The graphics card should be your primary concern when purchasing parts for a new gaming computer. After all, it’s responsible for the bulk of the game’s visuals, so it stands to reason that it would be high on your list of priorities.
- Sara Build Guide For Genshin Impact. Is Sara Worth It? Update 04/2024
- Pokemon Sword and Shield Guide: How To Use Evolution Stones? Update 04/2024
- How To Get New Passengers In Spiritfarer? Update 04/2024
- Anthem: Storm Javelin Best Loadout – Ultimate Guide Update 04/2024
- Ooblets Release Date, News, Trailer and Rumors Update 04/2024
However, there are many aspects to consider when shopping for a graphics card, from performance to compatibility, as is the case with any piece of computer hardware.
Bạn đang xem: How To Pick A Graphics Card Update 04/2024
If you think you fit into this category of PC builders, read on; we’ll cover the most frequently asked questions here.
What Is a Graphics Card?
Therefore, the graphics processing unit (GPU) is what people mean when they talk about a “graphics card.” The graphics card, like the motherboard, is a printed circuit board. In the case of discrete, or stand-alone, GPUs, these will also include fans, on-board RAM, a memory controller, a BIOS, and other features.
Even though there is a wide variety of graphics cards available, there are essentially only two kinds:
- Integrated graphics processing units (GPUs) share a single package with a central processing unit (CPU) or system on a chip (SoC). Most Intel CPUs have built-in graphics, while AMD CPUs are more hit-or-miss in this regard. You can use the integrated graphics for light gaming, surfing the web, checking email, and maybe even watching some videos. They also consume less energy than a dedicated graphics processing unit.
- A discrete graphics processing unit (GPU) is one that is not integrated into the central processing unit (CPU), but rather is plugged into an available expansion slot on the motherboard. Discreet graphics processing units (GPUs) are superior to integrated GPUs when it comes to performance in computationally demanding applications like high-end gaming, video editing, 3D model rendering, and so on. Some of today’s graphics processing units (GPUs) can eat up to hundreds of watts of power.
In most cases, a modern discrete GPU will outperform an older integrated GPU, but this depends on the generation of both the CPU and the GPU. The discrete GPU will be superior to integrated graphics if both pieces of hardware were made around the same time. For one thing, it has more available cooling and processing power to handle complex tasks.
What Components Does a Graphics Card Have?
While each model of graphics card has slightly different internals, most contemporary discrete GPUs include the following:
- Similar to a CPU, the GPU is a real piece of hardware.
- Memory: The graphics card has its own RAM specifically for facilitating operations; this RAM is called Video RAM (VRAM).
- PCI Express is typically located on the card’s underside and serves as the interface for graphics processing units.
- Video outputs include HDMI, DisplayPort, Digital Visual Interface (DVI), and Video Graphics Array (VGA).
- GPUs have fans and a heat sink to dissipate any excess heat that may accumulate while in use.
- Connectors for Electrical Current: Most current-generation graphics processing units (GPUs) need a six- or eight-pin power connector, and some may need as many as two or three.
- The Central Processing Unit (CPU) of the BIOS When you turn off your computer, the BIOS saves your settings and programs, including the settings for the voltages and memory.
How Does a Graphics Card Work?
Images, such as photos, videos, games, documents, your regular desktop environment, a file folder, and anything else, are rendered on a display by a graphics card. Video games, which require a lot of processing power, and even “simple” tasks like creating a new text document all necessitate the use of a graphics card.
To elaborate a little, your computer’s graphics card converts the data and instructions sent by your operating system and other applications into an image that can be displayed on your monitor. But a modern graphics card can process an incredible number of instructions at once, drawing and redrawing images tens or even hundreds of times per second to keep whatever you’re looking at or doing running smoothly.
Therefore, the central processing unit communicates with the graphics processing unit (GPU) to relay screen-related data. The graphics card then processes these directives locally, rapidly updating the data stored in its visual random access memory (VRAM) to reflect the current state of the screen’s pixels and the adjustments that need to be made to them. The data is then transmitted quickly from your graphics card to your monitor (via a cable, of course), where the images, lines, textures, lighting, and shading are altered.
If done correctly and without forcing the graphics card or other computer components to do things beyond their capabilities, it can look like magic. This is a very simplistic explanation. Though that only scratches the surface of the inner workings of a graphics card, hopefully you now have a better idea of how they function.
Who Makes Graphics Cards?
For graphics processing units, you can choose between AMD and Nvidia. Both AMD and Nvidia are household names in the graphics processing unit (GPU) industry, and while CPU juggernaut Intel has entered the fray with its Arc Alchemist graphics cards, it is AMD and Nvidia whose names are most often heard. It’s true that AMD and Nvidia don’t actually “make” graphics cards, but that’s not the whole story. They create the boards and either manufacture them themselves or license the design to companies like MSI, ASUS, Zotac, Palit, and so on who then produce the boards themselves.
However, we are not here to argue, so let’s focus on the two major graphics card manufacturers, AMD and Nvidia. One of these two designs will be what you see if you go to a website today and try to buy a new graphics card. Which GPU you end up buying raises the question: does it matter?
