Interview with Engineers - Optimization/Performance

Guys, there's an "Optmization News Troll Thread" just below this one, let's keep this thread with proper questions.

Move the store to an external application that pops up outside of the game. My computer sounds like its on fire when in the store browser.

If it helps performance, have store and loadout screen use 2d sprites instead of rendering 3d models, and if we want to see something in 3d have a button that says render in 3d.

I think the store and loadout screen render a lot of resources that isn't necessary. These screens should be less resource intensive than a first person shooter...

My requests:

• Optimize the death sequence. I'd love for it to be a bit more optional, meaning I can immediately skip it to start planning my loadout. Alternatively there are times I wish I could go back and double-check how I died but after it's gone it's gone.
• Optimize terminals. The delay before you can access (or hack) a terminal is a constant source of annoyance. It's become a little more irritating now that "E" not only opens loadout, but closes it -- because I'll walk up and tapping "E" the first time won't work, and I'll tap it a few more times, and the second press will exit the terminal. This should be reliable.
• Render distance. It'd be great if things rendered more reliably at range.
  • While in a vehicle, I should reliably see all other vehicles.
  • If I'm shot by something, I should start rendering the thing(s) that shot me.
  • If I shoot at something with a lock-on missile, it'd be nice for that vehicle to render a bit further/longer than normal too.
  • Turrets should never totally disappear.
  • Use a brief fade-out effect for things which stop rendering, instead of having them just pop out of existence. (This would make the experience feel a lot higher quality, but probably also works against your goal of optimizing performance. :/ )
  • Always show friendly Sunderer repair/ammo icons on the minimap. They're really intermittent currently.
  • Always render the closest ammo pack on the minimap.
  • (optional but awesome) When low on ammo, nearby sources of ammo of the appropriate type(s) (infantry/vehicle) should flash repeatedly.
• (really not related to perf at all) Squad Waypoints should be symbols (triangle, circle, square, diamond.) Currently I use all empire-color IFF indicators, which only leaves about one color for squadmates (to avoid blending with empire colors) which means waypoints blend together and get pretty noisy.

The most brazenly obvious things i can think of is an ETA.

Followed by "By how much has the roadmap been delayed as a result?"

And "Did you have to do this before fixing projectile rendering on live servers?"

One that i have is "My machine is crap. I barely meets the listed minimum requirements and as a result im stuck with 15fps in the larger fights. Will i see any performance increase?"

I think everyone else has pretty much covered what I wanted to ask. Main points of interest:

• Multi-core support
• 64 bit support (RAM for the RAM god! Bits for the bit throne!)
• HUD streamlining
• Expanded DirectX support
• Normalizing (preferably longer) render distances, especially for infantry and walls/spawn shields. In particular making sure walls render before anything else does.
• Better fight scaling (ie less FPS loss per person)
• More stable FPS. Technically the FPS I get is fine, but it would be nice if it would stay that way more.

When someone lands a drop pod in the middle of a group of players, will they be visible sooner and will they be able to see other players sooner?

How about a slider option for icons on screen. I feel like my screen is full of icons all the time and I think the current option is just to turn them off or on is not enough (correct me if wrong). But I'd like a slider that fades out icons after w/e distance chosen. So icons are shown, as it is currently, up to 30m then start to fade out after that for example. I have the map for further out important icon ques anyway. Is this possible and would that help performance?

As others have stated, i will like information on multicore support.

I would also like information on render distance, either improving its distance, or at least rendering objects that were fired outside of render distance once they enter a players render distance.

Pic to go with my comment above ^^
Is it possible? Would it help?

deleted

I would like to know

1: Is the players render distance will be improove or affected by the optimisation process

2: Is the mana turrets , view distance will be improove so we wont be target by invisible mana turrets and rockets , or is this is to CPU heavy to be truely handle or solve

3: Is physX particule will be FIX by that i mean much less crash and better performance while using PhysX particules systhem ?

4: How is the code and art team think they will be able to do with lesser frame to still provide fluid animations , i think the animation are already quiet simples , if you compare to BF3 or BF4 , is it truely achivable or do we have to expect some few downgrade to get better performances ?

5: wich kind of systhem will benefit the most from the overall optimisation so far ? 2 core 4 core or 6 to 8 core processsor ?


TIPS : Better minimaps icon ( for vehicules terminal and weapons terminal generator etc.. so we could remoove many floating icon in 3d space

Agreed, I've been playing Indar for close to a year now I know each square inch of the map. I don't need floting icons to tell me where things are anymore.

Will the optimization help Physx run more stable? As it stands now the game crashes way too often with this enabled.

Also how will some of the cpu optimization effect i7 users?

Can we expect client and server side optimizations? --> will hitreg improve on big fights?
Will the optimization also improve the effects like heals,repair.resupply not working in big fights?

Which rigs can expect a bigger performance boost - low/mid/high end systems?
High end CPUs/GPUs seem to run @ 40-60% load during very large fights? Will we see an increase in CPU/GPU load and better fps?

I'm interested in hearing about the strategy for managing the optimisation project from a code point of view.
So do you have some sort of branching and merging strategy for dealing with the optimisation, or is it being worked on from the core code base?
If you have branched off to deal with it, how do you deal with the merges back to the main code base?
And how are you going to ensure that the optimisation work that you have completed persists throughout all the future updates?
Will the final optimisation build be used as a baseline, and will you then be doing performance testing of each update to make sure that nothing has regressed, and comparing it back to the baseline?
Do you have any automated tests that deal with any of this?
And how will you be testing this against the myriad PC configurations?

I don't believe that crashing on exit is that big of an issue. (crashing is, but not when you are closing the game) I like to think of it as a girl you're breaking up with, "You're not closing on me, I'm going to close on you!" *crashes*

How are these questions?
Back on topic, I agree with the hotkey concern. Is it possible that we can have a hotkey toggle option?

I find it rather annoying that i have to go into task manager every time to close PS2 manually, i even wrote a script so it wouldn't make a crash report and send it over to SOE and just kills the process.

For many the game runs better when graphics quality setting is set to "high" but many also dislike some of the effects that are forced on by the setting.

- Can we expect any new graphics settings which are commonly found in modern games such as bloom and depth of field effects?
- Would you consider letting us disable your built in anti-aliasing like we were able to do in the past since this tweak improved frame rate?
- In early beta we could turn the game fog off by writing "/fog off" in the chat which also increased your framerate, is there any chance of seeing this return in any form?

optimisation and performance is all well and good, but it needs to be PERMANENT.

Going by the previous track record, a simple GU ends up dropping peoples FPS by large amounts. What's to say you get it fixed for this update, then the next it's back to square one. Also, it's taken way too long for this to have started to be addressed. It should have been looked at way back at GU2 or something.

Also while it needs to take priority, it doesn't mean everything else needs to stop? Which it seems to have. After all, you don't have every team working on optimisation do you? I'd be a bit worried if the art team is dealing with it. So yes, you're heavily invested in getting the performance right, that's good and all, but honestly, it should be done anyway and not really hold everything else up.