatory to have one. Let’s see what they bring us and what negative aspects they have:
Advantage
- They are much more powerful than an integrated GPU.
- They can be bought and traded whenever we want for a better one.
- It has its own GPU and its own memory.
- We can play any games we want and enable advanced and better filters without our team slowing down.
- They have their own integrated cooling system.
If it’s good, we can play the latest titles in the market, even if our team is old. There are many models, from the most powerful to the most ordinary, and almost all of them will perform better than a built-in card.
Disadvantages
- A bit Expensive
- They consume a lot of energy and we need power supplies over 500 W.
- They put more heat in our box. Either way, the pros are more than the cons, and if you want to play the last one, you’re bound to need it, and that’s why you’re here.
Graphics Specifications: GPU and Architecture
Well, there are several things to consider when purchasing a graphics card. Each of these has a lot of characteristics and numbers to determine which is better or worse, let’s start with your GPU or GPU. We will try to explain it from the technologies currently on the market.
Graphics processors have an infinite number of performance parameters and are also built under different architectures and manufacturers. In this list, we will only be able to use the latest technologies from each manufacturer, as well as the characteristics to be taken into account for each of them.
Turing Architecture (NVidia)
Its market name will be under the term “RTX”. Any graphics card with RTX in its name will be Turing technology, and it’s the latest technology from the brand that gives us the best performing graphics cards today.
If we want to play the latest, in the best quality, at the highest resolution, and with virtual reality, we will need it. Turing’s architecture manufactures processors with 12nm transistors and optimized for Ray Tracing, or Real-Time Ray Tracing, Virtual Reality (VR), and Artificial Intelligence. We are interested in the first two.
Real-time ray tracing means that in next-gen games and the latest titles we’re going to get higher graphics quality than anything we’ve seen before. Greater realism in shadows, reflections in water and soil, dynamic adaptation of the density of particles, to provide the most realistic final result possible. The same can be applied to virtual reality.
In the characteristics of Nvidia RTX processors, we can identify CUDA cores, Tensor cores, and RT cores, as well as the clock frequency of the processor. The higher the number of these cores and the higher the frequency, the higher the performance of the graphics card.
The architecture of the Radeon NAVI 10
This is the most recent technology from AMD, the big news AMD brings to these graphics cards is their architecture, one in which it claims to have completely redesigned the way to handle instructions and the processing of these. by graphic hearts.
Its name is RDNA (remember the previous one was called GCN) and there are two key features for the user: the first, an improvement in GPU IPC (operations per cycle) of up to 25% compared to the previous generation, and the second, an increase in overall performance per watt of up to 50%.
On paper, an RDNA GPU should offer up to 44% better performance than an identical GPU, but a weak GCN. This opens many doors for AMD to create more powerful and efficient cards. But we also have big absences, like real-time ray tracing or deep learning technology like DLSS on Nvidia. Obviously, it will be part of the next generation of games, so this is still a pending topic from AMD.
Pascal Architecture (NVidia)
Pascal is the architecture of the previous generation of Nvidia graphics cards. Even today these are very good graphics cards and located in the low, mid, and high range at the same time. In general, they have a lower cost than the new ones, and if we get a good deal, they will be very interesting as well.
We can easily identify them if the word “GTX” appears in the model and the numbers 1000, for example, 1050, 1060, 1070, and 1080. They also work with all kinds of games at 1080p, 2K, and 4K resolutions.
Polaris RX Architecture (AMD)
These are the previous generation of AMD graphics cards, although they are currently widely used as midrange and low-end components. These are graphics cards with good performance at 1080p and 2K resolutions at a really low price. In fact, they are even a better option than the Radeon Vega, with the transistor manufacturing process being 14nm
We will quickly distinguish them by the distinctive “RX” in their name, and we will always have to watch out for manufacturers with custom models like Asus, as the series models are quite poor and with poor cooling.
Intel HD Graphics
We mention it as a simple anecdote. This Intel HD technology is used by the manufacturer to name the graphics cores that make up its processors. In other words, they are integrated graphics cards and can be distinguished by the name “Intel HDxxxx” in the device manager.
It doesn’t have dedicated graphics cards, they cater for low-usage mid-range laptops for gaming and therefore don’t deliver performance to match an optimal gaming experience.
Read More: Best Budget Graphics Cards 2021
Specifications of a Graphics Card: Memory, Quantity, and Bus Width
Another fundamental aspect of a graphics card is the memory it has and of course, the width of the bus, the ability to store the game’s graphics data, the rendered environments, and the transfer capacity will depend on it.
Currently, the different architectures that we have seen mainly use three types of graphics memories, which will need to be known and known to evaluate their properties.
GDDR6 Memory
It is the fastest memory that exists today, but also the most expensive to produce. It is implemented by Nvidia’s Turing architecture graphics cards, and this is one of the main reasons the products have become more expensive compared to the previous generation.
This GDDR6 memory is capable of reaching a speed of no less than 14 Gbps. On almost all occasions, we usually find this nomenclature instead of the traditional GHz to define the speed of the memories.
