other games Gaming News: AMD's Mantle vs DirectX: The Benchmarks

[Armory Staff]

/files/u166440/pvzgw_001.jpg

 

How does Mantle stack up against DirectX in certain games? Find out here!

AMD has been partnering up with as many video game developers as it can to incorporate its Mantle API. From Crytek to Robert Space Industries, AMD is slowly making the rounds. Games such as Battlefield 4 and Thief allow users to switch from DirectX to Mantle while other games are expected to incorporate the API at some point.

 

With Mantle, AMD claims that it boosts a game’s performance for both low and high-end GPUs. Albeit, when it comes to high-end GPUs, the boost isn't as big. Curious, we decided to see how well Mantle stacks up to DirectX by playing games that support it. For now, there is only a small selection of games: Battlefield 4, Plants vs Zombies: Garden Warfare, Thief, and the latest Mantle introduction, Sniper Elite 3. 

 

In order to get the best results, we made sure, when possible, to benchmark the same section of each game with the same graphics settings. For the testing we used a PC with an AMD Phenom II X4 965 processer, 8GB of RAM, and an AMD Radeon R9 280 graphics card. 

 

 

/files/u166440/bf4_001.jpg

 

To help us record our results, we had to rely on benchmark demos or in-game benchmarking tools. Normally, we would use FRAPs, which would record real-time FPS and then provide us with the minimum, maximum, and average numbers, but the problem is that FRAPs is programmed for DirectX 11 and does not support Mantle just yet. The only software that will show FPS numbers is D3DGear. Unfortunately, it doesn’t record results like FRAPs does. 

 

Luckily, most of the games that support Mantle come equipped with either their own benchmarking demo or benchmarking tools. With that said, let's take a look at the benchmarks!

Battlefield 4 DirectX 11 vs Mantle

/files/u166440/bf4_avg_fps_001.jpg

First up was Electronic Arts’ Battlefield 4. We cranked the graphics settings up to High at 1920x1080 resolution, and then ran the same section of the game across both APIs. We decided to compare the first 30 minutes of the singleplayer campaign to use as our test scenario. We felt that Chapter 1 of BF4's singleplayer campaign was a good test as it went from an urban setting to more wide open vistas filled with gunfire, explosions, and fire to put both APIs through their paces.

 

Looking at the graph, you can see that there was a difference between both playthroughs. For the most part, both tests went smoothly though, with Mantle, we experienced some brief stutter in the gameplay. Despite that, Mantle still pulled ahead with an 11.8FPS advantage between it and DirectX 11. 

 

Sniper Elite 3 DirectX 11 vs Mantle

/files/u166440/sn3_avg_fps_001.jpg

Sniper Elite 3 was recently updated to support Mantle and with it came a new benchmark demo. With settings on Ultra at 1920x1080 resolution, we launched the demo. The demo is set up to showcase many of the game’s various scenes. It was replete with long distance views, lighting and shadows, close-up views of objects and terrain, a truck full of enemy NPCs with the wheels kicking up dirt, capped off with an exploding vehicle blowing up in slow motion.

 

/files/u166440/sniper_elite_3_001.jpg

But as we saw with Battlefield 4, Mantle had a slighter higher average FPS than DirectX 11 with a 13.6 FPS advantage.

 

Thief DirectX 11 vs Mantle

/files/u166440/thief_avg_fps_001.jpg

Unlike the previous games we have tested, which all took place in broad daylight, Thief's game features a lot of nighttime excursions. We set the graphics to Ultra High at 1920x1080 resolution and took to the rooftops, An interesting thing about this game is that it comes with its own benchmarking system designed to compare both APIs.

 

/files/u166440/thief_screenshot_001.jpg

So we ran the benchmark for DirectX 11 and Mantle. the demo is a sequence that runs through the town at night, while it is raining, with various NPCs walking around and interacting with each other. It is a simple enough sequence that was able to put both APIs through their paces. And, just like the previous two games, Mantle scored higher than DirectX 11 with an 11.5 FPS advantage.

 

Star Swarm DirectX 11 vs Mantle

/files/u166440/swarm_screenshot_001.jpg

Now if you want to really stress test both APIs, then definitely take advantage of the Star Swarm Stress Test. It is a free, real-time demo from Oxide Games that showcases the developer's Nitrous engine. Set in space, Star Swarm pitches thousands of fighters and ships into constant combat. For the demo, there were around 4,000 ships constantly firing lasers, exploding, and more.

 

/files/u166440/star_swarm_stress_benchmarks_002.jpg

The demo itself lasts for six minutes and we made sure to have the resolution set to 1920x1080 and the graphics on its highest setting which was labeled Extreme. And it certainly lived up to that name. Star Swarm made both APIs struggle and have convulsions. When the camera panned around to show thousands of ships fighting simultaneously, the FPS dropped into the single digits and the game slowed to an unbearable crawl. 

 

But as you can see from the results, Mantle was able to handle the stress test better than DirectX 11 with a 16.3fps advantage. 

 

/files/u166440/pvz_2014-10-07_19-12-56-09.jpg

Now we come to Plants vs Zombies: Garden Warfare. Unfortunately, while the rest of the games include their own in-game benchmarking tools, PvZ does not. So there is no way to accurately gain any results when we run it with Mantle. Until PopCap Games adds in some benchmarking tools, or software like FRAPs finally supports Mantle, we will have to skip it for now and come back to it later.

 

The same goes for future games that will support Mantle. We’ll be testing them and updating our story here so stay tuned for more games and more results!

 

Follow Sean on Google+, Twitter, and Facebook

 

View the full article