Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Valorant's 128-Tick Servers (2020)

technology.riotgames.com

Post reply on HN
Print view View original

111–120 of 136 posts

Re: Valorant's 128-Tick Servers (2020)

#111

Post by acron0 »

bakugo wrote: post #69
Thaxll wrote: post #39

Earlier quoted context omitted.

Fortnite has to handle 100 players on the same process, very different from 128hz @10players max.

Nowadays it's more like 20 players and 80 bots, so a lot less networking stuff going on, and the bot AI is so basic that I doubt it has a significant impact on server performance unless it's very badly implemented.

Also assumes that the bots coexist on the server. My first thought would just be to connect them like any other client, with compute of their own so the server doesn't even know it's a bot.

Re: Valorant's 128-Tick Servers (2020)

#112

Post by Fizz43 »

deathanatos wrote: post #7

128 ticks per second servers. (And lo, suddenly the article's thesis is inherently clear.) A "tick", or an update, is a single step forward in the game's state. UPS (as I'll call it from here) or tick rate is the frequency of those. So, 128 ticks/s == 128 updates per sec. That's a high number. For comparison, Factorio is 60 UPS, and Minecraft is 20 UPS. At first I imagined an FPS's state would be considerably smaller…

A better comparison would be another shooter. Counter strike has had 128 tick servers for a decade.

Re: Valorant's 128-Tick Servers (2020)

#113

Post by beAbU »

deathanatos wrote: post #7

128 ticks per second servers. (And lo, suddenly the article's thesis is inherently clear.) A "tick", or an update, is a single step forward in the game's state. UPS (as I'll call it from here) or tick rate is the frequency of those. So, 128 ticks/s == 128 updates per sec. That's a high number. For comparison, Factorio is 60 UPS, and Minecraft is 20 UPS. At first I imagined an FPS's state would be considerably smaller…

I'm completely speculating here:

I think for FPSes, the server relies on the client for many of the computationally intensive things, like fog of war, collision detection, line of sight and so on. This is why cheating like wall hacks are even possible in the first place: The client has the total game state locally, and knowing where to look for and extract this game state allows the cheater to know the location of every player on the map.

If the server did not reveal the location of other players to the client until the server determined that the client and opponents are within line of sight, then many types of cheating would basically be impossible.

Re: Valorant's 128-Tick Servers (2020)

#114

Post by AnssiH »

nemothekid wrote: post #28
greatgib wrote: post #25

This deep dive article is very nice. At any given time, ~50 of those games are going to be in the buy phase. Players will be purchasing equipment safely behind their spawn barriers and no shots can hurt them. We realized we don’t even need to do any server-side animation during the buy phase, we could just turn it off. That explains the current trend of "online" video game that is so annoying: For 10 minutes of play,…

This is not equivalent to "lobby time or end game animations" in other games. In Valorant (similar to Counter Strike), at the start of the game you have 60 seconds to buy your weapons and abilities for the round. Valorant/CS is typically a best-of-13, and before each round is a 60 second "buy" period.

> In Valorant (similar to Counter Strike), at the start of the game you have 60 seconds to buy your weapons and abilities for the round. Valorant/CS is typically a best-of-13, and before each round is a 60 second "buy" period.

Valorant's buy phase is 30 seconds, with +15sec at start of match and halftime.

Re: Valorant's 128-Tick Servers (2020)

#115

Post by Hikikomori »

beAbU wrote: post #113
deathanatos wrote: post #7

128 ticks per second servers. (And lo, suddenly the article's thesis is inherently clear.) A "tick", or an update, is a single step forward in the game's state. UPS (as I'll call it from here) or tick rate is the frequency of those. So, 128 ticks/s == 128 updates per sec. That's a high number. For comparison, Factorio is 60 UPS, and Minecraft is 20 UPS. At first I imagined an FPS's state would be considerably smaller…

I'm completely speculating here: I think for FPSes, the server relies on the client for many of the computationally intensive things, like fog of war, collision detection, line of sight and so on. This is why cheating like wall hacks are even possible in the first place: The client has the total game state locally, and knowing where to look for and extract this game state allows the cheater to know the location of ev…

Both CS2 and Valorant use fog of war, and so does LoL and Dota 2. It works because they have simple geometry in their maps, mostly straight corridors, 90 degree turns so it has blocking walls in all directions. They still have to start sending player information before they get to the corner to avoid pop in effects, so they'll use the map geometry and then adjust the fog of war mesh.

Open world games or if you have complex geometry like buildings with windows and such it's not really worth the effort to implement its useless in most areas of the map.

