Jump to content

Devnote Tuesday: Smooth Transitions


SQUAD

Recommended Posts

Not so hot on the reduction in garbage collection however; memory is something KSP is fast running low on. It's annoying but without clean up and now the reduction in scene changes - what does that do to overall memory use and release? Has any efficiency been sought to improve the release of memory? People seem to not want GC, but it's actually important to long term stability and the game's memory footprint.

I've been trying to learn some basic C# stuff for beginners. I'm slow, but interested and motivated. What Mu is talking about, I think, is a good thing. It will allow for more efficient use of memory by "grabbing the code by the ****", so to speak. In other words, instead of relying on Unity garbage collection to release previously used memory that is no longer needed, he is instead trying to close those "memory calls" himself by optimizing the code. Now that KSP is getting close to 1.0 and feature complete, he is taking the opportunity to condense and streamline. Just my guess. I think it's gonna make things way-better.

Link to comment
Share on other sites

Is there a way to disable that "feature"? Also, you guys should really unclamp the timewarp at low altitudes, it's terribly annoying to spin around Kerbin at 50x speed in a low parking/transition orbit...

Kerbin is not so bad. Landing on Gilly is what sucks the most. It's so bad, most of the time I just point nose down and burn, just so I don't have to wait so long.

I too would like to be able to warp at max speed through SOI transitions. Sometimes they are transitions you don't care about - ie gravity assist.

Link to comment
Share on other sites

55th! (Man this habit is so incredible stupid and useless)

Awesome Dev-notes! Thank you for all the in dpth information. Good to hear that everything is going well. I am really excited for 1.0. - honestly i cant wait till KSPTV this friday to see the new Kerbals. Keep it coming.

Excellent job SQUAD! :)

Link to comment
Share on other sites

Well written.

Liking what I read on the SOI fix, but not liking that this seemingly was a minor issue and quick fix. This in the sea of major issues, that imo, needs addressing first before any new features are added. But that is just me.

On my end, I’ve been mostly running around tying loose ends here and there, and fixing issues that came up during the integration of these recent features… Admittedly, not the most interesting thing to write about, but it did take up most of my time this week.

I like to contradict this. As these "not the most interesting things" could make or break 1.0. And having info on what is done under the hood, certainly helps elevate the concerns a lot of folks have about the technical aspect of KSP development.

What I would like to know is the following:

- The better performance

What exactly has been done in the memory management department and what effect can we see?

Has there been any significant improvements made on the GC or manual unloading of non used allocated data in memory?

Scene switching and unloading. Has this been partly fixed for 1.0 so it is not SO much of an issue?

- Degrading FPS on increased memory footprint

Has this been looked into?

- Orbital lag in regards to sea render

Has this been fixed in any way? As I find it rather odd the sea needs rendering at all when at X altitude.

So to SQUAD and Harvester, could you guys make a technical devenote, in which you inform us on all those non-interesting bug fixes and performance enhancements you made?

I for one would really like to read about this technical, more than on new features thb.

Link to comment
Share on other sites

I'm more concerned about the water physics.

Is that an issue?

I recently had a contract with 2 'surface' actions on the water, so I parachuted my jet down on to one of them then checked to see if I could get to the other by just going along the surface.

I was pleasantly surprised to find that I could easily skim along around 20m/s without a problem.

I was not surprised by the fact that I could not lift off form the water because that jet design can't even lift off form the runway until I hit the down-slope after the end.(I think I need more lift)

Of course having the bottom of my jet lined with Structural air in-takes may have had something to do with that.(They are not the same air-intakes featured in the videos I watched, but they are similar intakes and may have similar behavior)

Link to comment
Share on other sites

Unless they've been doing something silly with GC, the way that you reduce garbage collections is by reducing the amount of garbage you produce - so it's kind of implied that they've been improving on their memory usage (probably by reducing unnecessary allocations or introducing object pools or something similar).

Indeed--once you've reduced the garbage to collect, you can improve performance further by reducing the garbage collection itself.

Link to comment
Share on other sites

Alex (aLeXmOrA): After a really exotic weekend in which I got a wisdom tooth extraction surgery that gave me a hamster cheek (that I still have),

I just had mine out over spring break, its not fun. Hope you feel better soon.

Also, great devnotes, I'm really excited for this update.

Link to comment
Share on other sites

THanks for the updates, as usual. Hugely appreciated!

Great news about SOI change fix and improvements to memory management. Looking forward to trying out the new aero model. Any info on whether 1.0 will be overall lighter/heavier on the CPU than 0.90? My rig is already struggling; main reason I don't have FAR installed anymore.

Link to comment
Share on other sites

"In any case, however, there is also a new system in place which sets a time warp limit when approaching an SOI transition. This was something I added before this new fix, and even though it isn’t strictly necessary anymore, it is quite nice to have anyway, as it highlights the SOI transition event, and lets you know something important is happening."

This is great. Shows you and your team can turn a "bug" into an opportunity to add a feature. No doubt people still want to adjust after SOI transitions, or sometimes just before (especially if doing aerobreaking). So avoiding throwing the baby out with the bath water is a good idea. :)

Link to comment
Share on other sites

I just had mine out over spring break, its not fun. Hope you feel better soon.

Also, great devnotes, I'm really excited for this update.

It's good to get them out early. I never had mine out and both my lower ones .... well it was pretty bad for awhile and I still have them. After they decayed down to the gum line though, gum growth filled them in and now they're not so bad. There's still a few little bits of tooth sticking out but they don't bother me.

Link to comment
Share on other sites

Maybe it just stops existing timewarp when you're about to cross, but it doesn't prevent you from timewarping again (maybe after a second delay)
Maybe it could just be an option to work like it always has? I certainly don't want this game to control timewarping any more than it does, especially since I play using a realistic solar system and much higher timewarp rates. Even in the tiny stock system I'd prefer less timewarp clamps. That sort of thing is kind of annoying.
Link to comment
Share on other sites

Maybe it could just be an option to work like it always has? I certainly don't want this game to control timewarping any more than it does, especially since I play using a realistic solar system and much higher timewarp rates. Even in the tiny stock system I'd prefer less timewarp clamps. That sort of thing is kind of annoying.

Yeah, some of the limits already in game are somewhat problematic ... like when you're orbiting the smaller bodies, due to their smaller orbital velocities. I do understand why those limits are in game as it is ( if they should be needed at all is another discussion ) , but some of them look to have been in the fly stuff that would benefit from a redo.

Anyway, this was a very, very old issue, that is pretty much as old as the game. While it is good to see it fixed, it makes people wonder why did it took so long ... and now hopefully the derpiness of manouver nodes ( that somehow was exacerbated in recent patches ) could also get a fix, please :D ?

Edited by r_rolo1
Link to comment
Share on other sites

I do understand why those limits are in game as it is ( if they should be needed at all is another discussion ) , but some of them look to have been in the fly stuff that would benefit from a redo.
I agree. If anything the fix to SOI transitions should be a great reason to re-examine the time warp limits and why they exist since SOI transitions will now be handled properly. Hopefully that fix also applies to body impacts as well.

E: I mean, honestly, is there any sound reason why I can't blast timewarp in an 80km parking orbit to make my transfer window, or does it really require me to switch to the space center or some geostationary satellite? If there's a physics issue then that needs to be fixed, but clamps are just annoyances.

Edited by regex
Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...