HBM2 Memory
This memory is the main novelty of AMD’s Vega architecture, and while it doesn’t have speeds as high as GDDR, it has a much larger bus or interfaces width, reaching up to 2048. bits. Its speed is around 1.9 Gbps. Let’s say they provide raw speed for the bus width and transfer capacity. For convenience, they have a bandwidth similar to GDDR6 memories.
GDDR5 and GDDR5X Memory
GDDR5X memory is an evolution of the classic GDDR5 which is only implemented by Nvidia’s high-end GTX models of the previous architecture, namely the GTX 1080 and 1080 Ti. The speed of this memory reaches up to 10 Gbps.
GDDR5 memory is present in both Pascal generation cards and AMD Polaris RX cards, and we can find it with speeds ranging from 6 Gbps to 8 Gbps, of course, the more the better. is.
Amount of Memory
Regardless of the type of memory, we have the amount installed in the graphics card. 2 GB is not the same as 8, far from it. The more memory we have installed on a graphics card, the more graphics data we can store on it. In operation it is the same as in RAM, if it is small and fills up we will have a slower game and as we walk around the world we will notice an inconsistency in the objects around us, appearing suddenly or disappearing.
Viewing distance also influences, especially in open-world games, the more memory there is, the more distance will be represented in the world and the farther objects we will see. For good graphics to work properly in all games, you will need at least 4 GB of installed memory.
Memory Bus width and Bandwidth
The width of the memory bus represents the number of bits that can be transmitted and is measured in bits. This is called a word, the instruction that is sent from memory to the processor, the longer the word we can send, the larger the bus will be, and therefore the greater the processing capacity of the instructions we will have will be great.
Currently, graphics cards have a bus width of between 192 bits and 2048 bits that we have seen in HBM2 memories. The larger the width of the bus, the better, but always taking into account the bandwidth that we will end up with.
Memory bandwidth is the amount of information that can be transferred per unit of time and is measured in GB / s. The greater the width of the bus and the greater the frequency of the memory, the more bandwidth we will have. And that is why HBM2 memory has a final bandwidth similar to GDDR6.
While the HBM2 runs at 1.9 Gbps and 2048-bit bus width on the RX Vega 64, you get a total bandwidth of 483.8 Gbps.
The second has a bus width of up to 352 bits and 14 Gbps on the RTX 2080 Ti and achieves a bandwidth of 616 Gb/s. This means that the width of the bus is as important as the speed of the memory.
Power Connectors
The connectivity of a graphics card is very important, and of course the power connector it has, because the higher the TDP (Thermal Design Power) or the amount of Watts it consumes, the more connectors we will need.
Power Cable
We will start with the power connector that our graphics card will have. In general, one can find several types or, rather, quantities of connectors. Fortunately, all graphics cards will operate at the same input voltage, so in that sense, we won’t have any issues with the connectors on our power supply. Types we can find:
6 Pin Connector: This is the basic connector and every mid/high-end graphics card will have at least one. It is simply a cable with two rows of 3 pins. All sources have at least one.
6 + 2 pins: in addition to the 6 above, there will be two more, forming two rows of 4 connectors. Likewise, any power supply worthy of the name will unite these two pins with the 6 others in a removable manner.
8 + 6: we now turn to cards that have a TDP greater than 160W. In addition to the 8 pin connector (6 + 2), we will find another one, with another 6 pins.
8 + 8: Finally, the cards with the highest consumption and more than 200W will provide the complete set, which will be composed of two 8-pin connectors. Power sources over 500W should have two of these connectors, preferably separately on separate cables.
8 + 8 + 8: These are already just special configurations with integrated liquid cooling or a ridiculously exclusive board like the MSI Carbon
Very small graphics cards will not have connectors and power from the PCIe slot will suffice.
Graphics card Media Connection Ports
We are now turning to connectivity in terms of multimedia ports, which is becoming increasingly important for high-resolution monitors and virtual reality glasses. Next, we’ll go over which connectors we can find on a board and which ones we’ll need depending on the monitor we have.
HDMI Connector
High Definition Multimedia Interface is a communication standard for multimedia devices with uncompressed sound and pictures. It is an elongated connector with two grimaces at the ends. We have different sizes, HDMI, Mini HDMI, and Micro HDMI. We are interested in an HDMI connector and much more in the HDMI version it brings.
The HDMI version will influence the image capacity we can get from the graphics card. The latest version is HDMI 2.1, which will allow us to connect monitors with resolutions up to 10K and reproduce 4K at 120Hz and 8K at 60Hz.
Most cards come with HDMI 2.0b, which allows us to connect 4K monitors at 60Hz and dynamic sync. A self-respecting graphics card must bring at least one if we have a monitor with this type of interface.
DisplayPort Connector
It’s a very similar HDMI connector, but with only a scowl on one side. As before, the version of this port will be very important, and we will need it at least 1.4 since this version supports playback of content in 8K at 60Hz and 4K at 120Hz.
If we have a high-performance monitor, it will surely have such a connector, and you will need the graphics to have it as well to get the most out of our equipment.