GPU vs Graphics Card – What’s The Difference?
Before anything else, you should know that although the terms “GPU” and “graphics card” are often used seemingly interchangeably, they are actually quite different.
What is the distinction, anyway?
Actually, “graphics processing unit” (GPU) is shorthand for the actual graphics chip. Nvidia and AMD are the two most popular GPU manufacturers right now.
However, the graphics card is more than just the chip that does the rendering. In fact, it incorporates not only the central processing unit and memory but also the cooler, the external shell, and the connectors.
Currently, consumers can choose between purchasing graphics cards from either Nvidia or AMD. The vast majority of graphics cards, however, are manufactured by third-party companies like Asus, Gigabyte, MSI, EVGA, Sapphire, and others.
However, while a graphics card’s PCB and cooling system can be customized by the manufacturer, the GPU itself cannot be changed.
Now that we’ve established this distinction, we can move on to the meat of the discussion: settling on the best graphics card.
Finding The Right GPU
Performance
What resolution do you plan on playing games at, and what framerate are you shooting for? These are the two most important factors to consider when selecting a graphics processing unit (GPU).
The more pixels (represented by the resolution) that are displayed on screen, the more lifelike and detailed the game will look. The most common video game resolutions in 2022 are 1080p (Full High Definition), 1440p (Quad High Definition), and 2160p (4K), but higher resolutions necessitate more powerful hardware to play. Thus, higher resolutions will result in lower framerates.
Most modern mid-range GPUs are capable of running games at QHD at 60 FPS and 4K at 30-60 FPS. However, depending on the game’s requirements and optimization, performance can vary greatly. The good news is that you can easily find out how fast a given GPU will run a given game by doing a simple Google search.
The framerate, also known as frames-per-second (FPS), is a measurement of how frequently a given sequence of images is rendered and sent to the screen. It’s easy to see why some players (especially those who focus on competitive multiplayer) prioritize performance over visual quality, since more frames equals a more fluid and responsive gaming experience.
Thus, while the majority of gamers would be content with a consistent frame rate of 60 FPS, those who want the absolute best experience can benefit from monitors with even higher refresh rates. Most standard monitors and televisions can only display 60 frames per second, so if you want to see more you’ll need to spring for a higher refresh rate gaming monitor, such as a 144 Hz or 240 Hz display.
Check out our monitor buying guide for a more in-depth analysis of displays, including additional questions answered and excellent recommendations if you’re in the market for a monitor yourself.
VRR
Also related to screens is the issue of virtual reality reproduction.
What is VRR (variable refresh rate), and how does it work, since it depends on the monitor just like framerates do?
High-refresh-rate monitors use VRR technology to prevent the tearing of the screen and stuttering that can be distracting when playing video games. Monitors designed for gamers now typically feature either AMD FreeSync or Nvidia G-Sync.
Both AMD and Nvidia’s technologies have advantages and disadvantages. The complete article on the subject can be found here; however, I’ll summarize the main points for you:
- Since FreeSync is open-source and inexpensive to implement, it can be found in a wider variety of monitors, including some remarkably affordable ones. It’s important to note that the frequency range supported by FreeSync monitors is quite narrow, and that AMD’s site details which models support which ranges.
- Because G-Sync is a proprietary Nvidia technology and more expensive to implement than FreeSync, G-Sync monitors are typically more expensive than FreeSync displays. They make up for it with improved reliability and a few extra bells and whistles, though.
The issue of compatibility also arises. Until recently, only AMD GPUs could use FreeSync, while Nvidia GPUs could use G-Sync. There is more nuance to the situation now that some FreeSync monitors have been labeled “G-Sync Compatible,” meaning that they can support adaptive sync when coupled with an Nvidia GPU.
It is true that G-Sync is currently incompatible with AMD GPUs, but there is speculation that this will change in the near future as Nvidia is said to be working to make G-Sync more widely available to consumers who don’t own Nvidia GPUs.
Real-Time Ray Tracing
Every modern discussion of graphics will inevitably touch on real-time ray tracing as one of the most exciting new features. Why is this useful, exactly?
To achieve photorealistic lighting and reflections, it basically follows the virtual light rays as they travel through the game world. It was a gimmick until recently because it was only available on Nvidia RTX GPUs and had significant drawbacks. Since not all games properly implemented it, it was often very demanding and could cause significant drops in frame rate.
The new Xbox and PlayStation consoles use RDNA2-based GPUs, and now that AMD has added ray tracing to their arsenal with the new RX 6000 lineup, ray tracing is becoming a much more accessible, more mainstream feature.
Developers will eventually shift their attention to ray tracing, but Nvidia’s DLSS still gives it a leg up in the area.
Finding The Right Graphics Card
After deciding on a graphics processing unit, the next step is to choose the card itself. Now, as was previously mentioned, while Nvidia and AMD produce the chips, i.e. the GPUs themselves, their partners produce and sell the vast majority of graphics cards available today.
And here we’ll go over all the ways in which cards can differ from one another.
Design
The card’s visual appeal and design come first. Of course, if you’re more of a “content over form” gamer, you don’t give a hoot about any of this. With the advent of translucent cases, however, it is easy to see why hardware companies now prioritize design over function.
