Evolve PC Benchmark Performance Nikolas Nikolaou November 2, 2014 Benchmarks Today we are going to take a look at the benchmark performance of Evolve, a new multiplayer shooter game based on Cryteks CryEngine (4th Generation). I was lucky enough to get an invite to the Big Alpha which is a 3 day closed test to stress servers and help provide feedback to developers Turtle Rock Studios before the game’s official release on the 10th of February 2015. Evolve is a unique co-op multiplayer 4v1 action game. Four of the players (trapper , medic , assault and support) or else called the Hunters try to track down and kill the Monster before it evolves to stage 3 and destroys the power relay. Quite simple you may think, but there’s a lot going for it. Each of the Hunters as well as the Monster have unique abilities and get power-ups when killing NPCs that consist of creatures and deadly wildlife. Testing Components and Methodology Evolve Benchmark System CPU i7 3930K 4.3Ghz Motherboard Asus P9X79 Deluxe RAM Kingston Hyper-X 16GB 1600mhz HDD/SSD Crucial M550 256GB GPU Galax GTX970 EXOC 4GB , MSI R9-290 Gaming 4S Monitor Asus PB287Q 28″ 4K 60Hz 1MS PSU Corsair AX1200W OS Windows 8.1 Drivers Nvidia: 344.48 WHQL, AMD: Catalyst 14.9.2 Beta Worth mentioning is the results provided might vary to the official video game release after optimisations and driver releases. The benchmarks are tested in an Alpha state that, of course, will have various problems. Replicating a benchmark in a multiplayer game isn’t the easiest of things but I did my best to try to standardize whatever I could. There were various menu glitches as well as a few long loading times, but that’s fairly understandable. I used Goliath (the Monster) at the Fusion Plant in all benchmarks and evolved to stage 2 so I could incur more damage and continue to find the other players. As soon as the fight started with the Hunters I ran the benchmark for 60 seconds. The game did run crossfire on the AMD R9-290s but the performance was actually worse than a single card (due to bad scaling). For the Nvidia GTX970s, SLI didn’t work at all, so we avoided adding any multi-GPU frame rates and frametimes for either manufacturer. Benchmark Resolutions : 1920×1080 and 3840×2160 Benchmark Settings : Very High Preset, SMAA TX1 , Tesselation ON , Vsync OFF Evolve 1920×1080 Very High Benchmark For some odd reason, even though we had VSync off, we see the frame rate cap around 60 FPS. We are not 100% sure if this has to do with our 60Hz monitor or if its something to do with a limitation from the developers. The GTX 970 seems to be slightly ahead of the R9-290 by a small margin. Looking closer at frametimes we see that the Maxwell GPU performs smoother. Something that’s felt in-game even though the frame rates between the two graphic cards were very close together. Evolve 4K Very High Benchmark At the 4K resolution we see the R9-290 performing better than the GTX970, as it has better compute power as well as a larger memory bus (256-bit vs 512-bit). This is in addition to a higher TDP and a higher heat output. Looking at the frametimes on both GPUs the game is unplayable at 3840×2160. There was a lot of stuttering during gun fights that made it harder to kill opponents as well as navigate on the map. Conclusion Despite this being an Alpha version I was glad I benchmarked Evolve as it has a very unique angle with the stealth mode as well as powering up and evolving the Monster. I did take a shot at playing all of the other characters and they each needed their own unique style to be efficient at killing the monster. It’s like having 5 games in one. It’s an enjoyable experience with minor problems (mainly during loading and flickering) but we hope all this will be ironed out as we have a good 3 months before release. In terms of performance it’s demanding over the 1080p resolution, something you would expect from the CryEngine. To run the game maxed out at 1440p or 4K you are going to need two higher end graphic cards when (hopefully) CFX and SLI are supported. All in all it’s a great title that, in my opinion, just needs a few driver updates and multi-GPU support from Nvidia and AMD. Meet you in Shear!!! See Pricing If you have any questions about the benchmark please ask in the comment section below and I will be glad to answer. pong76 in order to get decent fps