Rodarg

Programmer by day, debugger by night

[Rodarg @RodargDev]

@Simon_Hypixel @ThatBram0101 @Hytale Any chance you could have a dev look into InteractivelyPickupItemEvent? It looks like a dead event right now which prevents me to add functionality whenever player picks up item.

[Rodarg @RodargDev]

@flash_canadian Do you still need help with this? You pretty much just move you singleplayer world folder and memories to your server and then your golden.

[Rodarg @RodargDev]

@Jamezbower @RuneteriaServer Any idea if something like pandora hallelujah mountains is doable?

[Rodarg @RodargDev]

@Simon_Hypixel I dont agree, i think the current mining mechanic is good but probaply just add a way to get more from a single ore in late game. I really dislike if i try to mine a block and the signature ability blows a hole into the wall.

[Rodarg @RodargDev]

@Simon_Hypixel isnt it 4:45 am right now for you?

[Rodarg @RodargDev]

@Simon_Hypixel @felinusfish

[Rodarg @RodargDev]

@Simon_Hypixel @not_mynes @_Hawkon @not_mynes

[Rodarg @RodargDev]

@not_mynes @Simon_Hypixel @_Hawkon Just let it be man, its really not that deep. Whenever the game releases thats the state we start with but currently we don’t have a finished game and everything is based on blogposts

[Rodarg @RodargDev]

@CottonTheFoxUwU @slikey @theRangerGilan @Hytale GPU doesnt really matter in this case i think, mainly CPU and RAM. But depending on the amount of players 16GB will probaply be fine for you and some friends.

[Rodarg @RodargDev]

Hytale Server Performance Comparison
Yesterday i asked @slikey a question about @Hytale server performance numbers so i could get a rough idea on what the performance would be like compared to Minecraft, since Minecraft's server performance is quite bad especially for larger SMP servers.

So let’s compare the performance numbers given by Slikey. I do want to add a big disclaimer here because the game is still in very early development, and as Slikey said himself, he believes the server can be optimized further. Also, I am a fan of Hytale and do not in any way want to bash the server performance. I was just curious how the servers in their current state compare to Minecraft servers.

TLDR
Hytale’s minimum server specs are likely higher than Minecraft’s for small player counts, but its server software appears to take better advantage of multiple CPU cores. If this holds true, Hytale should scale more efficiently for larger SMPs and minigames as player counts increase.

That being said, let’s look into the values.
Slikey mentioned that a 3 vCore, 8 GB server can sustain 6-8 people. The first thing I notice with that statement is that he is specifically naming the 3 vCores. In Minecraft, the game is single-threaded and only uses one core. Newer server implementations like Paper do offload some of the work across multiple cores, but still have to deal with the single-core problem. Since Hytale is built by Minecraft devs, who probably know how annoying the single-core problem is for scalability, I have a feeling they may have solved this in some way. Because it’s currently very unclear how the game is built and how it uses multiple cores, it’s kind of hard to make a good prediction based on the vCore value. The value we can use for a prediction is the RAM, though, since RAM is easier to compare than CPU behavior, but still highly workload-dependent.

After looking through multiple hosting providers and other sources for their RAM requirements for a 6-8 player count (vanilla Paper), I found that many hosting providers average the amount of RAM used to be between 4-6 GB. This means that default vanilla Minecraft is probably performing better than vanilla Hytale if the load is not very high. I do have a feeling that the testing team ran in different directions to see what the limits would be.

The comparison does shift, though, when looking at modded Minecraft servers. Because Hytale is built with modders in mind, I do have a feeling these values should be taken into account. When looking at modded Minecraft servers and their RAM requirements for heavy modding, they average about 8–12 GB. This means that, compared to a modded Minecraft server, the Hytale server might be performing just as well or even better.

Later in the tweet, Slikey mentions that on a Ryzen 7950X with 12 GB of RAM, they were able to hit 70 players, which seems like a very good player count for those specs. Looking around, I found that a 12 GB server is normally advised for 30+ players, but I have a feeling that hitting 70 players with 12 GB would be very difficult in Minecraft, especially with mods involved.

The exact scenario for these benchmarks is unclear, but the numbers suggest that Hytale’s server software is likely much more scalable than Minecraft, taking better advantage of multiple CPU cores. While the minimum specs are slightly higher than a typical Minecraft server, the ability to handle large player counts is impressive, Hytale servers could scale really well once fully optimized.

I hope the Hytale team will share a blog post about server performance, as it would be really interesting to see more detailed data. If not, I’m planning to try running some benchmarks myself on January 13th.

Another small disclaimer: This is by no means a thorough benchmark, just some quick research and my thoughts based on available information.

Sources:
https://t.co/bUOFV7chIX
https://t.co/kau4AVhgsH
https://t.co/n3IHwIJmh5

[Slikey @slikey]

I wouldn't call it a blog post. It is more of a technical manual with some example data. Just wrapped up a playtest with 70 people on various hardware until we finally crashed the server.
There is a lot of context such as "minigame server" vs "exploration mode SMP". We believe that a 3 vCore + 8 GB server can sustain 6-8 players. A Ryzen 7950x with 12 GB RAM server, we were able to push 70 players but TPS degraded until the server crashed (~45 minutes).
I think we will defintiely see 200 player minigames such as Mega Walls or Super Sky Wars. While vanilla SMP is going to require some beefy hardware to push 100 players in a single SMP.
We are continuing our effort to improve stability and further add performance optimizations to get those big anarchy servers.

[Rodarg @RodargDev]

@slikey @Hytale Thanks for the reply! Since the trailer i’ve always been very curious about the server performance because i really like the idea of a giant SMP with a load of players. Thanks for clarifying and good luck with working towards the launch!

[Rodarg @RodargDev]

@slikey So will we get some kind of blogpost about the server performance/details before the Early Access or will you guys leave it up to community to get some benchmarks done on the server? Im actually very curious about the performance numbers.

[Rodarg @RodargDev]

Hey @slikey is there anything you guys can say about client hardware and server performance?

[Rodarg @RodargDev]

@slikey @slikey is there any update on when we get more server details? Or will the server hosting manual be released at the same time as the game?

[Rodarg @RodargDev]

@Simon_Hypixel Please get @flash_canadian a code brother

[Rodarg @RodargDev]

@flash_canadian Lmk whenever you get one, think you do have the best content of the current YouTubers so would gladly use it.

[Rodarg @RodargDev]

@flash_canadian will you have a creator code at launch?

[Rodarg @RodargDev]

@Simon_Hypixel @slikey Any idea at what time we can purchase Hytale tomorrow?

[Rodarg @RodargDev]

@Simon_Hypixel @not_mynes @Prowl8413 Any idea then we will get the details on server performance, very curious about it

[Rodarg @RodargDev]

@Lyall_PD Still very curious about server performance, really hope you guys nailed it, so annoying with Minecraft right now