A graphics card’s design typically refers to its shroud, backplate (if present), and RGB lighting.
Of course, companies want their products to stand out from the crowd, so various graphics cards feature unique aesthetic choices meant to do just that.
In the past, it was common practice for companies to brand their products with distinctive colors, such as the orange accents on Gigabyte cards or the yellow ones on Zotac ones. To ensure their cards look good in as many PCs as possible in 2022, most manufacturers went with a more aesthetically generic design.
Naturally, that doesn’t mean the hue has vanished entirely. Colors other than red, green, and blue (RGB) are used instead, which are still vastly preferable to static decals. It’s easy to see why RGB lighting is so popular, not just on graphics cards but on a wide variety of other components and gaming accessories as well, given how easily it can be customized to match any aesthetic.
Backplates, which appear to be as common as RGB lighting on modern graphics cards, beg the question: why?
Backplates serve two main purposes: they support the card so that the PCB doesn’t get bent from the cooler’s weight, and they make the back of the card easier to clean. However, the main draw of backplates is their aesthetic appeal, as they are much more pleasing to the eye than a cluttered PCB.
Also, some backplates include thermal pads that allow them to function as heatspreaders, but most backplates don’t really make much of a difference in cooling.
Size
When looking for a new graphics card, one thing you’ll notice is that they come in a variety of sizes. Some appear quite small, while others are enormous blocks that make regular-sized graphics cards look puny.
What gives with that, then?
As we’ve established, though, it’s the manufacturer who designs the PCB, and there are several compelling reasons why “mini” graphics cards exist.
Xem thêm : Overwatch Torbjorn Guide. General Tips For Playing Torbjorn Update 04/2024
First, they are less expensive because they are easier to produce, and second, they are suitable for use in compact form factors and external GPU enclosures.
However, there is a catch with compact graphics cards: they require more energy to cool.
Because of their smaller form factor, compact graphics cards naturally feature smaller heatsinks and are typically cooled by a single fan. It causes the temperature to rise, the volume to increase, and the ability to overclock to be severely restricted.
To the contrary, the larger cards typically have a larger heatsink, which improves cooling efficiency and allows the manufacturer to install a triple-fan cooler (more on this below).
If you’re on a tight budget or you’re looking for a graphics card to fit in a Mini ITX case or use as an external GPU, a compact model may be the way to go. However, if you can accommodate its size, a larger card will likely have superior cooling and produce less noise.
Cooling
Let’s dive deeper into the cooling now that we’ve briefly mentioned it.
Open air, blower, and liquid cooling are the three primary methods used in today’s graphics cards.
The vast majority of gaming computers benefit most from open-air cooling. An open heatsink and one to three fans are used to circulate air over the chip on graphics cards cooled in this way.
As was previously mentioned, a graphics card may only need a single fan to operate at safe temperatures, but this solution comes at the expense of increased noise. Unless you need a particularly small card or are desperately trying to save money, a dual- or triple-fan card is likely to be the best option.
The use of blower cooling for graphics cards is not as common as it once was, and this is not without good reason. These cards use a closed heatsink and are ventilated by a single fan.
Blower-cooled cards tend to run both significantly louder and significantly hotter than those equipped with an open-air cooler, so it’s understandable why they aren’t as popular as those with an open-air cooler, but they can be useful for preventing heat buildup inside the case, which is helpful for smaller cases or workstations with multiple GPUs.
Similarly to blowers, liquid coolers are not widely used, but for a different reason. They are extremely expensive and should only be used with high-end GPUs if you plan to overclock the card and truly push the hardware to its limits. Despite their potential for unprecedented cooling efficiency (and lower noise generation), they are still prohibitively expensive.
Overclocking
Since we’ve already broached the subject, it’s only fair that we go into greater depth on overclocking. Just how crucial is it to overclock GPUs, then?
To put it bluntly, not very, at least not for your everyday gamer.
The reason for this is that it is impossible to achieve appreciable gains in performance by manipulating the settings of a graphics card in this way.
The clock speed of a typical air-cooled graphics card can only be pushed so far, usually by around 5-10%. This is in contrast to high-end graphics cards with quality liquid cooling, which can yield very noticeable performance improvements. Especially on low- and mid-end graphics processing units, the performance boost you’ll see from doing this is minimal.
Having this information in hand, you may be wondering why it’s important to care if the card has a good cooler or not.
The solution to this question is equally elementary: a better cooler on your graphics card will reduce noise, keep your system running cooler under load, and extend the life of your hardware.
Memory
As was previously mentioned, all modern graphics cards feature their own dedicated video memory (VRAM).
Right now, the only thing you need to worry about when it comes to video memory is its capacity. The standard VRAM capacity for modern graphics cards is 4, 6, or 8 GB; while 4 GB is sufficient for 1080p gaming, 6 GB or 8 GB is required for high-resolution textures or higher resolutions, such as 1440p or 4K.
