Oolite Bulletins

Norby wrote:

Please test it for sure and feel free to make a proper fix.

Thanks, I'll test it a bit . This being one the most used OXPs around, it's somewhat essential that the default install after download is as flawless as possible. My plate is quite full at the moment, but that status will change over time. At that point, if no one else has upgraded the XMapping component to use SystemInfo.setInterstellarProperty, I can take a look at it.

Thanks Cody and Day.
If it's just one moment its most likely Oolite setting up the new system. The shader is expensive and using discard may be another factor (there's no other way though).

Other options would be to change expensive code parts (atan2 > atan or even sqrt, length > sqrt or even dot, smoothstep > max or min) and helping the compiler by using scalar ops and changing the evaluation order. Before I left a new shader was nearly done (~50% less instructions), only some minor details were missing. Probably just a few minutes work to get the settings right.

@Norby, spara: If you are at it shaderEffectsLevel in BGS-M.js needs also to be changed .-)

Svengali wrote:

shaderEffectsLevel in BGS-M.js needs also to be changed

A simple replace to the new detailLevel is not enough so I do not do this atm.

1. The current bug is working as a help also which allow to see the nice docking effect regardless of the player set down the detail to get more FPS in combat.

2. There is another problem (at least at me on Intel HD) which display a static centered texture instead of the old docking animation when bgsDockingFX is false, which currently is solved by the mentioned bug and I do not want to fix something which cause another bug what I can not solve easily.

Svengali wrote:

Thanks Cody and Day.
If it's just one moment its most likely Oolite setting up the new system. The shader is expensive and using discard may be another factor (there's no other way though).

Other options would be to change expensive code parts (atan2 > atan or even sqrt, length > sqrt or even dot, smoothstep > max or min) and helping the compiler by using scalar ops and changing the evaluation order. Before I left a new shader was nearly done (~50% less instructions), only some minor details were missing. Probably just a few minutes work to get the settings right.

@Norby, spara: If you are at it shaderEffectsLevel in BGS-M.js needs also to be changed .-)

I have the stutter when arriving at the new system, too. So whatever is being done on this end would be very welcome!

Svengali wrote:

Thanks Cody and Day.
If it's just one moment its most likely Oolite setting up the new system. The shader is expensive and using discard may be another factor (there's no other way though).

Other options would be to change expensive code parts (atan2 > atan or even sqrt, length > sqrt or even dot, smoothstep > max or min) and helping the compiler by using scalar ops and changing the evaluation order. Before I left a new shader was nearly done (~50% less instructions), only some minor details were missing. Probably just a few minutes work to get the settings right.

Yum, seems quite fun coding :-p

Zireael wrote:

I have the stutter when arriving at the new system, too.

Perhaps you're just shy around pirates you haven't met yet?

Is OO_REDUCED_COMPLEXITY still available in v1.80+? Or any other attibute to allow control about it?

Layne wrote:

Zireael wrote:

I have the stutter when arriving at the new system, too.

Perhaps you're just shy around pirates you haven't met yet?

Rolling on the floor laughing...

Statter at the first jump, but no stutter if I directly perform a second jump.

I've uploaded the new files (shader+texture) -> https://app.box.com/s/m3p4nutld442eb30sv6hkigf0thdgnek (29.2 KB).

Performance should be much better now.

They need to be dropped into the latest version of BGS, I presume? In my slightly older version, they produce strange results.

I'll install the latest .oxz version of BGS and check that. Done - same result. Have I done something wrong?

I'm seeing the same strange results as well with the next shaders/textures.

I wonder how much of this stutter is due to oxps working their own operations?

Cody wrote:

In my slightly older version, they produce strange results.

I guess my driver isn't picky enough about NaN and silently 'optimizes'

So, next try -> https://app.box.com/s/mavp90c7p1jfo8l4038zzmut0hqie13n

Alas, still the same problem, amigo!