H1Z1 Benchmark Performance Nikolas Nikolaou February 27, 2015 Benchmarks The truth of the matter is that I wasn’t initially going to do a H1Z1 benchmark, but due to success over on Twitch and watching quite a few hours of Battle Royale on CohhCarnage’s channel (great guy btw), I got all hyped and decided to dive into Early Access. H1Z1, for those that don’t know is a survival zombie MMO sandbox game similar to Rust or DayZ. It was initially developed by a small team over at SOE and runs on the Forgelight engine (Featured in Everquest 3 and Planetside 2). Sony Online Entertainment was later sold to investment firm Columbus Nova and now renamed Daybreak Game Company running as an independant game studio. The current multiplayer gameplay modes available are PVP,PVE and BR (BattleRoyale). The concept in PVP and PVE is relatively simple – loot buildings, caravans, dustbins and the surrounding enviroment – to conserve resources (hydration, stamina, energy and health) survive NPC and attacks from other players. There’s a crafting system as well as basebuilding to create a simple shelter by yourself or an complex fort with friends. Battle Royale is a 200 player Last Man Standing game-mode where you are para-dropped into a map to scavenge for weapons and med supplies, taking out players along the way before the 20 minute time bracket is over. When the initial 20 minutes are over, a toxic gas cloud is released shrinking the map constantly, creating havoc amongst survivors in the process until ofcourse one man is left. The last few players get progressive rewards according to their position. As we have covered the basics, it’s time to get into detail with a performance analysis. I’m sorry to say but the GTX970s will not be continued in the benchmarks as I have RMAed them (due to the memory controversy). I was using them on my main rig for 4K and was facing severe stuttering issues when running over 3.5GB. I have added the GTX680s in their place and will also be testing with the R9-290s. SLI and Crossfire while being fully utilised on H1z1 causes burn in image overlap and very high frame times that deemed the game unplayable. I will be sticking to single GPU benchmarks at 1080p, 1440p and 4K in the test below. Benchmark Components H1Z1 Benchmark System CPU i7 3930K @ 3.8Ghz Motherboard Asus P9X79 Deluxe RAM Kingston Hyper-X 16GB 1600mhz HDD/SSD Crucial M550 256GB GPU Asus GTX680 Reference 2GB , MSI R9-290 Gaming 4S Monitor Asus PB287Q 28″ 4K 60Hz 1MS PSU Corsair AX1200W OS Windows 8.1 Drivers Nvidia: 347.25 WHQL, AMD: Catalyst Omega 14.12 Testing Methodology For the test runs I join a very populated PVP map, team up with someone and kill an NPC. The performance test is shown below in the video, a simple 60 second run recorded with fraps. For the 1920×1080 and 2560×1440 resolutions I use the Ultra preset with Ambient Occlusion OFF. For 4K the medium preset, again with Ambient Occlusion OFF. H1z1 1920×1080 Benchmark Even a 3 year old card like the GTX680 seems to run the game pretty well on the Full HD resolution on the Ultra preset. The dips are not visible on either of the cards and the gameplay is very smooth compared to other titles. 2560×1440 Benchmark At the 1440p resolution we see minor drops on both cards with the GTX680 hitting 21 FPS minimums but still holding 30 FPS average. The numbers shouldn’t fool you, the game is playable on both GPUs even with a lot going on on-screen. In most newer titles frame rate drops like these usually mean an inferior experience, the Forgelight engine acts slightly differently here with no visible stutter (of what I could see). 4K Benchmark Because 4K is a whole different beast I prompted to drop the preset down to medium. The GTX680 doesn’t seem to have the horsepower and is marginally playable pushing so many pixels (even at the lowest of settings). The R9-290 seems to handle the game a lot better than what it did at 1440p, with average frame rates over 15% better. CPU Usage Lowering the resolution to 1024×768 to test CPU usage, we see that only 2 threads are utilised on the far left of the graphs. One is just below 40% and the other is bouncing around 70-90% usage. Even an older dual core processor will handle the game relatively easily. System Memory Usage Looking at the memory requirements from SOE, the minimum states 4GB and the recommended 8GB. You can see above that just over 4GB are used in all the resolutions tested. You should be able to get away with running ~3GB on lower presets. VRAM Usage Many people will be happy as the games VRAM usage only touches 1GB, even at the 4K resolution. Most if not all older graphics card should be able to run the game without any VRAM bottlenecks. Conclusion H1z1 might not be as visually appealing as Dying light, most would say it looks more like an older console game (GTA San Andreas maybe?) but we are still in it’s Alpha state and things “might” change. In general the game runs fine when in either PvP or PvE mode but it takes quite a grind to get up to par with other players and their inventories. You will need to team up if you haven’t found any weapons yet and should be careful as players get overly aggresive and kill anyone to steal gear. Battle Royale is quite a mess in terms of server que at the moment, they have just released a patch and hopefully it will be easier to get into a game. The BR lobby fills up very quickly and at times took me up to 30 minutes to play. The game is currently sold for $19.99, albeit its final version release later this year (no ETA) as a free-to-play. Most will disagree with a price tag on a F2P for Early Access, but for now BR entries are unlimited while on release will be paid for or tickets found in-game. To each his own ofcourse. If you have any questions about the benchmark please ask in the comment section below and I will be glad to answer.