Unless you’re really strapped for cash and considering getting an older model to save money, it’s probably not a good idea if you want a more future-proof build and/or intend to game in a high resolution in 2022, as the latest mainstream graphics cards come with GDDR6 memory that offers roughly twice the data transfer rate that GDDR5 did.
Connectivity
What should you pay attention to in this regard, given that the manufacturers choose which ports the graphics card will have and how many of each?
HDMI, DisplayPort, and Dual Link DVI-D are the three most common types of connections found on modern graphics cards.
DVI is the oldest connector here, and its newest iteration, Dual Link DVI-D, is still used by some monitors and graphics cards because it supports both 1080p and 1440p at a refresh rate of 60 Hz. However, despite the fact that it will still have some use in the year 2022, it is not a top pick.
Most modern televisions and computer monitors use the HDMI connection, and both HDMI 1.4 and HDMI 2.0 are in common use. When comparing the two, what is the main distinction?
The primary distinctions lie in the maximum frame rates and resolutions that are supported; for example, HDMI 1.4 only allows for 1080p at 144 Hz, 1440p at 75 Hz, and 4K at up to 30 Hz, while HDMI 2.0 allows for 1080p at 240 Hz, 1440p at 144 Hz, and 4K at 60 Hz. Additionally, HDR support was added to HDMI 2.0a and HDMI 2.0b, with the latter being the version of the port typically found in modern hardware.
While HDMI 2.0 is compatible with older HDMI 1.4 devices, you’ll need an HDMI 2.0 port on both your graphics card and your monitor to take advantage of the newer technology’s enhancements.
Then there’s DisplayPort, which is widely used in current gaming monitors and the one you’ll probably end up with.
DisplayPort, like HDMI, has two different versions available now: 1.2 and 1.4.
Incredibly, the older 1.2 port also supports 4K at up to 75 Hz, along with 1080p at 240 Hz and 1440p at 144 Hz. Newer 1.4 ports go even further, allowing for 1440p at a whopping 240 Hz and 4K at a respectable 120 Hz.
Despite the focus on high refresh rates, DisplayPort 1.4 is also the only technology that supports Nvidia G-Sync and has support for high dynamic range (HDR), unlike DisplayPort 1.2.
DisplayPort is backwards compatible in the same way that HDMI is, but the graphics card and the monitor must have the corresponding ports in order to use the full capabilities of the newer version.
Finally, USB-C should be mentioned because some high-end graphics cards now also include this increasingly popular port, which is important to keep in mind if you plan on using a virtual reality headset on your PC.
The best graphics cards in 2022
1. Nvidia GeForce RTX 3080
SPECIFICATIONS
The number of cores in a CUDA GPU is 8,704.
Starter frequency of 1,440 megahertz
Boost frequency of 1,710 MHz
29.76 TFLOPs
In terms of storage, we’re talking about 10GB of GDDR6X here.
Clock rate of memory: 19GT/s
Bandwidth of memory: 760GB/s
REASONS TO BUY
Astonishing performance when compared to prior generations
Makes the 2080 Ti look like a mid-range card.
Now that ray tracing isn’t a hindrance,
REASONS TO AVOID
Requires a powerful power supply unit
in great demand
Although Nvidia’s new RTX 3080 graphics card is currently as hard to come by as pigeon eyelashes, there is no doubt that it is the best graphics card currently available. In comparison to the preceding RTX 20-series, it’s a significant performance upgrade of a generation. To put that in perspective, this nominally $699 card can not only match but massively outperform the $1,200 RTX 2080 Ti, which is impressive when compared to either the RTX 2080 or 2080 Super.
The improvement in ray-tracing performance is the most striking result of our tests. Previous ray tracing-capable graphics cards had such a severe impact on frame rate that most players avoided using it, but that is no longer the case with the latest generation.
You know you’re dealing with something new when you can achieve ray-traced performance that outpaces what you’d get from the flagship card in the RTX 20-series when not using it. Furthermore, Crysis(opens in new tab) can be played on an RTX 3080.
To accomplish this, Nvidia has updated the Tensor Cores (for more DLSS goodness) and the RT Cores (to make with the ray-traced pretty pictures) in this 8nm GPU, in addition to adding a whole lot more CUDA cores.
However, the RTX 3080 is currently the most sought-after graphics card on the market despite its higher power consumption (at least an 850W PSU is required) and difficulty of acquisition. Which is probably why obtaining it is so difficult.
If I had to give anyone advice about purchasing a GeForce RTX 3080, it would be this: if at all possible, you should aim to acquire the Founders Edition. The RTX 3080 Founders Edition is fantastic in any way you want to phrase it. By virtue of its cooler and reworked PCB, the Ampere RTX 3080 Xtreme is the pinnacle realization of the RTX 3080, making it significantly more than just the reference version of the ‘flagship’ Ampere graphics card. And I’m afraid you’ll feel a little hard done by if you wind up with another version of this fine-ass GPU.
Your best bet is to learn to improvise. However, third-party manufacturers such as Asus, MSI, Gigabyte, Palit, and Colorful all produce high volumes of their own versions of similar cards, making the Founders Edition cards scarce. Given the current RTX 3080 supply constraints, you should be satisfied with any model you manage to acquire. Most notably if you can find one for close to the suggested retail price. It’s not all bad, though; some of the cooling features on these aftermarket cards are quite impressive in their own right.
