Nerd Forum: Talk about Unreal Engine 4 and its impact on Mirage:AW.
I’m curious to see how TBS has worked around Unreal Engine’s only real downside (imo): the darn thing is so CPU intensive! Everyone loves how Chivalry was able to host massive amounts of players in one game. The frame rates were great, all things considering, but the biggest proverbial thorn in the side was CPU dependency. Between the collision detection system and dynamic lighting, most players couldn’t handle over 24 players comfortably. Their CPUs were not physically capable.
I remember my modest little 2550k chugging away at 100% usage darn near every game. Even when cranked 1.5GHz above stock, it just couldn’t cut the mustard sometimes. Meanwhile my 970 was taking a nap in the background.
How has TBS dev team overcame these issues? Can we the community expect to nerdgasm at 32 and 64 player matches, or has the CPU intensity of UE4 forced our hands to go ham on the hardware side?
Forever your loyal fan from the beginning, Sushi.
P.S. My 4930K is waiting to crush the beta!
Oy last edited by
With all the intense magic effects and also just from what we’ve seen so far, I’m expecting the game to only really be 6-12 player skirmishes, the class variety seems limited and you can probably only have 1 of each character per team, meaning multiple small teams or 2 slightly bigger teams. If this game gets mod tools we may see large-scale battles at some point but the duplicate characters running around from not having a very large roster means their abilities will overlap in bad ways. Think of how a group of archers could only be beat by another group of archers. Well how would a group of maceguys or daggerchicks be countered?
It just seems to me that this game, unlike chivalry, was designed for smaller battles with unique classes countering each other like TF2, whereas chivalry gave everyone a more-or-less equal loadout and was skill and team dependant. Mirage will be the same, but different.
DevilsAdvocate last edited by
I’ll copy and paste what i said on another thread.
I’ve fooled around with UE4 a bit now that it’s free to download, while the CPU is definitely a factor it seems to be mostly GPU heavy, it also is really scalable. I’ve already been playing some UE4 alphas/betas (Paragon, Fortnite, and ARK) and I can max out Fortnite, almost max Paragon, and run ARK on Medium settings with GTX 760. intel i5 4670k, 16GB ram on Windows 10.
The graphics on this game look like they are between Fortnite and Paragon (ARK is not a fair comparison because it’s the only open world game)
And I’ll add that UE4 is the best performing non custom engine I’ve seen and in the games I can’t max out the CPU is not the bottleneck.
TBPhoenix last edited by TBPhoenix
Oy’s right, Mirage is designed to be 6v6 (for gameplay, not optimization reasons).
Because of the nature of our combat, CPU is always a concern because animation accuracy matters a lot to the experience of Mirage (telegraphing attacks, drags, etc.). In addition, we have a new challenge with the abilities because the particles can be quite heavy, both on the GPU and CPU.
That said, we’re currently working on some optimization, and UE4 gives us some very good tools to evaluate performance.
edit: in terms of clarity between UE3 and UE4, the comparison is pretty complicated. We aren’t trying to do the same things as other games, so it’s hard to get a good comparison between them and us, as well as between 3 and 4.
Wilt last edited by
I do recall a similar speculation as mr Oy, on the other thread about teams and sizes, plus also having spent many months
in UE3 making my angels weep map ( which I then completely forgot how to even use the editor due to a very long netflix trip)
that a while ago I dabbled in UE4 with an uploaded level of my fps game, that the cpu was not as loud or thrashed as it
used to sound even in the UE3 Editor , let alone the game sometimes. Overall TB made the right move to UE4 as I did read
its easier to get bugs and find them. plus they got to learn new exciting stuff, this time around its a game they made from
scratch more or less.
I have no doubt UE4 is far superior to UE3. Factoring particle effects, CPU physics, and all the other jive, twelve man games seem a very viable choice from a technical standpoint. I like the interoperability of UE3 and how it allowed for cross game texture packs and more. I look forward to hearing Dev updates regarding UE4 and everything TBS is doing from a technical standpoint.
I do want to stress that the twelve person games are a design decision rather than a technical one. That’s where we’ve found the abilities remain manageable and deliberate, it gets very chaotic as the player counts scales up.
Beyond optimization, there’s a lot of reasons that UE4 is a great choice for us over UE3. 4 has a lot of features that allow a small team like ours to work much more flexibly, even allowing non-programmers to fully prototype or implement gameplay and rendering features as well as complex shaders. The rendering engine is now physically based (very nice reflectivity and better grounded materials and surfaces), and supports some great lighting features that have added some life to our environments.
I’m sure other folks have reasons they enjoy it, but from the art side, it’s visually head and shoulders above 3.
Huge battles and chaos was what made chivalry a lot of fun, seems like mirage is going for a more team focused and tactical approach with gameplay this time. This makes me wonder are we going to see a ranked mode similar to league of legends? I know the competitive community and competitive players like myself would love to have something like that! (Imo thats what makes league so popular)
@2ndchallenge I have no doubt there will be high player count servers much like Chivalry; I am looking forward to it! It also depends on the map. Playing Hillside on Chiv with more then 24 people can get crazy. Imagine a bunch of mages with AOE spells guarding a choke point.
AnoniMouse last edited by
@2ndchallenge Surely the huge battles were great in Chiv (and I still’d like to see them in Mirage) but one should remember it’s a far faster game with all the magic. I still cannot decide what I’d like more Chiv on UE4 with even bigger battles or Mirage… ^^
Hey 2ndchalllenge, comp play is a huge focus for us, but we can’t really talk about matchmaking much yet. Also, in Chiv, it’s rough when you exit the new player servers and enter the general pool of players that are just on a whole other level and pulling moves that are totally new. A lot of people bounce off at that point, so it’s an experience we want to improve.
As to the player count, despite the glorious chaos of 32 (or even 64) player servers in Chiv, only a few players can engage each other at a time just by virtue of the fact that it’s primarily a melee game. Mirage is very different because players can be disruptive or lethal from a large range of distances. Our focus in combat is on readability and reactability, so keeping the player count lower than Chivalry enables us to make more dynamic and interesting abilities that remain reactable.
Its Just Milk last edited by
@TBPhoenix are you able to say whether or not it will be possible for community servers to increase the player count?
Sorry, can’t confirm at this time.