ScatterJoin Date: 2012-09-02Member: 157341Members, Squad Five Blue
Just been thinking about cysts and infestation lately. Is the cyst still the best way of spreading this or are there potentially other methods that could save us fps?
With performance as big a sticking point as it is, I would love to see us ruthlessly cut dubious mechanics like cysts. People don't play the game for cysts alone but they do leave the game for performance.
Are cysts an issue any more? I thought i remember them fixing that? I do agree with what you are saying though. I got a feeling people would stick around more if they were getting higher fps. Personally speaking I played around 300 hrs at sub 30 fps and it got tiring to the point I would have quit playing if couldn't have afforded to buy a new pc.
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
edited May 2013
I'd actually be curious to see screen shot comparisons of any processor that's over 3.5 ghz (because let's be honest here, the remaining fans and comp players have mostly OC for this game already - this also rules out the insufficient core 2 Duos mostly) lose more than 10 fps* with cysts in the room compared to not.
I have a four year old i7 920 and i can't see more than a ~3 fps difference in a giant room like cargo.
Now rich infestation on the other hand is different, but
A) comp players and therefore those that typically post in here don't use it cysts are the ones that keep getting brought up and getting the bad rep
C) removing cysts as just suggested this page would still leave said rich infestation in use
*My line of thinking is : there is way bigger culprits out there for fps loss that add up to much more than 10 fps at the end of rounds.. So why would you target such a small contributor if performance was truly the focus of your argument for removal?
Are cysts an issue any more? I thought i remember them fixing that? I do agree with what you are saying though. I got a feeling people would stick around more if they were getting higher fps. Personally speaking I played around 300 hrs at sub 30 fps and it got tiring to the point I would have quit playing if couldn't have afforded to buy a new pc.
They improved them, but I don't think they "fixed" them. I don't think they CAN fix them, tbh.
It might not be the cysts in a room per se, but rather the ever-increasing logic demands it places on rounds as more networks exist? I'm not technical enough to know for sure and I can't be arsed to test it but my anecdotal experience from xDragon's NS2c mod was that the removal of grid/infestation really improved the performance on my PC - that was a while ago however.
Without getting into the debate here, I would advocate the removal of those features irrespective of their affect on performance so you're absolutely right in identifying the not-so-hidden case I'm making. I just think that even +3-5fps (and I suspect it's probably more on long rounds) makes a difference at this stage!
Make Biomass give 8% of base HP per amount instead of the 2% or the like. Two full biomass hives would give as much HP as Carapace and Biomass currently gives. Remove Carapace and replace it with Reactive Armor. Any time you would take lethal damage or over 250 damage in .2 seconds, you gain 15/30/45% damage reduction instead for like .2 seconds. Would counter Shotguns/Railguns/Grenaderlaunchers and would be an interesting test.
It would be a way to remove Carapace and make it not the most important upgrade in the game.
I'd actually be curious to see screen shot comparisons of any processor that's over 3.5 ghz (because let's be honest here, the remaining fans and comp players have mostly OC for this game already - this also rules out the insufficient core 2 Duos mostly) lose more than 10 fps* with cysts in the room compared to not.
you are delusional if you think most NS2 players overclock
At one point I devs said that the code is not really event based because that would open completely different cans of worms than it would help performance. I have no clue to which extent this is true and I did at no point peek at the code to see how it actually works internally.
But based on that statement above, I really wonder if something is going on like
class Drifter
// executed once every frame
function tick()
{
....ForEach curEntity in allEntities
....{
........if (curEntity.isA(AlienStructure))
........{
............if (vSizeSq(curEntity.location - self.location) < drifterGrowRadius)
................growStructure(curEntitiy);
............elseif (vSizeSq(curEntity.location - self.location) < maxAutoGrowCheckRadiusSquared)
................moveTo(curEntity.location);
............break;
........}
....}
}
or instead still a somewhat event-based approach
class KhammanderInterface
function droppedStructure()
{
....[...]
....ForEach curEntity in allEntities
....{
........if (curEntity.isA(Drifter)
............Drifter(curEntity).checkDrifterGrow();
....}
}
class Drifter
// called once on all Drifters only whenver a structure is dropped by the khammander or the Drifter changed it's location
function checkDrifterGrow()
{
....ForEach curEntity in allEntities
....{
........if (curEntity.isA(AlienStructure))
........{
............if (vSizeSq(curEntity.location - self.location) < drifterGrowRadius)
................growStructure(curEntitiy);
............elseif (vSizeSq(curEntity.location - self.location) < maxAutoGrowCheckRadiusSquared)
................moveTo(curEntity.location);
............break;
........}
....}
}
The first example would perform distance checks on every tick and try to find close stuff. This works fine for things like Crags that are supposed to interact with a lot of dynamic things. Armories might do the same, though native collision hash checks could work as well to notify about close entities instead of checking on their own.