The performance boost over the previous generation is massive and, pretty much, unprecedented, regardless of which card you end up with. Perhaps we have grown accustomed to smaller generational changes, such as that between the GTX 1080 Ti and the RTX 2080. It bears repeating that the $699 RTX 3080 is a massive improvement over the $1,200 RTX 2080 Ti, but we’ll keep banging that Ti drum.
With the RTX 3080, Nvidia has gone all out, and the result is a phenomenal gaming card that revolutionizes the performance of high-end 4K gaming and ray tracing. We could all benefit from having access to the best GPU of this generation if only Nvidia and Samsung would produce more of them.
2. AMD Radeon RX 6800 XT
SPECIFICATIONS
Number of ribosomal DNA (rDNA) cores: 4,608
The initial clock speed is 1825 MHz.
Peak frequency: 2,250MHz
Quantified in terms of TFLOPs, the number is 20.74.
RAM: 16GB GDDR6
A 16Gbits-per-second memory clock
Storage transfer rate of 512 GB/s
REASONS TO BUY
4K Superb
More affordable than an RTX 3080
REASONS TO AVOID
Average ray tracing speed
Contrast with the 3080, which is faster at 2K, but not 4K.
The RX 6800 XT is among the best red team alternatives to Nvidia’s high-end graphics cards. With such a small performance gap between it and the RTX 3080, it’s safe to say that any PC gamer with 4K aspirations would be wise to give this card serious consideration.
The RX 6800 XT was the first of AMD’s RDNA 2 GPUs to hit the market, and while we’ve had plenty of other cards since, this one is the brightest and most economically sensible. If you consider the MSRP, at least, and considering that AMD’s GPUs are expected to return to reasonable price levels before Nvidia’s, the red team may actually have an advantage in the near future.
Nvidia and AMD’s main battleground this generation has been memory, in terms of both bandwidth and capacity. A total of 512GB/s of bandwidth is available on the RX 6800 XT thanks to its 16GB of GDDR6 memory and 256-bit bus. That puts AMD on equal footing with Nvidia’s 10GB RTX 3080 in terms of capacity, but behind in raw bandwidth, which is measured at 760GB/s for the RTX 3080.
In terms of throughput, AMD has an ace up its sleeve in the form of its Infinity Cache, which significantly increases the card’s ‘effective bandwidth. An estimated 1,664GB/s, an increase of 3.25 times the raw bandwidth of the RX 6800 XT. When discussing video games, this means that similar performance can be expected despite the vast technological differences between the various platforms.
It’s a toss-up between the RX 6800 XT and the RTX 3080, but Nvidia just barely edges out AMD with some slick finishing touches. Compared to the RTX 3080, the RX 6800 XT saves you fifty dollars while providing superior 4K performance and significantly more video memory. One could reasonably argue that the additional $50 spent on the RTX 3080 is money well spent, given the benefits of higher 4K performance, enhanced ray tracing, and DLSS. All of these options are readily available now and have a lengthy history of developer support.
Nonetheless, AMD’s FidelityFX Super Resolution has gained a lot of traction among developers, and it provides respectable upscaling that is well worth enabling in games that are compatible with it. In addition, Deathloop’s (opens in new tab) implementation of FSR 2.0 hints at exciting times ahead.
With the RX 6800 XT, AMD is in an excellent position to continue growing; the company now has a compelling offering for gamers right at launch and has the potential to attract the attention of the entire industry. It comes as no surprise that the collaboration between Zen and RDNA engineers also contributed to this success.
Xem thêm : Pubg Reshade Guide Settings Update 04/2024
That’s why the RX 6800 XT is such a fantastic graphics card: AMD improved upon an already promising architecture in RDNA. And the speed with which it has nearly caught up to Nvidia in terms of performance is completely unimpressive. While there is still work to be done, RTG can confidently cross “build a high-end GPU” off their list of to-dos thanks to the release of the RX 6800 XT.
We’re big fans of what AMD has accomplished with the RX 6800 XT, a red team alternative for any high-end gaming PC build that represents a return to form for the Radeon Technology Group.
3. Nvidia GeForce RTX 3060 Ti
SPECIFICATIONS
number of CUDA cores: 4,864
Starter frequency of 1,410 megahertz
The maximum frequency of the booster is 1,665MHz.
Number of teraflops: 16.20
RAM: 8GB GDDR6
Term describing the speed of memory: 14 Gigatransfers per second
Bandwidth of memory: 448GB/s
REASONS TO BUY
Finally! Something with a lower price tag
Outstanding work! 2080
Under $400
REASONS TO AVOID
An annoyance is the 12-pin power connector.
-It’s not quite up to 4K standards.
Advertisement
The RTX 3060 Ti is highly similar to the RTX 3070, making it the best value Ampere product to date. The GA104 GPU is shared between the two, as is the 8GB of GDDR6 memory spread over a 256-bit bus (though the RTX 3060 Ti has fewer SMs enabled).
