Dragon Age Inquisition Benchmark Mantle Vs DirectX Nikolas Nikolaou November 25, 2014 Benchmarks Continuing this month’s crazy AAA title releases I have another PC performance analysis for you. This time around it’s Dragon Age: Inquisition benchmarked in both Mantle and DirectX. DA: Inquisition is the third release of the action-RPG series Dragon Age, developed by Bioware and published by EA. This year’s version is based on the Frostbite 3 engine that’s been used in Battlefield 4 , Need for Speed Rivals and the upcoming Battlefield Hardline to name a few. Suprisingly, after customising my character and loading up the game to get a feel for any bugs and glitches, other than a few odd cutscenes being laggish and a few connection errors after exiting and re-entering the game, the general experience was smooth in comparison with most of the latest game releases. Before jumping into the benchmarks I wanted to warn anyone changing settings to first exit and re-enter the game to have a stutterless experience. When changing settings without exiting you might encounter performance issues that wouldn’t be present otherwise. As both AMD and Nvidia have released SLI and Crossfire compatible drivers I went the extra mile and added those to our single GPU results for both DirectX and Mantle on all the tested resolutions. Benchmark Components Dragon Age Inquisition Benchmark System CPU i7 3930K Motherboard Asus P9X79 Deluxe RAM Kingston Hyper-X 16GB 1600mhz HDD/SSD Crucial M550 256GB GPU 2XGalax GTX970 EXOC 4GB , 2XMSI R9-290 Gaming 4S Monitor Asus PB287Q 28″ 4K 60Hz 1MS PSU Corsair AX1200W OS Windows 8.1 Drivers Nvidia: 344.75 WHQL, AMD: Catalyst 14.11.2 Beta Testing Methodology There’s a built in benchmark tool for Dragon Age: Inquisition that can be found in the Extra menu. The benchmark tool run though is very small (15 seconds) and has a big difference in terms of performance compared to actual gameplay, especially on the minimum frame rates. I decided to run a custom 60 second benchmark for DirectX and Mantle on all cards, the run can be seen below in the video. All resolutions used – 1080p,1440p and 2160p, are run on the Ultra preset. Dragon Age 3 Inquisition 1920×1080 Benchmark The GTX970 in single and SLI perform marginally better in both average and minimum FPS against the R9-290 and Crossfire R9-290s. Mantle however seems to be slightly behind in both avg and min frame rates. This, of course, is mainly due to the API being useful for lower end hardware boosts allowing for more draw calls and reduced CPU overhead. As I use a powerful 6-Core Intel processor there were no actual limitations in CPU performance and of course, no performance boosts due to this. For those with GCN AMD GPUs and low or mid-tier CPUs, you should see performance boosts of 10-45% using Mantle instead of DirectX. 1920×1080 Screenshot 2560×1440 Mantle Vs DirectX Benchmark The Mantle results seem to reflect what I said above. We can see the R9-290s bit bus at 2560×1440 starting to kick in with the results on par in the single GPU scenario while with multi-GPUs there’s a slight lead over the SLI GTX970s. 2560×1440 Screenshot Dragon Age 4K Benchmarks At 4K we start to see different results when comparing our Mantle and DirectX frame rates. In Mantle the R9-290 hits higher averages while in Crossfire it holds better minimums. 3840×2160 is a beast of a resolution in most modern games with everything maxed out so your system has to be well balanced and beefy in both CPU and GPU power. The Ultra preset has 2xMSAA which is a big frame rate killer. Those wanting a performance boost should set AA off and set post processing and tessellation to medium or high. This will ensure FPS gains while not reducing image quality that much. 4K Screenshot Dragon Age : Inquisition CPU Usage Being a Frostbite 3 Engine game we see equal usage across multiple threads (30-40%) while gaming, with a few of the cores working slightly harder. The 85%+ usage spikes seen above are from the menus. Hopefully Bioware will fix this bug as there’s reports of menu glitches for people with 4-core Intel and lower end AMD CPUs. VRAM Usage The first thing that probably slapped you in the face is the close to 6GB VRAM usage at 4K while our VRAM test cards the R9-290s in Crossfire only have 4GB (mirrored) . This is probably due to a bug on the driver side, so it should be taken with a grain of salt. The VRAM usage was monitored for DirectX , as in previous Mantle enabled games as well as DA usage was off. At 1440p we see the 4GB buffer being filled with no bottlenecks, as there’s no single digit FPS drops in the charts. The game uses VRAM very efficiently so kudos to Bioware. DA Total System Memory Usage The memory usage seems to be on par with the system requirements from Bioware. The minimum required 4GB should be fine for resolutions at or below 1920×1080 while 2560×1440 and 4K will need 6 or 8GB to run perfectly as per recommended system requirements. Conclusion Even though the frame rates seem low on Ultra at higher resolutions, it’s one of the only games I could actually play (due to my intolerance to frame drops) without going full retard. The experience was smooth in both Mantle and DX and the small glitches here and there didn’t affect gameplay in any bad way. Mantle helps in situations where you’re CPU bottlenecks especially on lower end hardware or uber resolutions with multiple GPUs. The character models are very detailed with textures, tessellation and shadows being some of the best for a modern open world RPG, even considering it’s a multi-platform (new and old gen) release. The character customisation and LOD are something to brag about and a paradigm for what other games should do in the future. The combat and tactical view takes some getting used to but nothing too difficult. I’m not one of the series’ fans but I’ll probably be sinking a couple of hundred hours into this one. EA seem to be getting better at catering for what their customer base really wants. See Pricing If you have any questions about the benchmark please ask in the comment section below and I will be glad to answer.