But the second approach seems so much better for performance since it only executes code and checks for stuff after something changed in the gameplay that would make it necessary to do so.
You could extend this kind of approach to other things as well: instead of checking every tick whether or not a cyst is still connected to it's parent (and checking recursively whether or not that parent still has a connection as well) you would only check the entire cyst network whenever a new cyst or hive is planted or whenever a cyst or hive gets destroyed.
Or for power nodes: instead of letting every structure check on every tick whether or not it's in a location with power, have the power tell the structures in the associated room about the changed state. (If it's not already done that way)
Again, I have no clue about the implementation details in Spark and these assumptions are only based on something I recall a dev saying in the past.
Btw, I didn't see anyone talk about the latest changes:
You seem to gain 0.1 Biomass levels per alien structure that you build, in addition to the usual upgrades at the Hive for 25 and 40 res. I am not 100% sure though if this "fake" Biomass is usable for upgrades or only as HP extension, though. I think it's the latter.
Also, T Res cost for alien eggs has been increased:
Gorge - 10 res
Lerk - 40 res
Fade - 70 res
Onos - 100 res
This surely combats powerful early lifeform strategies. I didn't pay any attention to whether or not the Biomass requirement for Fade and Onos eggs was changed accordingly, but I assume that was the goal of this change in the end. Still, that is quite a steep investment for pretty fragile lifeforms. May I assume that T Res drops from the marine commander are also increased in price, so we don't get the inflation of 30 T Res Exos?
Btw, I didn't see anyone talk about the latest changes:
You seem to gain 0.1 Biomass levels per alien structure that you build, in addition to the usual upgrades at the Hive for 25 and 40 res. I am not 100% sure though if this "fake" Biomass is usable for upgrades or only as HP extension, though. I think it's the latter.
Also, T Res cost for alien eggs has been increased:
Gorge - 10 res
Lerk - 40 res
Fade - 70 res
Onos - 100 res
This surely combats powerful early lifeform strategies. I didn't pay any attention to whether or not the Biomass requirement for Fade and Onos eggs was changed accordingly, but I assume that was the goal of this change in the end. Still, that is quite a steep investment for pretty fragile lifeforms. May I assume that T Res drops from the marine commander are also increased in price, so we don't get the inflation of 30 T Res Exos?
Having multiple ways to accumulate biomass is adding too much complexity. The system was already borderline too hard to understand. If anything I would rather that building things was the only way to accumulate biomass.
Marine bullets being bigger.
-it rewards random hits at long range, because of the loose spread on the rifle
-instead tighten the rifle spread, and reward true aim
Fade new movement.
-heavily reduced stamina
-slower mid/late game
-less spectator friendly because of restricted movement
*I don't think there is anything wrong with vanilla fade movement.
Skulk new movement.
-more like pre b240
-seems good, needs more playtesting, vs accuracy/bigger bullets
Haven't tried lerk, but I don't think anything is wrong with vanilla lerk, other than the amount of blinding tracers coming at you.
Just played for a bit again and I completely changed my mind about the movement.
Fade and Lerk movement is amazing right now and should stay exactly as it is. It's highly responsive and has tons of options. Now while the pseudo bhop thing is very enjoyable for the Fade, Skulk movement is still a little dodgy. The tolerance of how quickly you lose your speed when bumping into things or touching walls (unless it's head-on) needs to be way higher, as there's no way to quickly gain momentum again, at least not until Leap is up.
As for pretty much all other changes, I say bring 'em on.
I haven't really been following the discussion here so I´m sorry if this has been talked about.
We played a match yesterday with the new RT HP / build time changes, I have mixed feelings about this but wanted to post some input.
The game felt very good to play, it was a lot more fast paced than the current vanilla games we are used to, but perhaps a bit too much so, as marines we didn't really care that we lost a lot of RT's while pushing since a single recapper could recap half of the map in such a short time, also we felt the aliens were at quite a disadvantage since it was hard to setup effective ambushes due to the short building time.
I think the change is very interesting, the game went from a "slow" tactical positioning feel to a much more action feel, RT's felt less important so the focus was a lot more on attacking specific areas of the map and holding map control then chasing after recappers and defending RTs.