The RTX 3060 Ti may have 17 percent fewer processing cores than its more expensive sibling, but it more than makes up for this with its increased GPU Boost frequencies. This helps to explain why, despite having a silicon disadvantage, the RTX 3060 Ti is only 17%-11% slower than the RTX 3070. Not bad for a card that costs $399 (if you can even find it for that price).
If you haven’t worked out the numbers yet: The RTX 3060 Ti, at $399, is 20% less expensive than the RTX 3070, so its performance per dollar is better. That’s why we like it so much: the second-generation RT Cores give it respectable ray tracing capabilities on top of being a great GPU for all resolutions.
When compared to the GTX 10-series, the RTX 3060 Ti’s gaming performance is downright terrifying. This is especially true when considering the generational gains it offers over even the RTX 20-series. Impressive 1080p and 1440p performance in a compact design; just a few years ago, a setup like this would have cost you close to twice as much. That’s fantastic, by the way.
It’s possible to experiment with high-quality 4K gaming with this card, though you may need to tweak the graphics settings in order to achieve a stable 60 frames per second in more taxing games. It takes a lot of fiddling to keep your GPU above 60 frames per second, and that’s just the average. That can diminish the initial appeal of playing games in 4K. At 1440p, things run much more smoothly, and at 1080p, you can crank up the settings and go.
The supplemental Nvidia features, such as DLSS, Reflex, and Nvidia Broadcast, can help in this regard as well. RTX and the Nvidia package are built on a foundation of great features, and Ampere has only made them better in every way, especially the ray-tracing performance.
The availability problem that has dogged every other RTX 30-series launch to date remains. Its original $399 MSRP was its selling point, but the price increase has made the card less appealing.
However, if the RTX 3080 or RTX 3070 are out of your price range, the RTX 3060 Ti is a very respectable alternative. Perhaps the most compelling evidence of this card’s accomplishments is its performance in comparison to the 20-series generation: it consistently outperforms the RTX 2080 Super.
4. Nvidia GeForce RTX 3070
SPECIFICATIONS
5,888 cores of CUDA
Starting frequency is 1.5GHz
In terms of frequency, the boost clock is set at 1,725MHz.
20.37 TFLOPs
Disk Space: 8GB GDDR6
The memory clock is 14GT/s.
Bandwidth of memory: 448GB/s
REASONS TO BUY
Superior to Turing in terms of frames per second
Power consumption is low
The FE card is hot
REASONS TO AVOID
-Painful 12-pin power plug
The RTX 3070 is remarkable because it can compete with the flagship Turing graphics card, the RTX 2080 Ti, for less than half the price. It is also the only high-end Ampere that comes close to being reasonably priced.
While $499 is a hefty sum, especially when compared to the cost of today’s top graphics hardware for PC games, it is not outrageous when weighed against the price of the next-generation consoles. To make up for it, you get a graphics card capable of 4K resolution that doesn’t need extensive tweaking to achieve playable, if not high, framerates. At 1440p, it will absolutely decimate the competition.
Surprisingly, the RTX 3070 Founders Edition retains the RTX 3080’s 12-pin power connector, so only a single 8-pin to 12-pin adapter is included in the box. With this FE card, Nvidia has once again delivered a super-compact PCB, resulting in a much more compact shroud than the RTX 3080.
While the card’s two front-mounted fans do a good job of keeping the smart little GPU cool, the spinner at the back of the card does a better job of directing air over the heatpipes.
The RTX 3070 features Nvidia’s stable driver stack, the GeForce Experience app, ShadowPlay’s recording, and the new Broadcast suite, as well as the impressive Reflex latency-spiking(opens in new tab) setup and the latent potential of RTX IO(opens in new tab), bringing the most potent feature of next-gen consoles to the PC. We also have DLSS improvements that are constantly being refined, and ray tracing is already a thing.
The RTX 3070 already boasts impressive raw performance, but this is a significant improvement that will appeal to a wide audience.
The RTX 3070, as part of Nvidia’s xx70 series of cards, is consistently near the top of the performance charts for PC gamers. That’s the Ampere card, offering a real performance boost for most of us at a price we can almost afford. If you’ve been sitting on a GTX 1080 or GTX 1080 Ti through the RTX 20-series, now is the time to make the jump.
5. AMD Radeon RX 6900 XT
SPECIFICATIONS
5,120 RDNA-cores
The initial clock speed is 1825 MHz.
Peak frequency: 2,250MHz
Performance in TFLOPs: 23.04
Total Memory Capacity: 16GB GDDR6
Time-to-restore memory: 16Gb/s
Memory data transfer rate: 512 GB/s
REASONS TO BUY
There is occasional RTX 3090 performance…
but at a lower cost!
REASONS TO AVOID
-Occasionally slower than the RTX 3080.
-Unimpressive ray-tracing speed
While NVIDIA’s RTX 3090 may have debuted as the undisputed king of graphics processing power, AMD’s RX 6900 XT, or “Big Navi,” wasn’t far behind. The RX 6900 XT’s goal is to knock Ampere’s best off its pedestal and send it tumbling to the ground. And it comes close, too, with 4K performance that’s just a hair behind the RTX 3090’s for a third of the price.
