N-body physics

[Incarnation of Chaos]

6 hours ago, Nuke said:

i dont have a problem with n-body on select objects and for any player crafts (provided difficulty setting). it doesnt make any sense to try to move any of the full size planets, but if i wanted to turn gilly into a space ship i should be able to do that. 

The video literally shows why Gilly isn't being moved anytime soon; you either need a box of fuel of similar size and mass or you need to change the laws of physics themselves lel.

[Nuke]

1 hour ago, Incarnation of Chaos said:

The video literally shows why Gilly isn't being moved anytime soon; you either need a box of fuel of similar size and mass or you need to change the laws of physics themselves lel.

or install a thousand orion drives.

[Incarnation of Chaos]

2 minutes ago, Nuke said:

or install a thousand orion drives.

Still unlikely to have anywhere near the intended effect; i'd have to do the math to see for sure though.

[GoldForest]

1 hour ago, Nuke said:

or install a thousand orion drives.

F (newtons) = m (kilograms) x a (m/s^2)

 

To move Gilly:

F = 1.2420363×10¹⁷ kg x 9.81 m/s^2 (1G of acceleration for argument sake)
F = 124,203,630,000,000,000 kg x 9.81 m/s^2
F = 1.21843761E+18 N
or
F = 1,218,437,610,000,000,000 Newtons of force to move. 
1,218,437,610,000,000 Kilonewtons
1,218,437,610,000 Giganewtons
And that's where it stops because apparently there's nothing bigger than a giganewton, unless you want to use the other orders of magnitude. 

Now, let's take the F-1A with its 8,003.800 kN of thrust.

1,218,437,610,000,000 Kilonewtons / 8,003.800 kN = 152,232,390,864.34 or 152,232,390,865 F-1s continuously firing to move Gilly at a comfortable 1G of acceleration. And keep in mind, that's with REAL WORLD F-1s. Not the toned down KSP versions. 

You would need 902,546,377,778 (Rounded up) KE-1s to move Gilly at 1G of acceleration.

Moving planets is super hard. That's why it takes a planet crashing into one to dislodge it from orbit or something with greater gravitational force, say a black hole. 

You are NOT moving any planets under N-body physics or any other physics. 

Edited October 4, 2019 by GoldForest

[Nuke]

25 minutes ago, GoldForest said:

F (newtons) = m (kilograms) x a (m/s^2)

 

To move Gilly:

F = 1.2420363×10¹⁷ kg x 9.81 m/s^2 (1G of acceleration for argument sake)
F = 124,203,630,000,000,000 kg x 9.81 m/s^2
F = 1.21843761E+18 N
or
F = 1,218,437,610,000,000,000 Newtons of force to move. 
1,218,437,610,000,000 Kilonewtons
1,218,437,610,000 Giganewtons
And that's where it stops because apparently there's nothing bigger than a giganewton, unless you want to use the other orders of magnitude. 

Now, let's take the F-1A with its 8,003.800 kN of thrust.

1,218,437,610,000,000 Kilonewtons / 8,003.800 kN = 152,232,390,864.34 or 152,232,390,865 F-1s continuously firing to move Gilly at a comfortable 1G of acceleration. And keep in mind, that's with REAL WORLD F-1s. Not the toned down KSP versions. 

You would need 902,546,377,778 (Rounded up) KE-1s to move Gilly at 1G of acceleration.

Moving planets is super hard. That's why it takes a planet crashing into one to dislodge it from orbit or something with greater gravitational force, say a black hole. 

You are NOT moving any planets under N-body physics or any other physics. 

what about when tweakscale 2 comes out? 

muahahahaha. ive always wanted to use tsar bombas as rocket fuel. kerbals should get to type one. more bombs bigger bombs better bombs!

Edited October 4, 2019 by Nuke

[GoldForest]

1 minute ago, Nuke said:

what about when tweakscale 2 comes out? 

muahahahaha. ive always wanted to use tsar bombas as rocket fuel. kerbals should get to type one.

Real world Tsar Bomba: 209,199,999,999,997,630 Newtons
209,199,999,999,997.63 Kilonewtons

1,218,437,610,000,000 Kilonewtons / 209,199,999,999,997.63 kilonewtons = 5.824271 ~ 6 IRL Tsar Bombas to move Gilly. By how much? Eh, not sure. 

KSP version? Eh... not sure. KSP works off a 1/8th to 1/10 scale, correct? 

