Page 1 of 2

Physical simulation of the space engine

Posted: 17 Sep 2018 01:23
by xingqiu1
Can you tell me what physical simulations have been implemented in the space engine?

Physical simulation of the space engine

Posted: 17 Sep 2018 01:28
by xingqiu1

Do you think that the universe in the spac engine is as big as the universe in the Universe Sandbox 2?

[font=Roboto, sans-serif]Some people say that in the Universe Sandbox 2, the black hole can be swallowed up to hundreds of times to swallow the galaxy. It is also said that the Universe Sandbox 2 is as big as the space engine. It is 1:1. I don't believe it, so can you go back to the right one?[/font]

Physical simulation of the space engine

Posted: 17 Sep 2018 06:53
by Quarior
xingqiu1 wrote:

Do you think that the universe in the spac engine is as big as the universe in the Universe Sandbox 2?

[font=Roboto, sans-serif]Some people say that in the Universe Sandbox 2, the black hole can be swallowed up to hundreds of times to swallow the galaxy. It is also said that the Universe Sandbox 2 is as big as the space engine. It is 1:1. I don't believe it, so can you go back to the right one?[/font]

Is like Space Engine in size but with many objects, performance decrease.

Physical simulation of the space engine

Posted: 17 Sep 2018 07:35
by Mosfet
xingqiu1 wrote:
Source of the post Do you think that the universe in the spac engine is as big as the universe in the Universe Sandbox 2

I don't know how big it is in US2, but SpaceEngine universe is a cube with 10 billions parsec per side, or 1000 Gpc3.

Physical simulation of the space engine

Posted: 17 Sep 2018 20:55
by Watsisname
xingqiu1 wrote:
Source of the post Can you tell me what physical simulations have been implemented in the space engine?

I think Space Engine involves more emulation than simulation.  For example it never simulates the geological processes behind terrain building or erosion.  It instead emulates surface features by procedural algorithms.  It didn't form the cosmic web structure of the universe by running an n-body simulation, but rather generated their distribution to look like the real universe on large scales.

The difference between simulation and emulation is important, because a simulation can only handle so many interactions before it's just not computationally feasible on a home computer.  This is what makes Space Engine and Universe Sandbox so fundamentally different.  Universe Sandbox is a powerful simulator and can portray things like planetary impacts that Space Engine cannot, but by the same token, it cannot simulate anywhere near as many objects as what Space Engine contains.  

Space Engine is a 10x10x10 Gpc cube, and there are about 10 galaxies per cubic Megaparsec.  That comes out to roughly 1013 galaxies in SE.  If each one contains on average 100 million stars (I'm not sure the actual number), then that's 1021 star systems, and maybe about 1022 planets, not counting the rogue ones. :)

Physical simulation of the space engine

Posted: 18 Sep 2018 05:37
by vlad01
USB2 has infinite space as with many 3D game engines, it just indefinitely scales with zooming or movement.  But it can only support a few 1000s objects at most due to a hard limit set in it for computation reasons as it is computing interactions between all those objects in every combo, so interactions grow exponentially with object count in a simplified way of explaining it.

You will see what that is like if you run the internal benchmarks. like the one with 5000 moons. That on my 8 thread CPU runs at 2-3 FPM, yes frames per minute!  Forget cinebench R15 or ashes of the benchmark, run the ones in USB2 lol

Physical simulation of the space engine

Posted: 19 Sep 2018 15:18
by SpaceEngineer
Watsisname wrote:
Source of the post If each one contains on average 100 million stars

100 billions actually :)

Physical simulation of the space engine

Posted: 25 Sep 2018 02:26
by Salvo
vlad01 wrote:
That on my 8 thread CPU runs at 2-3 FPM, yes frames per minute!  Forget cinebench R15 or ashes of the benchmark, run the ones in USB2 lol

LOL
Also KSP is very good for that!

