They been doing it for years now with the proving grounds. FW can’t live up to it’s ideal so they needed another tool to give 1v1s?
Just use FPS limit.
It uses 100% of gpu because you are telling your computer to generate as many frames as it possibly can. Which means 100% of hardware utilization.
4 Likes
FW was a compulsion loop injected into organic ecosystem by CCP. That was not player made content . And low sec did not die because of players…
Blaps,blops,blings,blocks you give a name… none of these caused the end of low sec…
They clutter and inject so many things into game…instanced PvP, instant dungeon finders… you give the name …
They Turned game into theme park … which destroyed the very fabrique specialities makes this game great .Once upon a time there were more player made organic content and of course more players… more they introduced themes , they lost more players… they will never be able to reach out those numbers again …
as thumb rule industry crafted ■■■■ might look nicer and away from impurities… but hand crafted and self created everything weigh and value more in so many cases… both physical and emotional lvl …
if they dont achieve to create something totally different from all these broken pieces… and circulate by refreshing most of the playerbase with a totally new “Tabula Rasa” player base .
So cocroaches, bugs like me must go . Then Will be possible to have a new age for them.
assets take the same time as search probably just because someone at ccp changed to open asset’s search tab per default instead of ‘all items’
The way I would explain the changes in a simple way (I like simple explanations) -
A big part of the new changes are designed to streamline the process of sending messages (information) from the CCP servers to client (player) computers.
Think of as writing lots of letters/messages to many people.
The current way it is done - Python, GIL:
- The tiny Oompa Loompas in the server write a message, which is the main thing.
- They have to fold it and put into envelope, and make sure the address lines up with the little ‘window’.
- They have to lick the envelope and seal it.
- They have to find the right amount of stamps depending on the destination, and lick, then stick.
The new way CCP is working on - Python, GIL:
- The Oompa Loompas write the message.
- They throw it in a box. (buffer)
Easy peazy. Less steps, so now they can write a lot more messages in the same time.
The rest of the steps still have to be taken care of to be able to send the messages, so there is now a 2nd separate department/software with Gerbils dedicated to this - gRPC:
- The Gerbils take each message out of box. (buffer)
- Fold and put into envelope.
- Lick and seal.
- Get stamps, lick and stick.
Each department (software) focuses on their specific area - and things get done faster… Theoretically.
1 Like
More likely that our complete list of assets isn’t sitting in a cache, so when you call it, a database query is needed.
For someone like Tipa, who has been around for a while, that query takes enough time that she noticies a client freeze because the client is waiting for the response and is blocked until that happens. It appears not to be asynchronous.
Also possible that the list is sent in several packets and the client is waiting to receive them all before displaying the assets in the window.
It seems like one of the least optimal parts of the code, but has been around a long time too.
Yeah, I‘m not even asking to have my mess sorted faster (there is hardly a lowsec system where I have no stuff dumped), but that this happens in another thread asynchronously and not freezing the game client.
2 Likes
Just a wee tip:
A gigabit line, with large packets won’t take up a full core.
A gigabit line with small packets can easily hose a system.
It’s not just about the bandwidth. packet counts are important too.
The new character sheet is not done with this new approach, is it? Because it freezes the client hard every time it has to calculate the wealth values.
As all popups freeze the client, I think it’s more a client design fault not being multi-threaded. If the popup comes up fast, it isn’t noticeable though.
I think you are right about this, when I upgraded my pc the industry window became way more responsive, especially when first opened.
will this also allow cosmic signature scans to be saved over downtime or logging in general?
I don’t think that all these messaging systems are meant for realtime applications.
I didn’t either, but when we see average case scenarios of 4ms overhead with strict tech like RabbitMQ it makes us wonder what something like NATS could cut that down to.
1 Like
Oh my!
Breaking down the monolith with event-driven microservices!
My favorite architecture pattern in my favorite game!
Best of luck with this, guys. So excited to see more.
For a server, which has to take the flood from from a lot of clients, yes, it’s possible.
For a client, which was actually discussed in the post, no, not even maximum reasonable amount of packets multiplied by 10 would amount to any significant load to a modern CPU.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.