Because of this, any PC gamer who doesn’t have any pro-creator bias should consider buying it instead. However, it is not the best graphics card on the market right now for one simple reason: it is rarely, if ever, noticeably better than an RTX 3080. For an additional $300 over the RTX 3080, you might expect improved performance in both rasterized and ray tracing applications. But it will always be behind the curve in terms of ray-tracing speed.
Within AMD’s product lineup, there is no compelling reason to choose this over more reasonably priced alternatives. The RX 6900 XT is $53 more expensive than the RX 6800 XT but only provides a 5.25% performance boost in games, making it a poor value at a price premium of $350.
So it’s odd that AMD promotes the RX 6900 XT as a “gaming first” card. The RTX 3090 is priced significantly higher than the second-run gaming card below, but even Nvidia didn’t make that mistake by marketing it as a professional creator card.
Although it has some advantages over cheaper cards, the RX 6900 XT is hard to recommend due to its high price and lackluster performance boost.
I’m a big fan of RDNA 2, and the progress AMD RTG has made in such a short amount of time to get it to compete at this level is remarkable, but the RX 6900 XT is not the card’s finest form. For even the most ambitious of gamers, it’s not the most universally appealing recommendation. The RTX 3080 remains the best graphics card available at the moment.
On the other hand, the RX 6900 XT provides guarantees that the RTX 3080 doesn’t, such as its 16GB of GDDR6 memory, which is 6GB more than the 10GB of (faster) GDDR6X memory on the green team card. This means that the RX 6900 XT is a more affordable alternative to the RTX 3090 for pure gaming. Unfortunately, its high cost still works against it.
6. Nvidia GeForce RTX 3080 Ti
SPECIFICATIONS
10,240 CUDA cores
The starting frequency is 1,365 MHz.
Increased speed of 1,665MHz
34.10 TFLOPs
GDDR6X memory, 12 GB
Clock rate of memory: 19GT/s
Bandwidth of memory: 912.4GB/s
REASONS TO BUY
Outstanding 4K resolution and performance.
Save nearly the cost of an RTX 3090 while getting nearly as much performance
Possesses superior ray tracing abilities
REASONS TO AVOID
The RTX 2080 Ti is a much more affordable option.
-Requires a lot of energy, consuming 350W
-It gets pretty toasty in here
Nvidia’s RTX 3080 Ti isn’t their most powerful GeForce graphics card. This distinction was narrowly won by the RTX 3090 and later taken over by the RTX 3090 Ti(opens in new tab). After extensive testing, however, we conclude that the RTX 3080 Ti is the best graphics card for any PC gamer looking to pull out all the stops with their next build.
While the arrival of cards from the Ampere generation helped reduce the awkwardness or finickiness of 4K gaming, the RTX 3080 Ti completely eradicates the problem. High-performance 4K gaming setup is still expensive, but you no longer need to be as careful as we were when purchasing games individually. You can play games at 4K resolution with high or ultra settings and experience playable frame rates.
When playing at 4K, the last thing you want to do is lower the graphics quality settings, and with the RTX 3080 Ti, you shouldn’t have to do that nearly as often as you would at 3840 by 2160.
The RTX 3080 Ti typically lags behind the RTX 3090, but it appears the more expensive card only sees marginal benefit for an increase in core counts and memory—the RTX 3080 Ti is only a few percentage points behind the leader.
The RTX 3080 Ti is every bit as much a gaming graphics card as its ancestor, the RTX 3090, thanks to its near-identical performance at 4K. This card’s massive number of CUDA Cores and lightning-fast GDDR6X memory make short work of any game you can throw at it. With 80 RT Cores, it is also more than capable of real-time ray tracing.
Despite some rigidity in the price, the GeForce RTX 3080 Ti remains an impressively powerful graphics card; it will allow some to save money on ultra-high-end gaming PC builds without sacrificing performance. Simply cramming more of the fantastic Ampere architecture into that unassuming black and (polished) grey casing makes for a fantastic gaming processor.
As a result, it accomplishes its goal: to satisfy gamers’ aesthetic preferences while also breaking the back of any game, API, or graphics-intensive task, even at 4K. When comparing price and performance, the RTX 3080 Ti FE is slightly more cost-effective than the RTX 3090 FE, though neither card is a clear winner at the high end.
However, its high price prevents us from placing it higher on our list of recommended graphics cards. Its starting price of $1,200 places it close to the $1,499 RTX 3090. Despite the fact that the price is astronomically high or there is none in stock.
7. Nvidia GeForce RTX 3060 12GB
SPECIFICATIONS
3,584 CUDA cores
The initial clock speed is 1,320 MHz.
Rate of gain: 1,777MHz
12-74 TFlops
RAM: 12GB GDDR6
Clock rate of memory: 15Gbits per second
Throughput of 360 GB per second in memory
REASONS TO BUY
Supports 60 frames per second at 1440p or higher
Fair performance boost from generation to generation
Pre-assembled computers are more widely available for purchase.
REASONS TO AVOID
-Maybe not noticeably quicker than an RTX 2060
At $399, the RTX 3060 Ti is the superior option.
