Red Guard
What if culling is unfixable?
Culling probably can’t be eliminated. It’s a game of numbers that is stacked against the developers; simply sending the positions of all the players on the map would likely blow the pipe.
However, it can very likely be significantly improved. I suspect the improvements they are talking about might involve things like variable-length encoding and so forth, but these things take time – especially in the QA department: they can’t just release what they have, mostly because threads such as this one would explode everywhere complaining about “how they broke WvW.”
Epistemic.8013: Guys this is bullkitten a sentient plant creature is hitting these
wooden doors with fireballs and it’s working.
If culling can’t be fixed…
I would be disappointed. However, while there is still beer and pizza in the world I don’t think it will cause me to jump off a ledge any time soon.
Looney vids at http://www.youtube.com/feed/UCRhCtfrF9GhxU1CoeZSN0kQ/u
Midnight Mayhem
Culling is absolutely a fixable problem, and as Habib mentioned in his MMORPG interview, we are planning to completely remove it from the game in march.
Why is it taking so long? Well, I think the best way I can describe how we’ve had to approach this problem is that it’s almost like changing a flat tire on a car that’s still moving. That’s a really rough analogy, but hopefully it helps illustrate just how difficult this would be in terms that are easier for everyone to understand.
Something like the way players render is extremely core to the engine and any change to that sort of thing could cause all kinds of other issues throughout the game. There are a lot of really complex technical problems that we needed to solve, then make sure that our changes don’t actually cause any of those odd side effects.
That being said, we’re confident that been able to solve those problems on the technical side. We still have a bit more testing to do, but we’re not expecting any major issues to pop up at this point, so we’re really looking forward to flipping that switch and getting rid of culling for good!
We’ll have a blog post coming up sometime in the next month to go into more details about the culling fixes and what you can expect.
So…you guys spend your time doing this? http://youtu.be/l2rC0EMQeS0
Tarnished Coast [TC]
Could we expect to get better fps too?
Pretty much no.
Culling improves performance by processing less stuff that needs to be drawn (primarily), drawing less (primarily), and requiring less network traffic (secondary).
No culling = your PC will be working harder on several fronts. In some network conditions you may experience more frequent “rubberbanding” or other incidents of latency/packetloss.
Culling probably can’t be eliminated. It’s a game of numbers that is stacked against the developers; simply sending the positions of all the players on the map would likely blow the pipe.
No offense, but you don’t know what you’re talking about there. Other MMORPGs have maintained several hundred players on-screen without issues. WAR and DAOC had to exceed 300-500 players in close proximity to start crashing things (although lag would occur with fewer than those numbers).
There’s virtually no way a single borderland or EB could tank using hardware and systems 5-10 years more advanced than those games with only ~150 people in the area.
If that does happen, then ANet’s technical competence is a joke.
Sidenote: Based on my recollections of JQ v. SBI v. ET and HOD v. JQ v. SBI in T1, the borderlands do hit excessive lag with >150 people in close proximity. So yeah, ANet has their work cut out for them. Hopefully they’ve worked out what they needed to and the removal of culling will not result in tons of excessive lag.
I’m glad we have all these experts to tell us what’s gonna happen/not happen…
On a game they didn’t program
On a system they didn’t build
On servers they didn’t install
I have a programming degree and I’m not arrogant enough to assume what is/is not going to happen here.
Now Musty Britches since someone decided Shortbus Rider was offensive… [LoS] [NSP]
(edited by Musty.3148)
I’m glad we have all these experts to tell us what’s gonna happen/not happen…
My favorite so far is the person who claimed to know the exact specifications / capabilities of ANet’s servers.
I’m glad we have all these experts to tell us what’s gonna happen/not happen…
On a game they didn’t program
On a system they didn’t build
On servers they didn’t installI have a programming degree and I’m not arrogant enough to assume what is/is not going to happen here.
^^ this
On the topic of performance in WAR… the Fort fights in the first 6 months of the game were not even playable. “lag” doesn’t even begin to describe having your character respond every 15-20 seconds and Mythic claimed they weren’t even aware of these problems when they first started… meaning they hadn’t even play-tested a fort siege with anything close to real numbers of participants.
Let’s let ANet address this and we can let them know after the patch if it’s working or not. Pre-supposing solutions and potential problems is useless.
Culling is absolutely a fixable problem, and as Habib mentioned in his MMORPG interview, we are planning to completely remove it from the game in march.
Why is it taking so long? Well, I think the best way I can describe how we’ve had to approach this problem is that it’s almost like changing a flat tire on a car that’s still moving. That’s a really rough analogy, but hopefully it helps illustrate just how difficult this would be in terms that are easier for everyone to understand.
Something like the way players render is extremely core to the engine and any change to that sort of thing could cause all kinds of other issues throughout the game. There are a lot of really complex technical problems that we needed to solve, then make sure that our changes don’t actually cause any of those odd side effects.
That being said, we’re confident that been able to solve those problems on the technical side. We still have a bit more testing to do, but we’re not expecting any major issues to pop up at this point, so we’re really looking forward to flipping that switch and getting rid of culling for good!
We’ll have a blog post coming up sometime in the next month to go into more details about the culling fixes and what you can expect.
I’m afraid this will cause even more severe FPS drops than many of us, even those with high-end hardware and CPUs clocked beyond 4,5GHz, currently experience, did you work on this as well?
http://i.imgur.com/I1f1SC1.jpg
GPU usage drops to ~25%, CPU is probably more loaded than the graph shows… or there’s something else limiting in the client, memory isn’t a problem. There’s something very wrong, playing around with core parking and HT doesn’t help, further overclocking does, my Core i7-920 won’t be stable beyond 3,7GHz though and I doubt clocks beyond 4,5GHz should be a requirement for such a game. Changing the graphics options does pretty much nothing at all.
On the topic of performance in WAR… the Fort fights in the first 6 months of the game were not even playable. “lag” doesn’t even begin to describe having your character respond every 15-20 seconds and Mythic claimed they weren’t even aware of these problems when they first started… meaning they hadn’t even play-tested a fort siege with anything close to real numbers of participants.
Fort fights in the first 6 months had 200 players in sight range. Over a hundred players within bow shot.
Also, GW2 has already had incidents of skills being delayed 10-15 seconds.
Just saying.