209,199,999,999,997.63 Kilonewtons / 8 = 26,149,999,999,999.60 Kilonewtons
209,199,999,999,997.63 Kilonewtons / 10 = 20,919,999,999,999.70 Kilonewtons

1,218,437,610,000,000 Kilonewtons / 26,149,999,999,999.60 Kilonewtons = 46.59 or 47 KSPified Tsar Bombas going off at once to move Gilly
1,218,437,610,000,000 Kilonewtons / 20,919,999,999,999.70 Kilonewtons = 58.24271558 or 59 KSPified Tsar Bombas going off at once to move Gilly

Disclaimer: I don't know if any of my math is wrong or right as this is really my first time calculating Force and what not. So, take all my numbers with a bag of salt. 

[Nuke]

17 minutes ago, GoldForest said:

Real world Tsar Bomba: 209,199,999,999,997,630 Newtons
209,199,999,999,997.63 Kilonewtons

1,218,437,610,000,000 Kilonewtons / 209,199,999,999,997.63 kilonewtons = 5.824271 ~ 6 IRL Tsar Bombas to move Gilly. By how much? Eh, not sure. 

KSP version? Eh... not sure. KSP works off a 1/8th to 1/10 scale, correct? 

209,199,999,999,997.63 Kilonewtons / 8 = 26,149,999,999,999.60 Kilonewtons
209,199,999,999,997.63 Kilonewtons / 10 = 20,919,999,999,999.70 Kilonewtons

1,218,437,610,000,000 Kilonewtons / 26,149,999,999,999.60 Kilonewtons = 46.59 or 47 KSPified Tsar Bombas going off at once to move Gilly
1,218,437,610,000,000 Kilonewtons / 20,919,999,999,999.70 Kilonewtons = 58.24271558 or 59 KSPified Tsar Bombas going off at once to move Gilly

Disclaimer: I don't know if any of my math is wrong or right as this is really my first time calculating Force and what not. So, take all my numbers with a bag of salt. 

they are big enough to break fp math anyway lol. 

[Incarnation of Chaos]

14 minutes ago, Nuke said:

they are big enough to break fp math anyway lol. 

You haven't seen a long long double have you?

(I'm not even joking these are a thing)

[Arugela]

Aren't quad doubles a common thing now? Or quads.

 

Qaud double would be an oct or just a double qaud... I would assume.

Edited October 4, 2019 by Arugela

[Incarnation of Chaos]

1 minute ago, Arugela said:

Aren't quad doubles a common thing now? Or quads. Qaud double would be an oct or just a double qaud... I would assume.

Hm? i haven't seen them in person but i wouldn't imagine they don't exist; how else would you implement 128,256,512 bit extentions to current CPU instructions. But for us consumers i doubt we're looking at anything more than a 128-bit FP number since the more extreme ones generally end up on server chips due to die size and cost.

[Nuke]

16 minutes ago, Incarnation of Chaos said:

You haven't seen a long long double have you?

(I'm not even joking these are a thing)

i dont think there is hardware support for that. honestly the last time i saw one of those was on a vax mainframe. the x87 could do an 80 bit data type. but you are going to lose performance due to the second register op even on a 64 bit system (on a 32 bit system it required 3). im not even sure if the x87 instruction set is still a thing anymore or if it was replaced by something else. theres avx but those are for vector types.

Edited October 4, 2019 by Nuke

[Incarnation of Chaos]

Just now, Nuke said:

i dont think there is hardware support for that. honestly the last time i saw one of those was on a vax mainframe. the x87 could do an 80 bit data type. but you are going to lose performance due to the second register op even on a 64 bit system.

They've been part of the C  standard since 89; iv'e also written code with them on multiple consumer machines ;o

But damn; i didn't expect you to have experience with old mainframe tech xD

[Nuke]

2 minutes ago, Incarnation of Chaos said:

They've been part of the C  standard since 89; iv'e also written code with them on multiple consumer machines ;o

But damn; i didn't expect you to have experience with old mainframe tech xD

the anchorage school district kept one around for historical reasons. my computer science class let us use it for programming if you asked nicely. they strung ethernet cable half way across the city so a few nerds could play with it. knowing the way the state of alaska does things its still running. 

Edited October 4, 2019 by Nuke

[Incarnation of Chaos]

Just now, Nuke said:

the anchorage school district kept one around for historical reasons. my computer science class let us use it for programming if you asked nicely.

That's awesome; did it have a dummy terminal or did you have to use punch cards?

[Nuke]

Just now, Incarnation of Chaos said:

That's awesome; did it have a dummy terminal or did you have to use punch cards?

there were a couple we could use, but we used a terminal application that ran on an old skool mac. 