It seemed, and still seems to many, odd that NVIDIA would equip the RTX 3060 with 12 gigabytes of GDDR6 memory. After all, it has more memory than the RTX 3080, albeit slower stuff, and it also outperforms the RTX 3060 Ti. The RTX 2060 is a powerful card, but the RTX 3060 knows how to use it to its advantage, defeating it in most games by a significant margin and leaving it far behind in the few that it doesn’t.
A professional’s take on the best pre-built gaming PCs (opens new tab)
Laptops optimized for mobile gaming (opens new tab): best gaming laptops
If the alternative is only 6GB of VRAM, we’re not going to turn our noses up at that.
The announcement of a hash rate limiter(opens in new tab) for the Nvidia RTX 3060 was perhaps the most noteworthy change for the card, as it was created to prevent the card from operating at its full capability while mining Ethereum, the cryptocurrency of choice for GPU mining. By making its new CMP line, composed of GPU scraps deemed unfit for gaming use, available at a lower price, Nvidia hoped to divert cryptocurrency miners away from the company’s flagship GeForce graphics cards. Even so, the card quickly ran out of stock.
The RTX 3060 12GB is a compelling graphics card, with a respectable generational improvement and plenty of speed at 1080p and 1440p.
It is also a cheaper option than the RTX 2060 was on its release day, though this distinction is becoming increasingly difficult to find. However, you can find this card for a reasonable price in many pre-built gaming PCs (opens in new tab).
The RTX 3060 Ti continues to be the best midrange graphics card. The Nvidia RTX 3060 deviates too much from the rest of the high-end Ampere pack to continue making waves in the same way it did in previous generations. Nonetheless, it’s worth noting that the RTX 3060 is a solid upgrade from a 10-series or older card, especially if purchased for close to the MSRP, and will probably carry you confidently through the next few years of major game releases.
8. Nvidia GeForce RTX 3050
SPECIFICATIONS
2,560 cores of CUDA
Initial Clock Rate: 1,552MHz
Enhanced frequency of 1,777MHz
9 TFLOPs
RAM: 8GB GDDR6
Transmission rate of 14Gbits per second in memory
Transmitting data through the memory: 224GB/s
REASONS TO BUY
Support for DLSS and ray tracing
Better performance than the RX 6500 XT.
REASONS TO AVOID
We’re talking RTX 2060 minimum performance here.
Long-term stock and pricing uncertainty
In our opinion, the only sub-$300 graphics card worth considering is the GeForce RTX 3050. However, it didn’t take much to beat AMD’s Radeon RX 6500 XT. Even with its improvements, the RTX 3050 is still not an easy GPU to enthusiastically recommend because it is neither widely available nor as inexpensive as it should be.
The RTX 3050 has a significantly lower graphics processing unit (GPU) than the RTX 3060 12GB does, and we, like any decent PC gamer, would advise you to go with the latter. However, if you simply cannot spend any more, the RTX 3050 will provide performance comparable to that of a GeForce GTX 1660 Ti while adding DLSS.
In most cases, this will be a graphics card capable of 1080p gaming at playable settings, though some games may force frame rates down well below the 60fps target many of us aim for. It’s true that this is the least expensive Nvidia graphics card of its generation, but it’s still not chump change if you want to play your games at their native settings.
The RTX 3050 supports ray tracing but its frail internals won’t be able to handle much.
Instead of the RTX 2060 performance we were hoping for, we got something closer to the GTX 1660 Ti’s frame rates with some RTX icing on top. It’s tasty where it sticks, but it’s just another example of a card that was built and priced to meet the needs of the market during a pandemic and GPU shortage, rather than showcasing Nvidia’s full capabilities.
If you are looking for a prebuilt gaming PC on a budget, the RTX 3050 is the best option currently available. However, if you’re willing to wait, you can find excellent RTX 3060 12GB PC deals on a regular basis and pick up a graphics card with significantly more power.
Graphics card FAQs
How do I get a GPU in the the graphics card shortage?
HotStock, an app we developed, has been a huge success for us in the UK; Stock Informer(opens in new tab) and other sites provide a similar service in the US; we have not used these sites to score stocks ourselves, however.
Just like the popular StockDrops server(opens in new tab), there are a plethora of free Discord servers with dedicated stock alert bots and eagle-eyed community members.
Also, Twitch broadcasts should not be overlooked. Those who are determined to locate your stock will often sound a loud klaxon as soon as stock is spotted. For Nvidia graphics processing units, we suggest Falcodrin on Twitch(opens in new tab), though there are many generous people who provide this service.
Is there an easy way to get a new graphics card?
Which is better GTX or RTX?
Is ray tracing only for RTX cards?
Is SLI or CrossFire still a thing?
Do I need a 4K capable graphics card?
What’s a Founders Edition graphics card?
Conclusion
And with that, we have reached the end of our discussion of the various considerations you should make when selecting a suitable graphics card.
If you’re in the market for a new graphics card and want to see what we consider the best gaming GPUs to be right now, we recommend checking out our comprehensive buying guide.
Nguồn: https://gemaga.com
Danh mục: Guide