Sending all to all clients is the simplest and easiest implementation, all other options are much harder to implement and adds a lot of load on the server, especially if you have more players than 10.

Re: Valorant's 128-Tick Servers (2020)

#116

Post by greatgib »

nemothekid wrote: post #28
greatgib wrote: post #25

This deep dive article is very nice. At any given time, ~50 of those games are going to be in the buy phase. Players will be purchasing equipment safely behind their spawn barriers and no shots can hurt them. We realized we don’t even need to do any server-side animation during the buy phase, we could just turn it off. That explains the current trend of "online" video game that is so annoying: For 10 minutes of play,…

This is not equivalent to "lobby time or end game animations" in other games. In Valorant (similar to Counter Strike), at the start of the game you have 60 seconds to buy your weapons and abilities for the round. Valorant/CS is typically a best-of-13, and before each round is a 60 second "buy" period.

If you look at their own stats, they says that for a server that host 150 games, at any given time, they have around 50 games that are in the buy phase.

I don't know how long game rounds last, but if you tell me that you have 60s only locked in that state, that means that the playable game round last around 2 minutes.

And anyway, that you will spend on average 1/3 of your time not playing when you came to play.

Numbers look strange, as 2 minutes looks very low to me for a party, but I don't know how you can explain the ratio of 50/150 otherwise?

And just as an aside, the article in itself tells you that they capitalize on you being in the buying phase to reduce server costs. So looks like that even if they could do something on the game play side to improve that, there would be financial incentives not to do so.

Re: Valorant's 128-Tick Servers (2020)

#117

Post by simjnd »

greatgib wrote: post #116
nemothekid wrote: post #28

Earlier quoted context omitted.

This is not equivalent to "lobby time or end game animations" in other games. In Valorant (similar to Counter Strike), at the start of the game you have 60 seconds to buy your weapons and abilities for the round. Valorant/CS is typically a best-of-13, and before each round is a 60 second "buy" period.

If you look at their own stats, they says that for a server that host 150 games, at any given time, they have around 50 games that are in the buy phase. I don't know how long game rounds last, but if you tell me that you have 60s only locked in that state, that means that the playable game round last around 2 minutes. And anyway, that you will spend on average 1/3 of your time not playing when you came to play. Numbe…

> you will spend on average 1/3 of your time not playing when you came to play

The buy phase is playing. It's coordinating with your team on what loadouts to get to best counter the enemy team. The decision making is fun and is play. Because you're not shooting at other players does not mean it is not playing?

Re: Valorant's 128-Tick Servers (2020)

#118

Post by robertlagrant »

charcircuit wrote: post #68
Thaxll wrote: post #39

Earlier quoted context omitted.

Fortnite has to handle 100 players on the same process, very different from 128hz @10players max.

That's not an excuse to give up on performance. The map is also much bigger which spreads people out.

That's like saying Valorant have given up on performance at only 128 ticks for 10 players.

Re: Valorant's 128-Tick Servers (2020)

#119

Post by baxuz »

wyldberry wrote: post #81
Havoc wrote: post #66

You can mess with the code all day long, but you're not getting away from raw latency. The modern matchmaking approach groups people by skill not latency, so you get a pretty wild mix of latency. It feels nothing like the old regional servers. Sure the skill mix was varied, but at least you got your ass handed to you in crisp <10ms by actual skill. Now it's all getting knife noscoped around a corner by a guy that rub…

Good thing they thought of that. Disclaimer: I was at Riot During some of the Valorant dev cycle and the stated goal in this tech blog [0] was a huge goal (keeping latency This was only really doable because Riot has invested significantly in buying dark fiber and peering at major locations worldwide [1][2] [0] - https://technology.riotgames.com/news/peeking-valorants-netc... [1] - https://technology.riotgames.com/ne…

This is excellent material!

Re: Valorant's 128-Tick Servers (2020)

#120

Post by philipallstar »

Dylan16807 wrote: post #103
forrestthewoods wrote: post #102

Earlier quoted context omitted.

Also twice as many dollars in the future term! The downside very much does not decrease “exponentially”. It would be interesting to know what Valve’s server costs look like over time. They definitely spend the pretty penny. And any business would prefer to spend one penny rather than two.

> Also twice as many dollars in the future term! The downside very much does not decrease “exponentially”. The downside measured in dollars continues to decrease exponentially. If it's $100 today, soon it will be $30. It's still "twice", but the thing you're twice-ing is smaller and smaller and smaller.

I'm not sure this is true any more. Long gone are the late 90s when computing power per dollar would increase that fast.
Post reply on HN