Date: Thursday , May 24, 2018
There are many games that were launched years ago with earlier generations of GPUs. When these games were launched on the earlier hardware of the time we could not enable all of the in-game graphics options up to the highest level. In addition, we could not enable special graphical options that up the ante on image quality, but also cause a large burden on GPU performance that these games featured. Perhaps a game supports a unique feature like PCSS or HFTS shadows, or a high detail slider, or extra distance slider, which turning these features up would cause a great demand on performance.
We recently saw this video about playing older games in 4K, which got us thinking about this, but more from a performance perspective, and image quality perspective.
Considering that we are now at least a couple of years since these games were released, and we now we have newer GPUs with much more mature drivers we wanted to see what was not afforded to us in terms of better performance and IQ. The NVIDIA GeForce GTX 1080 Ti and the AMD Radeon RX Vega 64 are the fastest single-GPU video cards from both camps with mature drivers in 2018. The question is, can we now play these older games and turn up all the image quality features at the highest levels possible and enjoy an improved gaming experience compared to when these games were launched?
In order to test this, we are going to take the GeForce GTX 1080 Ti, and the Radeon RX Vega 64 with the latest drivers and take a look at some older games. To be considered an older game we are looking at games at least one and a half years old. The oldest game we have in this article will be Grand Theft Auto V released to the PC on April of 2015 on the PC. At this time the fastest GPU from NVIDIA was the GeForce GTX 980 (Maxwell) and from AMD the Radeon R9 290X (Hawaii) video cards. The most recent game we have in this article is Watch Dogs 2, released in November of 2016. In November of 2016 the fastest GPU was the GeForce GTX 1080 and AMD Radeon RX 480 and AMD Radeon R9 Fury X. The GeForce GTX 1080 Ti wasn’t released yet.
We are going to focus on 1440p and 4K performance with the goal of improving the gameplay experience beyond what was possible at the time these games were first released.
1440p is going to be our baseline, if all the in-game options can now be turned on and up at 1440p, we will consider that successful. 4K is being tested because that is an instant way to achieve a better experience in games. We have to understand even at 4K some games may still be too burdened to turn up all the options, even on the GeForce GTX 1080 Ti. 4K is there more of a "what if" scenario, what if you played these older games at 4K, what would be playable today in that game? In the past, 4K was not a common resolution to test performance at, but today, it certainly is. Playing an older game at 4K is one way to instantly improve the gameplay experience so we need to see what is playable today, it may be a good experience with older games.
When we talk about turning on all the in-game options, we are referring to the graphics settings in the game that affect image quality and improve the gameplay experience and are native to the game. For example, we are not going to consider MSAA or SSAA or TXAA settings part of this. Those settings can kill performance, and it wouldn’t be fair to the game to consider MSAA or SSAA or TXAA as part of this. We will use shader AA though like FXAA or SMAA as it does not affect performance much. What we are focused on are all the other in-game settings, and special graphics features a game may support to improve the gameplay experience. Features like PCSS or HFTS shadows, or extra detail sliders, or extra distance sliders etc.
We will test MSAA and SSAA and TXAA when supported in-game, just to give you the data and information as a comparison, but we won’t consider that as part of a game being playable at the highest settings as successful or not. We will explain all of this in more detail in each game, and it will make sense as you read each game page as we describe the game settings available in each game.
Keep in mind is that we are not comparing the performance of GeForce GTX 1080 Ti to Radeon RX Vega 64. We go into this evaluation knowing that GeForce GTX 1080 Ti and Radeon RX Vega 64 are in two different price and performance categories. This evaluation is about each game in question, taken separately, and finding out what is playable on today’s fastest single-GPU video cards from NVIDIA and AMD. Using the latest drivers our goal is to find out how each video card performs in each game, individually. Instead of comparing by video card, we are comparing by game features. This will also show you Radeon RX Vega 64 performance in older games which we have not tested yet because that GPU did not exist when these games were released. Therefore, it should also give you some new GPU performance we haven’t shown before.
In order to accomplish these goals each game will be represented on two pages. On those pages we will look at three overall big topics, or questions to ask and answer. The first big topic to find out is if we can now maximize a game’s graphics settings on either video card? That is, to use all the special features a game may support, like PCSS Shadows, or HFTS Shadows, or VXAO Ambient Occlusion or Extra Detail sliders, or Extra Distance sliders etc.
The second thing to test is if we cannot run at the highest settings, then what is playable, and what do we have to sacrifice in image quality?
Finally, the third test is to compare the special graphics features performance in games and find out what the graphical cost is in performance enabling those. This will show us why a game might not be playable at the highest settings, and how much of a GPU performance cost a feature is.
From all this data we can then talk about what kind of performance may be required to finally maximize these old games using all the graphics options they possess at the highest in-game settings. We can see if the performance is there now to really enjoy these games at 4K, improving the experience.
We recognize there are many older games out there we have not included in this article. We only went back to the beginning of 2015, and there are certainly older games than that that deserve to be tested. We can only include so many games per-article since the testing involved in this kind of format takes a tremendous amount of time. Remember, as always, this is REAL gameplay, not canned benchmarks that can put run in a few short minutes. You may not see the game you wanted to see. If this evaluation proves popular we may explore more games in the future with more articles testing more games in this way.