[Incarnation of Chaos]

1 minute ago, Nuke said:

there were a couple we could use, but we used a terminal application that ran on an old skool mac. 

Oh thank god; i couldn't imagine having to "Program" with punch cards xD

[Arugela]

Can you imagine sending letters via telegraph!

 

You could even try smoke fires... That will get to the point faster! 8)

Edited October 4, 2019 by Arugela

[Nuke]

41 minutes ago, Arugela said:

Can you imagine sending letters via telegraph!

 

You could even try smoke fires... That will get to the point faster! 8)

this is why they used to charge per letter. it was kind of like twitter. 

 

wait what was this thread about? lol.

Edited October 4, 2019 by Nuke

[Incarnation of Chaos]

5 minutes ago, Nuke said:

this is why they used to charge per letter. it was kind of like twitter. 

 

wait what was this thread about? lol.

Spoiler

N-body physics apperently. Yeah; we better stop before each of us gets a banhammer .

 

[Arugela]

Is there any correlation between these older techs and the logic of how to implement n body physics. Or any take aways to help do it. You'd be surprised.

[Guest]

1 hour ago, Arugela said:

Is there any correlation between these older techs and the logic of how to implement n body physics. Or any take aways to help do it. You'd be surprised.

Difficulty of implementation isn't the problem, the effects on gameplay are. The devs expect some careers to span thousands of years, and with n-body physics entire planetary systems could become unstable, never mind orbital stations or such. There are other similar difficulties as well.

[Incarnation of Chaos]

1 hour ago, Brikoleur said:

Difficulty of implementation isn't the problem, the effects on gameplay are. The devs expect some careers to span thousands of years, and with n-body physics entire planetary systems could become unstable, never mind orbital stations or such. There are other similar difficulties as well.

Well take our solar system as an example; it's a series of resonances that's been keeping it stable for billions of years. So creating stable systems isn't impossible; just limits what you can create in a way. Orbital stations can be reboosted or otherwise corrected; so that's not an unsolvable issue either.

At the end of the day if they wanted N-body they would have it, but Star Theory decided not to. And i think it really comes down to two elements, primary gameplay like you're saying. N-body creates a ton of busy work with stations, supply and complicates orbits significantly; no longer can you just go "Hmmm; 80K is good" you have things like the distribution of mass making "Lumpy" gravity fields, preccession making it harder to keep a constant inclination and etc. This is more realistic, but does it really enhance the KSP experience? Not really; especially if most of these things are automated away. 

But i also suspect it comes down to time; depending on what we look at KSP2 has had about 2 years of development with another 6 months in the more generous estimates. So considering they're building the game on the newest Unity version from the ground up with a new codebase, optimization features KSP never had along with everything else it's understandable why they didn't go for N-body. They're familar with the Patched Conics system; they know how it's supposed to behave and that helps tremendously with troubleshooting. They're already tackling a big project with not enough time; N-body would just be more work for a feature that may end up turning away much of the audience or irritating the ones that stay.

This is a very smart choice; especially since they're going to support mods anyway. 

[Guest]

Yeah there are other gameplay issues as well. Like with the timespans. If you have a game spanning thousands of years, that means timewarping through centuries. You can easily compute the state of a patched conics system at any point in the past or future without having to compute the intermediate states, so you can time-jump rather than time-warp. With n-body (or even a simplified version where planetary bodies are on rails but craft are affected by their gravity fields) you can't do that, you would have to run through the whole simulation. Simulating, say, 100 years of a system with, say, 100 in-flight objects is gonna take a while even on a fast computer.

[OutOfAblator]

I'm going to throw in my two cents here, I have 200-ish hours on KSP and when i tried it, Principa seemed like a lot more busywork than what benefits it had.

[mcwaffles2003]

4 minutes ago, Brikoleur said:

Yeah there are other gameplay issues as well. Like with the timespans. If you have a game spanning thousands of years, that means timewarping through centuries. You can easily compute the state of a patched conics system at any point in the past or future without having to compute the intermediate states, so you can time-jump rather than time-warp. With n-body (or even a simplified version where planetary bodies are on rails but craft are affected by their gravity fields) you can't do that, you would have to run through the whole simulation. Simulating, say, 100 years of a system with, say, 100 in-flight objects is gonna take a while even on a fast computer.

Universe sandbox seems to do fairly well with it, also i dont see why there couldnt be a combination of on rails mechanics with n-body added for vehicles or even reduce it to a 4 most significant bodies approximation per vehicle so as not to waste computations on finding the gravitational effects of eeloo on a kerbin orbiting craft