I think it needs a bit of tweaking, perhaps a 50% drop is too much, but some drop is good, maybe fiddle around with different numbers to see how they feel, longer building times but less HP etc...
All in all I think the change is good, but if left as it is now it changes the gameplay too much towards just being "another action shooter" rather then a highly tactical positioning game, although its hard to judge from only one game, tactics come with time and trial so it might create a interesting new meta game where RT's play a lesser "attack / defend" role
JektJoin Date: 2012-02-05Member: 143714Members, Squad Five Blue, Reinforced - Shadow
People should start streaming any organised games on the BT mod and posting them imo. I'm very interested seeing how certain controversial changes play out. How the cheap and low hp extractors work, tres egg drops, 1 cc jetpacks, drifter building, cheap gorges etc. etc.
RT's felt less important so the focus was a lot more on attacking specific areas of the map and holding map control then chasing after recappers and defending RTs.
I haven't played the BT mod yet, but what you've described sounds like a good change to me.
I get tired of constantly hearing "Skylights!" and "Ventilation!" and "Observation!" and having people running off to go rescue or rebuild them. I try to tell my teammates that the way to save Skylights is to take and hold Overlook. If the current BT makes running after RTs less valuable and encourages tactical play to secure territory, then I'm all for it.
The problem I have with the lower RT health is that it almost guarantees that the RT will go down before someone can reach it. This by extension means there will be less (meaningful) engagements as people run around each other in circles taking out RTs. Sure, biting or building an RT on your own takes a long time in vanilla and some may call it boring. For me those are some of the most intense moments in the game.
Well presumably higher mobility means faster response times. Skulks obviously go real fast - marines have permasprint now, not sure how much that helps.
AsranielJoin Date: 2002-06-03Member: 724Members, Playtest Lead, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow, Subnautica Playtester, Retired Community Developer
I just played and i really liked it. except for the lower res tower health. it was very frustrating to defend them. But i love running around as skulks, its so much fun to try to improve your movement good work
Yea, I must say, I've been a skulk at (empty) terminal, two marines came by and I did something I couldn't since 240 due to decreased speed - running up the wall, leaping long distance to bite a marine, leaping on the wall, dodging fire for some time to regain health and then repeating it until both marines are dead. It's the speed that allows to do those things quickly, disorient enemies and make me feel ultra-awesome. This feeling is now gone in vanilla, you will probably die as soon as you start climbing a wall.
Also, I have a complaint. Sewlek's aim is too good, it pushes balance towards marine side.
Comments
I have a four year old i7 920 and i can't see more than a ~3 fps difference in a giant room like cargo.
Now rich infestation on the other hand is different, but
A) comp players and therefore those that typically post in here don't use it
C) removing cysts as just suggested this page would still leave said rich infestation in use
*My line of thinking is : there is way bigger culprits out there for fps loss that add up to much more than 10 fps at the end of rounds.. So why would you target such a small contributor if performance was truly the focus of your argument for removal?
Disclaimer : honestly not trying to inflame anyone, just genuinely curious if cysts really are that much of a fps loss for some of you..
And so you know that i am not lying about my results, here's the screenshots with fps in the top left:
http://i.imgur.com/wZb2v1J.jpg
http://i.imgur.com/grhydTn.jpg
http://i.imgur.com/YWsE85B.jpg
They improved them, but I don't think they "fixed" them. I don't think they CAN fix them, tbh.
Without getting into the debate here, I would advocate the removal of those features irrespective of their affect on performance so you're absolutely right in identifying the not-so-hidden case I'm making. I just think that even +3-5fps (and I suspect it's probably more on long rounds) makes a difference at this stage!
Make Biomass give 8% of base HP per amount instead of the 2% or the like. Two full biomass hives would give as much HP as Carapace and Biomass currently gives. Remove Carapace and replace it with Reactive Armor. Any time you would take lethal damage or over 250 damage in .2 seconds, you gain 15/30/45% damage reduction instead for like .2 seconds. Would counter Shotguns/Railguns/Grenaderlaunchers and would be an interesting test.
It would be a way to remove Carapace and make it not the most important upgrade in the game.
you are delusional if you think most NS2 players overclock
But based on that statement above, I really wonder if something is going on like
or instead still a somewhat event-based approach
The first example would perform distance checks on every tick and try to find close stuff. This works fine for things like Crags that are supposed to interact with a lot of dynamic things. Armories might do the same, though native collision hash checks could work as well to notify about close entities instead of checking on their own.