at 4k on a single card, settings for most games will have to be at medium or even lower than that. however, it still will look very good (better than max/ultra at 1080p IMO) and a constant 30 fps can be achieved with the right hardware configuration. Nikolas Nikolaou Most games at 4K can run perfectly fine on the highest preset aslong as you drop Anti-Aliasing, that ofcourse with the higher end GPUs (R9-290X,GTX980,Titan,GTX970,R9-290) if they don’t hit any vram limitations that is. We tested Evolve with SMAA 1TX (its the default AA) on both resolutions. And yes its alot better than 1080p as all the jaggies disappear (most of the time) even without AA. pong76 I agree, AA makes a big difference in performance. Though there are other things you can sacrifice at 4k, like shadow quality, particle quality, etc. and leave some form of AA on. It really depends on the game as well. Fiddling around with settings is half the fun of pc gaming anyway. But as you say correctly, SLI/CFX is needed and we only can hope that devs will get their act together on this. jcpro What tool was used to perform the benchmark? Nikolas Nikolaou There’s no internal benchmark tool. As mentioned in the methodology section we recorded during the first encounter with the Hunters at the Fusion Plant. We used Fraps to get the frametimes and frame rate , then used Frafs Benchviewer to get the frametime percentiles. Its a lifesaver rather than doing formulas in excel. Bob I’ve got a decent computer and this game runs like sh** for me. I have a quad core i7 @3.2ghz, 6gb RAM, which I know isn’t great, but serviceable, and a gtx760 or 770, I don’t remember off the top of my head. But I generally don’t have problems maxing out most games I play but evolve is basically unplayable for me, even on high settings. Very annoying Nikolas Nikolaou If you are running High-Very High and your GPU is 2GB (760-770) that is probably the reason.I have a GTX680 (same chip as the GTX770) laying around but didn’t test it on Evolve. What resolution are you running and are you getting strange stutter? Bob 1920×1080. But like I said, it’s basically unplayable. Like the game will freeze, be really smooth for a few seconds, freeze up, etc. I don’t have problems like this with other games, including other cryengine games. And I just checked, it’s a 760 Nikolas Nikolaou 2GB I guess ? Smells like a vram bottleneck . Lower AA or setting it OFF should solve the problem. Bob I’ve tried these things, they didn’t work. I had the same problems in the previous alpha test as well. It’s got me stumped. But yes, 2gb. Nikolas Nikolaou Seems your not the only one http://www.youtube.com/watch?v=3SYLvxcgleA Bob So picture that, but WAAAAAY worse. That’s my Evolve experience so far haha Shaun Walsh How about a test with vsync off. Turning it off in the settings doesnt work. You have to go into the ini settings file and turn it off there. It will have setting called maxframerate or something to that effect… Nikolas Nikolaou Shaun the tests were run in the Big alpha, I RMAed my GTX970s a few days ago so sadly I can’t replicate the results in the retail release. I would of liked to be able to add SLI and CFX results. Trent Foley “This is in addition to a higher tdp and heat output”. What is the point of this statement? The 290 performs better at 4k in this benchmark while the 970 performs better at 1080p. Nikolas Nikolaou The MSI R9-290 I have uses more power and runs hotter than a GTX970, most if not all R9-290s run hotter and need more power. It’s a stronger raw performance card after all. “it has better compute power as well as a larger memory bus (256-bit vs 512-bit)” overlord 970’s unable to fully utilise 256 bit memory bus in 4k due to missing L2 cache causing it to nosedive when hitting 3.5GB optimised vram. Unoptimised .5GB vram lowered effective memory bus to 224 bits. Nikolas Nikolaou I play at 4K, the exact reason I RMAed mine. Gladly the store I purchased them from had a no questions asked return. overlord Your findings are appreciated. Gr8 site. Linked you here, hope you don’t mind http://wccftech.com/amd-hiring-cpu-engineer-gpu-driver-optimization/ Any thoughts on testing APU’s or 860k, Mantle & DX12 may impact low end CPU’s most? Nikolas Nikolaou No problem at all. Hopefully I can get some lower end hardware to test by the time we have a game or two for DX12.