Should you buy an Nvidia or AMD video card?

The Nvidia vs AMD war is a never-ending heated debate. Let’s try to find out which side is really better.

Better AMD or Nvidia video card is the eternal topic of technology people. Gamers are passionate about technology and there is no better piece of hardware than video cards. So what is the difference between AMD and Nvidia video cards? Who is the better manufacturer, who is the best GPU maker, and which should you buy?

These are the things that we will learn in this article.

HISTORY SUMMARY

AMD

Let’s take a brief look at the history of the two companies. AMD (Advanced Micro Devices) has been around since 1969, almost 50 years ago. Headquartered in Santa Clara, California, the company started out making microchips, mostly for other manufacturers. Over the years, AMD has acquired other companies and sold off parts of its business. The two most famous deals are the acquisition of ATI Technologies in 2006, AMD’s GPU business, and the sale of its manufacturing business in 2008 to GlobalFoundries. This is the AMD people are most familiar with, the company that designs the CPUs and GPUs, and the production takes place in multiple locations – TSMC, GlobalFoundries or Samsung. AMD’s main products are currently the Ryzen (CPU) and Radeon (GPU) brands.

Nvidia

Nvidia is not that old. Founded in 1993 and also headquartered in Santa Clara, Nvidia has been focusing on graphics cards since the beginning. The first major product was Riva TNT in 1998, followed by TNT2 later in the year. They were considered the most successful all-in-one 2D and 3D video card solution at that time. GeForce 256 in 1999 became the first GPU (Graphics Processing Unit) thanks to hardware support for T&L (Transform and Lighting) algorithm. Nvidia’s GeForce brand has been around for nearly 20 years, and is now in its 17th generation. Nvidia is also a company that designs, not manufactures itself, mainly relying on TSMC to manufacture GPUs, but Samsung is also involved. manufactures some chips.

The video card industry is now largely dominated by ATI/AMD and Nvidia, with Intel also slightly involved thanks to the integrated GPU segment. (Intel also plans to enter the discrete GPU market in 2020, but it’s too early to say anything more.) Both AMD and Nvidia have had many segments over the past decade, including chipsets, mobile devices, and the debate between the two companies mainly revolves around GPUs and graphics products. And the article will only revolve around those two topics.

How does your GPU work

The early challenges in the video card market included competition on standards. 3dfx makes the Glide API (Application Programming Interface) a low-level method of communicating with hardware, increasing performance, but it can only run on 3dfx hardware. Having a platform that can work with any hardware drives software development. As a result, DirectX and OpenGL soon rose to prominence.

API

These days we have low level ‘popular’ APIs like DirectX 12 and Vulken. Whatever the API, we can understand the GPU as a black box. The API defines the information needed to perform a function and output a result, but how the results are produced depends on the driver and hardware. If the company can think of a more efficient way to implement the algorithms, it will gain a major competitive advantage. There is a second way to gain an advantage, which is to create an extension to the main API so that it can implement new algorithms – and if the extension becomes part of the API itself, it will be even better. .

In general, both AMD and Nvidia GPUs do the same main algorithms, but the difference lies in the application. AMD’s GPUs have been an example of asynchronous computing since the early GCN architecture (Graphics Core Next) in 2012, but it didn’t really matter until Windows 10 came out in 2015 with DirectX 12. Asynchronous computing. The set makes the video card’s command processing more flexible. In particular, there are three instructions defined in DirectX 12, and the GCN architecture has three distinct hardware associated with them. Nvidia’s GPU cards don’t have separate hardware, instead applying instruction processing through drivers.

Difference between AMD and Nvidia GPUs

AMD’s GPUs usually have more processing cores when compared to rival Nvidia. For example, the AMD RX Vega 64 has 4,096 GPU cores, while the Nvidia GTX 1080 has 2,560 cores. Same goes for RX 580 (2,304 cores) with GTX 1060 6GB (1,280 cores). Nvidia compensates for the lack of cores with higher clockspeeds and better performance.

The sub-algorithm, for other things, is Nvidia’s strong point – or proprietary if you will. As a major GPU maker (below), Nvidia has added a lot of technology over the years. PhysX and the vast majority of GameWorks games are designed for Nvidia GPUs. In many cases, they can’t even be used with non-Nvidia cards. The GeForce RTX Turing architecture pushes it to the next level, helping to pave the way for real-time ray tracing, deep learning, and other features. Will AMD be compatible with DirectX Ray Tracing (DXR) in the future? Possibly, but we won’t know until Navi comes out, in 2019 at the earliest.

The REAL QUESTION: POWER AND PRICE?

Ultimately, the argument over which company is better must be decided by its performance and cost. Many people really love sports cars, but very few people are willing to pay for them. Competition between companies creates a positive price signal, and conversely, when one side is in the lead or in power, it often leads to higher prices. In the past, the fastest GPU throne has changed hands many times. GeForce DDR, GeForce 4, GeForce 980 and later always take this title to Nvidia, but Radeon 9700/9800 Pro, HD 5870, HD 7870 and R9 290X are also great options.

