Camelot Unchained surpasses last week’s test, putting nearly 3500 bots and humans on a battlefield

    
9

Were you impressed by Camelot Unchained’s news last week that it had managed to successfully put nearly 3000 humans and human-like bots (ARCs, autonomous remote clients that consume the same number of resources as a real person) on the same battlefield? Be more impressed now, as beta one has expanded that range further.

“Believe it or not, among all the other ongoing work this week, we managed to push that number upwards of 3000! (Actually, even further to 3500 before things began to degrade.),” City State Entertainment’s Tyler Rockwell wrote in the latest dev missive.

The team notes it’s continuing work on the ability system, battlegroups, mages, valkyries, siege effects, Luchorpán animations, UI buttons, and giants.

9
LEAVE A COMMENT

Please Login to comment
  Subscribe  
newest oldest most liked
Subscribe to:
Reader
yoh_sl

That number just keeps on going up doesn’t it? What does it take to satisfy you lot.
On a slightly more serious note, what exactly was the change that caused improvement? Was it net coding or is it client side?
Trying to wrap my head around the under the hood mechanics of it all.

Reader
Mark Jacobs

Hehe, yeah, we are definitely overachieving right now. I really didn’t expect we’d hit 4K until we did some tuning work on some of our servers/net code. But as people know who were in our very Impromptu test yesterday, we broke 4K before we broke the servers. :)

As to the change, we did a few things differently in this test including running it on newer AWS instances. Now, these are the same instances that weren’t as performant in other tests so we were surprised at that. The reason, we think, is just a number of improvements that have gone into the code in the last month.

As to being satisfied, we are definitely at that point now. The test did confirm an issue we had with thread contention once we got to 3.5K and above that we need to look into. And at some point during Beta, we need to convert some of our servers from Windows to Linux as that will save a bunch of money (cloud services using Windows are in general 2x+ as expensive) and may boost performance. We also want to optimize those same servers which will also yield some cost savings.

We’ll also be running tests on Azure and GCP to see which of the major cloud providers has the best performance for this game. With cloud services it’s not as simple as comparing the hardware because the software that runs the cloud can have its own headaches, add lag, etc. We’re also factoring in whether the provider has its own fiber from US Europe so that we can be sure that our European Backers who want to play on US servers will have the best experience possible. I promised them during the KS that they would never be treated as 2nd class citizens and like all promises I make, I keep them as we’re shown with our GDPR compliance, testing hours, etc. We’ll also begin testing our servers in Europe during Beta 1 once we’re satisfied with their performance here since it’s obviously safer to test a version somewhere else once you know it’s working the way it needs to here.

Time to get geared up and head into the office. CU later! :)

Reader
Nick Smith

Wow!

Reader
Mark Jacobs

Thanks. It really is a sight to behold. As I’ve told some people who are skeptical (can’t blame them one bit), even if you just saw the players running around in the same tiny space at 60 FPS that would be a nice achievement but 3K is ridiculously cool.

Reader
Kickstarter Donor
Tandor

The questions is – will the human-like bots be paying subscriptions when the game launches ;)?

Hamblepants
Reader
Hamblepants

I thought the human-like bots were off playing P2W F2P cash grabs? :P

Reader
Mark Jacobs

Absolutely! We’re going to charge AWS for their subscriptions. I was inspired by Monty Python (again) and their legendary and brilliant way to raise more money for England.

Unfortunately, saying “We’ll sub all ARCs running amok” doesn’t have the same ring as “I’d tax all foreigners living aboard.”

:)

Reader
TheDonDude

Sounds good, but as usual the proof will be once players get their chance to try it out.

Reader
Mark Jacobs

We had 150+ players in one test this past week but I want to get it up to 1K players before we can say that we have no doubt that the engine can support that many in the same battle. Until then, we have to say that it should be able to support that many players in a battle. We’re pretty confident because our ARCs use the same ability system and their connections are treated the same by the servers but as I always say, here, there, and everywhere…

Time will tell!

:)