DVI Connector
This interface is unlikely to be found on current monitors, although graphics such as the RTX 2060 still have one. There are different versions of the DVI connector, although the most common today is DVI-D. It has a 24-pin connector and a flat horizontal plate that serves as a ground. It supports resolutions up to 4K, but it is not recommended if we have any of the previous connectors.
USB Type C Connector
This is one of the new additions to the connectivity of next-generation graphics cards. This connector is going to be very important from now on, especially for laptops and virtual reality devices.
This USB stick has DisplayPort Alternate Mode, which is nothing more than the functionality of DisplayPort 1.3, with support for displaying images at 4K resolution at 60Hz. So this port will be very interesting for ultra-thin laptops that don’t have one DisplayPort connector and we want to get an external monitor with that interface.
But it doesn’t end there, another of the great uses of this port is to provide a connection for virtual reality glasses, as these normally provide this type of connectivity today. Especially those from Nvidia with VirtualLink. So if we want to use the graphics card for VR, it will be better to have this port.
See More: Best NVidia Graphics Cards 2021
Size of a Graphics Card: Length and Locations It Occupies
Another section that we need to consider is graphics card measurements, as there are chassis that certain card configurations do not fit. So we always have to look at the width, length, and height of the card, and compare it with the specifications of our chassis, or take a meter straight and measure it ourselves.
Almost every graphics card will have its own measurements, and it is very difficult to categorize them into standard measurements. If we do, we’ll basically have three types:
Extended or ATX Size: These cards are the longest configuration of the three, and will almost always be over 220mm long and can go up to 300mm or even longer. We will have to pay particular attention to these measurements and those of our frames. They will be easy to identify as they almost always have three fans.
Normal Size: These are cards that will measure a maximum of 220mm in length and will fit virtually any chassis. These are almost always a dual fan or turbine cards.
Compact size or ITX: these are the smallest of all, although they are no less powerful for that. It is normal to find one of these versions in each model, with a little less power than the long configuration. Their measurements will be approximately 120mm wide by 150mm long or less and are intended for small ITX towers.
Another very important factor is the height of the cards, as the current trend is to make them bigger and bigger, with bigger heatsinks and taking up more space. This space can be measured by slots or expansion slots. We all know what it is. The higher it is, the more unused slots there will be on our motherboard.
1 Slot: almost forgetting, for a graph to occupy only one location, it should only be 2 cm high and it is very rare to find one that works for something.
2 slots: equivalent to a height of 4 cm or 40 mm, and yes we can find many of them limited to this height.
3 Slots: This makes a height of over 40mm, reaching as high as 54 in some models, and more than we’ll see soon.
Types of heatsinks and which one is best
One very important item that many gamers overlook is the graphics card heat sink. Graphics cards are components with processors that require a large amount of energy for the enormous mass of information they are capable of processing, and this of course generates a lot of heat inside. This is why having a good radiator is very important, even more than that of a CPU. There are different types of GPU heat sinks:
Custom or vertical heat sink
These heatsinks are by far the most efficient. They are composed of a copper base that is in contact with the GPU and the power phases. Certain heat pipes integrated into this base, are responsible for capturing all the heat from it and distributing it between a finned heat exchanger which is installed over the entire extension of the PCB of the card, or even exceeding this measurement and extending to the -of the.
These heat pipes can use a steam chamber to optimize the distribution and speed of thermal transmission. This block will be constructed of aluminum or copper and will be bathed by a vertical airflow generated by one, two, or up to three fans installed in a cover above this block.
The result will be a cage that completely covers the PCB of the board to facilitate heat dissipation of the components. In addition, in the upper area of the plate, one can also have a backplate which also covers this entire area.
Blower type heat sink
The design of these heatsinks is older and rare to see on current models because the airflow is poorer and the dissipation much less optimal. These heat sinks have a copper core that comes in contact with the GPU and VRM to pick up heat and distribute it to the top area where a finned block will be located. This block can also contain a steam chamber so that the heat transfer will be of better quality.
Well, this block will be bathed by an axial flow of air coming from a centrifugal fan (it captures the air vertically and expels it horizontally). All of this will be covered with an outer cover to prevent airflow from dispersing before passing through the fins.
Liquid refrigeration
In some cases, we also have liquid cooling settings for custom graphics cards. The operation is the same as in the case of CPUs, a block is installed in the GPU where a liquid will circulate in a closed circuit. This will reach an exchanger with fans to transfer the heat from the liquid to the finned block where it will ultimately be transferred to the air.
Passive heat sink
Well, it’s basically an aluminum or copper fin block that is installed on the GPU and dissipates heat without a fan. Of course, this is the worst, and only low-end or entry-level graphics cards install it.
Final words
Without a doubt, the field of graphics cards is one of the most complex in terms of choosing the right models. There are a lot of them and it’s hard to find the best of them, and it’s also one of the most sought after by gamers. We believe this model list covers most of the needs of low, mid, and high-end gamers well.
In general, these are great graphics cards with great features that will allow us to play pretty much anything today. Of course, keeping in mind the limitations of the low range, but it is precisely for this reason that it is called low range. Would you like to put another model on this list that caught your attention?
