FIrstly, the purpose of this post is just to go over what I've found about the FSS scan times, and what I think can be deduced from that, so that it informs further discussion and possible changes to improve the situation. It's not a bug as such so I haven't put it in a bug report, and it also comes up quite a bit so I thought the info and findings might be of interest to some others too.
A bit of background initially - I've only got onboard graphics (Intel HD 520). As a result of that, GPU intensive things become very obvious, and bottlenecks or apparent bottlenecks can be seen.
A couple of pertinent examples:
There's many more but those are the most relevant.
So as a result of that and a great many FSS scans, I've had a bit of a hypothesis about what's going on with the scan times. I've then sat down and timed a variety of different scans to actually get some data. (Please note, these are all for my setup - I would expect the times to be different for other people, but the general pattern to remain. It would be interesting if the pattern varies for others though.)
Here's the results:
The time for the scans on landables with vulcanism varies a lot, and it's not in proportion to the number of signals. See some examples below:
I was just concentrating on scan times at that point and didn't note down other details about the bodies in question as there's far too many to note conveniently. I do have screenshots though. I don't think that's really necessary for these purposes though.
So key points:
Supplemental info (for readers, not FD):
- to the best of my knowledge, the location and numbers of geological sites / signals is produced as part of the planet generation (as opposed to being something that's added according to criteria once the planet generation is complete)
Here's my conclusions as to what's happening.
Effectively this is all the stuff that has always gone on when approaching a planet (and approaching a geological site if there is one), but whereas mostly it would go unnoticed for people with half decent graphics cards because it's all happening in the background while in flight, when using the FSS it all has to be done when a body is zoomed on, and so the time taken suddenly stands out.
In terms of what could be changed to improved things, here's my take on it:
- for bodies without vulcanism, I don't have any suggestions for FD. The stuff that's happening is still going to need to happen. Hopefully though, some performance tweaks can be found which can make it work quicker. (For me personally, it's just another sign of needing to get myself a bit of a hardware upgrade!
)
- for bodies with vulcanism, this is where the biggest part of the issue is and which would also seem to be most apt to changes being made to improve things.
I've already suggested it in other threads, but the best solution would appear to me to be to drop the resolution of numbers of geological signals from the FSS and leave that to to be done as part of the DSS scans, and just report the presence of geological signals based on whether a body has vulcansim (which is info that's effectively available instantly).
One slight issue is that not all bodies with vulcanism have geological sites. However, it's a very low % and I'd personally be more than happy with 1 out of every few hundred bodies turning out not to have any sites, in order to save the time on the FSS scans.
For FD, I hope that's all useful info which helps towards improving this issue!
For anyone else reading, I hope that all makes sense, and please feel free to either corroborate or provide contrary data if your experience differs!
Fly safe, cmdrs. o7
A bit of background initially - I've only got onboard graphics (Intel HD 520). As a result of that, GPU intensive things become very obvious, and bottlenecks or apparent bottlenecks can be seen.
A couple of pertinent examples:
- There's always been an issue for me with generation of Geological POIs that isn't there for other types
- The time taken for planet generation when starting on a planet has always been very noticeable
There's many more but those are the most relevant.
So as a result of that and a great many FSS scans, I've had a bit of a hypothesis about what's going on with the scan times. I've then sat down and timed a variety of different scans to actually get some data. (Please note, these are all for my setup - I would expect the times to be different for other people, but the general pattern to remain. It would be interesting if the pattern varies for others though.)
Here's the results:
| Body type | Action | Time taken |
| Non-landable (all) | Resolve image of body and complete scan | Instantaneous (effectively) |
| Landable | Resolve image of body | ~4 - 4.5 seconds |
| Landable, no vulcanism, no signals | Complete scan | ~ 5 seconds |
| Landable, no vulcanism, signals (non-geological) | Complete scan | ~ 5 seconds |
| Landable, vulcanism | Complete scan | ~30 - 50 seconds |
The time for the scans on landables with vulcanism varies a lot, and it's not in proportion to the number of signals. See some examples below:
| Number of Geological Signals | Number of Biological Signals | Time taken for scan to complete |
| 8 | 6 | 29 seconds |
| 10 | 5 | 43 seconds |
| 23 | 0 | 51 seconds |
| 23 | 0 | 31 seconds |
| 36 | 0 | 41 seconds |
I was just concentrating on scan times at that point and didn't note down other details about the bodies in question as there's far too many to note conveniently. I do have screenshots though. I don't think that's really necessary for these purposes though.
So key points:
- there's 4 to 4.5 seconds for the planet to resolve to the level of image needed for the FSS zoom view.
- the time for scan to complete is always about 0.5 seconds on top of that for bodies without vulcanism.
- for bodies with vulcanism the time taken for the the scan to complete is much longer and is variable.
Supplemental info (for readers, not FD):
- to the best of my knowledge, the location and numbers of geological sites / signals is produced as part of the planet generation (as opposed to being something that's added according to criteria once the planet generation is complete)
Here's my conclusions as to what's happening.
- zooming on a landable invokes the planet generation system
- planet generation has to reach a certain point in order for the zoomed image of the planet to be produced
- a slightly further point then has to be reached for planets without vulcanism for any non-geological signals to be determined
- for planets with vulcanism, the planet generation has to reach a much further (and more detailed) stage for the presence and numbers of geological sites to be generated, and the scan can't complete until that stage has been determined
Effectively this is all the stuff that has always gone on when approaching a planet (and approaching a geological site if there is one), but whereas mostly it would go unnoticed for people with half decent graphics cards because it's all happening in the background while in flight, when using the FSS it all has to be done when a body is zoomed on, and so the time taken suddenly stands out.
In terms of what could be changed to improved things, here's my take on it:
- for bodies without vulcanism, I don't have any suggestions for FD. The stuff that's happening is still going to need to happen. Hopefully though, some performance tweaks can be found which can make it work quicker. (For me personally, it's just another sign of needing to get myself a bit of a hardware upgrade!
)- for bodies with vulcanism, this is where the biggest part of the issue is and which would also seem to be most apt to changes being made to improve things.
I've already suggested it in other threads, but the best solution would appear to me to be to drop the resolution of numbers of geological signals from the FSS and leave that to to be done as part of the DSS scans, and just report the presence of geological signals based on whether a body has vulcansim (which is info that's effectively available instantly).
One slight issue is that not all bodies with vulcanism have geological sites. However, it's a very low % and I'd personally be more than happy with 1 out of every few hundred bodies turning out not to have any sites, in order to save the time on the FSS scans.
For FD, I hope that's all useful info which helps towards improving this issue!
For anyone else reading, I hope that all makes sense, and please feel free to either corroborate or provide contrary data if your experience differs!
Fly safe, cmdrs. o7
