DirectX 11 vs. DirectX 12 – Radeon RX Vega 64 and Geforce GTX 1080

If the last two years have been a disappointment for those looking for graphics cards in the performance segment, all the more focus has been on interfaces. A gradual shift towards DirectX 12 and Vulkan has meant an increasing focus on the more modern solutions – for better or worse.

Not everything is glittering gold, to use a worn term. DirectX 12 in particular has been really problem-laden, and hardly convincing except in occasional game titles. Opportunities for higher frame rates on weaker processors in all glory, most are not prepared to take the step if it happens at the expense of stability.

What does it look like in the autumn of 2018, a year after the Radeon RX Vega saw the light of day? With plenty of numbers on hand for upcoming major graphics card tests, SweClockers is taking the opportunity to pick out a host of results, which reflect the performance of both DirectX 11 and DirectX 12.

Components of the test systems

As a basis, the editors are rolling out the autumn’s new test system – more about these in a separate article. In short, however, these are really great gaming machines with overclocked Intel Core i7-8700K in the center, plenty of DDR4 memory and fast storage in the M.2 connections.

This is how we carry out the tests

For the actual testing, parts of SweClocker’s fresh game suite are used, which will be presented in more detail in an upcoming graphics card test. There are a total of six different titles running in both DirectX 11 and DirectX 12, with the highest or near highest settings. For this scenario, the editors choose to pick out the resolution 2,560 × 1,440 pixels.

The drivers used are the newest at the time of testing from AMD and Nvidia, respectively. These are Geforce 398.82 WHQL and Radeon Adrenaline Edition 18.8.1. The frame rate is captured with the OCAT software for all titles. Note that the best results per graphics card are highlighted in yellow, not for the entire line.

DirectX 11 and DirectX 12, at 2,560 × 1,440 pixels

The result that is painted is very clear. Out of six titles tested, AMD Radeon RX Vega 64 performs better under DirectX 12 in five. For the Nvidia Geforce GTX 1080, the opposite is true – only one game shows better DX12 results, and in several cases the DX11 is significantly faster.

There is more than the numbers show

So far, so good. DirectX 12 gives better performance with AMD Radeon, and Nvidia is the worst – or? If it were that simple anyway.

Although much better than before, DirectX 12 still involves some uncertainty, regardless of graphics maker. There are a little more graphics bugs, a little more crashes, a little more unexpected performance losses or strange menus. A little more unpolished, quite simply.

The big mourning child, for both AMD and Nvidia, is Total War: Warhammer II – more so than the numbers show. The DirectX 12 feature is clearly marked as beta, and luckily for you – the game crashes, does not work properly with some cards, gives sometimes crazy performance losses and sometimes more than doubled charging times.

But, should I run with DirectX 12 or not?

An attempt at the rule of thumb is that if there is an AMD Radeon under the hood – test, by all means. You can very well earn a few frames per second, without actually losing that much – especially with a slightly slower processor.

Those who sit on an Nvidia Geforce, on the other hand, can in good conscience leave the slider set to DirectX 11, in cases where the choice exists. The performance is in almost all cases equal or better, at the same time as we duck a whole lot of small problems.

The above two summarizes quite well the status of DirectX 11 and DirectX 12, early fall in the year of grace 2018. The challenges around the new interface are many, and sometimes the benefits are unclear – especially for developers with limited resources. It will probably take many years before we can add the 11th to the story.