Performance

Now, when it comes to performance only, we don’t argue too much about the fastest GPU. Nvidia’s GTX 1080 Ti and higher series easily beat AMD’s fastest GPUs, with the new RTX 2080 Ti being the “king”. Across 12 benchmarked games, many of which carry the AMD logo, the RTX 2080 Ti still has an average 87% lead over the RX Vega 64. Nvidia also won the next three spots in terms of performance: RTX 2080 ‘only’ 45% faster than Vega 64, GTX 1080 Ti 40% faster, RTX 2070 18% faster, and even GTX 1080 5% faster. But the price of the RTX cards is…

Price

The RTX 2080 theoretically starts at $699, which is the same as the GTX 1080 Ti, with slightly faster performance than current games and possibly faster future releases. However, the cheapest RTX 2080 is currently priced at $770 – 10% more than the original asking price and $1,199 for the Founders Edition, while the 2080 Ti is currently out of stock. At least the $499 RTX 2070 cards are still on sale, and they’re still faster than the fastest AMD cards in terms of overall performance. Meanwhile, the Vega 64 starts at $450, and the slightly slower RX Vega 56 costs $380.

If you’re aiming for better performance cards, you’ve got your answer, but $200 cards sell for more than $500 cards. Nvidia may lead in performance, and performance per dollar in high-end models, but what about the mainstream? Nvidia’s GTX 1060 6GB is a bit slower than AMD’s RX 580, and the AMD card is about $20 cheaper. Same thing with RX 570 4GB vs. GTX 1060 3GB. But now, very few people want to buy a card with less than 8GB of VRAM.

So which brand should you buy?

Supporting smaller firms has no small benefits – competition drives more competitive prices. Let’s take a look at GeForce RTX to understand this better. Given the dominant position, the price of GeForce RTX cards is obviously higher.

But the vast majority of gamers play a lot of games, so average performance and value become the main factors in deciding which company to buy.

High-end segment

Nvidia dominates the high-end GPU market. With any video card more expensive than $350, Nvidia usually wins in terms of value as well as performance, and Nvidia’s GPUs are more efficient at each price point. GTX 1070 Ti when compared to RX Vega 56, has better performance, is a bit cheaper and uses about 50W less. In the higher segment, AMD is not the optimal choice. Geforce RTX cards are currently at the top in terms of performance. However, AMD still plans to continue releasing high-end graphics cards.

Mid-range segment

For the mid-range segment, the market is more competitive. GPU prices have returned to normal after the mining craze of early 2018, AMD’s RX 570 and RX 580 beat Nvidia’s GTX 1060. AMD still lags in overall performance, taking more power (30-50W), but at least you get better performance and price. There are Nvidia games that lead in terms of performance (like Assassin’s Creed Origins and Odyssey, Grand Theft Auto 5, and Total War: Warhammer 2), so it’s unclear who wins in this segment, but AMD is still a great choice for mid-range gaming PCs.

Low-end segment

The low-range segment is quite similar to the mid-range. The GTX 1050 Ti is faster than the RX 560 4GB, but costs as much as the RX 570 – a GPU that the GTX 1050 Ti can’t beat. Meanwhile, the GTX 1050 often lags behind the 4GB RX 560, due to only 2GB of VRAM. If you’re aiming for a $100-$125 graphics card, AMD is currently leading the low-end.

Nvidia is still dominating

Overall, we should be pleased that AMD leads these segments. In theory, there are more mid-range gamers with limited wallets than high-end gamers. However, that has not stopped the GPU market share that Nvidia is dominating. According to Steam’s latest hardware survey – perhaps the best document we have on the PC gaming community – the results are: 75% Nvidia, 15% AMD and 10% Intel. But things are even worse.

Now we’re only looking at ‘recent’ GPUs that are compatible with DirectX 12 – AMD HD 7800 or later and GTX 760 or later, excluding Intel and older GPUs. In this case, AMD’s game market share fell to 7%, while Nvidia accounted for 93%. Continuing, when comparing AMD’s RX and Nvidia’s later GPUs: GTX 900 and 1000, Nvidia’s market share increased to 97%.

The top 15 most popular DX12 GPUs come from Nvidia – and AMD’s Vega isn’t on that list. That is, Vega accounts for less than 0.3% of the “remaining” products. If you want to buy an AMD GPU to support them, this is the time to do it, at least when you are looking for a mid-range or lower graphics card.

Future

AMD is contemplating launching entry-level and entry-level graphics cards in the coming months, hoping to increase market share in new GPU sales. People are still recovering from the mining craze, which took a big chunk out of AMD’s revenue in GPUs. And if this trend continues, the only hope for the competition of the graphics card line is Intel’s GPUs coming out in 2020, so AMD’s revival now is of concern. AMD plays an important role in the industry. While people often recommend that you buy Nvidia cards, that doesn’t mean we want to see Nvidia take a monopoly.

By PC Gamer