But the second approach seems so much better for performance since it only executes code and checks for stuff after something changed in the gameplay that would make it necessary to do so.
You could extend this kind of approach to other things as well: instead of checking every tick whether or not a cyst is still connected to it's parent (and checking recursively whether or not that parent still has a connection as well) you would only check the entire cyst network whenever a new cyst or hive is planted or whenever a cyst or hive gets destroyed.
Or for power nodes: instead of letting every structure check on every tick whether or not it's in a location with power, have the power tell the structures in the associated room about the changed state. (If it's not already done that way)
Again, I have no clue about the implementation details in Spark and these assumptions are only based on something I recall a dev saying in the past.
You seem to gain 0.1 Biomass levels per alien structure that you build, in addition to the usual upgrades at the Hive for 25 and 40 res. I am not 100% sure though if this "fake" Biomass is usable for upgrades or only as HP extension, though. I think it's the latter.
Also, T Res cost for alien eggs has been increased:
Gorge - 10 res
Lerk - 40 res
Fade - 70 res
Onos - 100 res
This surely combats powerful early lifeform strategies. I didn't pay any attention to whether or not the Biomass requirement for Fade and Onos eggs was changed accordingly, but I assume that was the goal of this change in the end. Still, that is quite a steep investment for pretty fragile lifeforms. May I assume that T Res drops from the marine commander are also increased in price, so we don't get the inflation of 30 T Res Exos?
If Marines get Exos on 30 res and Aliens get nothing comparable till 70 res, I am so endlessly confused.
Having multiple ways to accumulate biomass is adding too much complexity. The system was already borderline too hard to understand. If anything I would rather that building things was the only way to accumulate biomass.
edit: Seems to only be the case for the onos?
Marine bullets being bigger.
-it rewards random hits at long range, because of the loose spread on the rifle
-instead tighten the rifle spread, and reward true aim
Fade new movement.
-heavily reduced stamina
-slower mid/late game
-less spectator friendly because of restricted movement
*I don't think there is anything wrong with vanilla fade movement.
Skulk new movement.
-more like pre b240
-seems good, needs more playtesting, vs accuracy/bigger bullets
Haven't tried lerk, but I don't think anything is wrong with vanilla lerk, other than the amount of blinding tracers coming at you.
Fade and Lerk movement is amazing right now and should stay exactly as it is. It's highly responsive and has tons of options. Now while the pseudo bhop thing is very enjoyable for the Fade, Skulk movement is still a little dodgy. The tolerance of how quickly you lose your speed when bumping into things or touching walls (unless it's head-on) needs to be way higher, as there's no way to quickly gain momentum again, at least not until Leap is up.
As for pretty much all other changes, I say bring 'em on.
New biomass system frees up too much tres for alien the tech trees. Fairly easy to put crags/shifts/shades all over without giving up tech.
This, I'd really love for it to be 18 slots or less, I prefer 16 slots
Sounds awesome, I always had a skulk hit me and stop me in my tracks >_>
We played a match yesterday with the new RT HP / build time changes, I have mixed feelings about this but wanted to post some input.
The game felt very good to play, it was a lot more fast paced than the current vanilla games we are used to, but perhaps a bit too much so, as marines we didn't really care that we lost a lot of RT's while pushing since a single recapper could recap half of the map in such a short time, also we felt the aliens were at quite a disadvantage since it was hard to setup effective ambushes due to the short building time.
I think the change is very interesting, the game went from a "slow" tactical positioning feel to a much more action feel, RT's felt less important so the focus was a lot more on attacking specific areas of the map and holding map control then chasing after recappers and defending RTs.
I think it needs a bit of tweaking, perhaps a 50% drop is too much, but some drop is good, maybe fiddle around with different numbers to see how they feel, longer building times but less HP etc...
All in all I think the change is good, but if left as it is now it changes the gameplay too much towards just being "another action shooter" rather then a highly tactical positioning game, although its hard to judge from only one game, tactics come with time and trial so it might create a interesting new meta game where RT's play a lesser "attack / defend" role
I haven't played the BT mod yet, but what you've described sounds like a good change to me.
I get tired of constantly hearing "Skylights!" and "Ventilation!" and "Observation!" and having people running off to go rescue or rebuild them. I try to tell my teammates that the way to save Skylights is to take and hold Overlook. If the current BT makes running after RTs less valuable and encourages tactical play to secure territory, then I'm all for it.
Also, I have a complaint. Sewlek's aim is too good, it pushes balance towards marine side.