Physical simulation of the space engine

Posted: 25 Sep 2018 03:39
by vlad01
Salvo wrote:
vlad01 wrote:
That on my 8 thread CPU runs at 2-3 FPM, yes frames per minute!  Forget cinebench R15 or ashes of the benchmark, run the ones in USB2 lol

LOL
Also KSP is very good for that!

I recorded it to put on youtube as a self reference to when I upgrade my PC to something like a 2700X later on but the video would not be accepted by youtube as the whole 10 sec video only was 1 or 2 frames. It kept thinking the video had errors but was fine playing back on my PC. It was as it was so I concluded it was just youtube complaining, audio was perfect so I know it recorded fine.


Some of the experiments I set up with tweaks in the debug panel, like no decay, huge particle limits and what not I left for hours to run and came back later in the day to see the results as it took minutes for each physics step to be computed.

They have updated it now so the same decay settings and particles can't be set as I could before so i's much faster but less detailed and complex now, but still many of the high body count experiments break your PC badly.

Physical simulation of the space engine

Posted: 08 Mar 2019 17:34
by ethos
Is there something "off" with the scale of things? For example I'm trying to recreate the transit of Mercury as seen in this actual image.[img]https://i.ytimg.com/vi/Gibaxh9x7O0/maxresdefault.jpg[/img]

however is SE, if I move away from Mercury far enough to make it appear as small as that in the image, the Sun is only slightly larger than Mercury is. Am i doing something wrong?

Physical simulation of the space engine

Posted: 09 Mar 2019 00:17
by JackDole
ethos wrote:
Source of the post Am i doing something wrong?

Yes.
You do not have to change the distance, but the FOV!
(But also the distance a bit, so that Mercury is visible before the sun.
And I recommend putting the 'Night side lights' down a bit.)
scr00058.jpg

Physical simulation of the space engine

Posted: 09 Mar 2019 02:53
by N0B0DY
vlad01 wrote:
USB2 has infinite space as with many 3D game engines, it just indefinitely scales with zooming or movement.  But it can only support a few 1000s objects at most due to a hard limit set in it for computation reasons as it is computing interactions between all those objects in every combo, so interactions grow exponentially with object count in a simplified way of explaining it.

You will see what that is like if you run the internal benchmarks. like the one with 5000 moons. That on my 8 thread CPU runs at 2-3 FPM, yes frames per minute!  Forget cinebench R15 or ashes of the benchmark, run the ones in USB2 lol

I remember running the one that simulates the local group of galaxies (just a few couple of them). My 8-core / 16-thread CPU was put to its knees:
~ 2,3 FPS,100% usage, temps reaching mid 80s Co. The mouse pointer was barely moving... And I thought my system was powerful behh..
US2 is the perfect CPU benchmark stress utility!
SE on the other hand is, as Watsisname said, a realistic emulation & representation of the current state of the universe. And a very good one on that!

Physical simulation of the space engine

Posted: 09 Mar 2019 06:34
by ethos
[quote="JackDole"][quote="ethos"][post]27450[/post] Am i doing something wrong?[/quote]
Yes.
You do not have to change the distance, but the [b]FOV[/b]!
(But also the distance a bit, so that Mercury is visible before the sun.
And I recommend putting the '[b]Night side lights[/b]' down a bit.)
scr00058.jpg[/quote]

Cool, thats much better. Is there a FOV that shows actual scale? Meaning if I was far enough away from Mercury in real life to appear a certain size to me, the Sun would be accurately be shown in size and scale as well?

Physical simulation of the space engine

Posted: 09 Mar 2019 09:02
by Mr. Abner
ethos wrote:
Source of the post Is there a FOV that shows actual scale?

That depends on the size of your monitor and how far away your eyes are from it. Basic trigonometry.

Physical simulation of the space engine

Posted: 10 Mar 2019 13:27
by ethos
[quote="Mr. Abner"][quote="ethos"][post]27462[/post] Is there a FOV that shows actual scale?[/quote]
That depends on the size of your monitor and how far away your eyes are from it. Basic trigonometry.[/quote]

It's 13.5 in tall and I'm about 25 in away from it.