5
$\begingroup$

Note: I'm working on a far future sci-fi setting for a novel series which aims to blend elements of both "soft" and "hard" science fiction concepts. I'm aiming for the series to be as grounded and consistent as possible whilst still allowing some elements of speculative physics etc. Megastructures, including Dyson-scale, are quite common and the majority (>90%) of people in any given star system live in space.

One civilisation in the setting has its capital in a binary star system, with the primary being an F6V star which supports one habitable planet and associated habitat swarms. The star has a partial Dyson swarm, primarily used for power generation/beaming. The star's distant companion is a relatively hot and massive white dwarf star (roughly analogous to Sirius B for now until more details are figured out) with a surface temperature of roughly 25,000K.

The society in question is known for its high integration with an AI system housed in a Matrioshka Brain around the white dwarf, used for housing uploaded minds, defence and economic coordination, scientific research and universe simulation. The system also supports the civilisation's e-democracy by managing the government and tallying votes from the system's population of several hundred trillion.

It is my understanding that a large Matrioshka Brain offers greater computational capacity but suffers from signal propagation delay due to the light speed limit at which individual elements of the brain can communicate with each other. My question is that would the White Dwarf strike a good balance between capacity and computational speed due to the more compact nature of any Swarm/Brain, or would it be more feasible for this structure to exist around a small red dwarf star instead?

Thanks!

Improve this question
$\endgroup$

4 Answers 4

Reset to default
7
$\begingroup$

I believe it does not matter in your case.

If a civilization can build a matrioshka brain, any element in it may possibly have more computational power by itself than everything humanity ever built so far combined. That should be enough for... well, anything we can think of, really.

Being closer together in this scenario means that the swarm layers will save a few seconds to minutes per message when exchanging data. So when deciding whether the balance is good, consider if that makes a difference to your world's inhabitants (and plot). When you are computing the meaning of life, the universe and everything, this time saving hardly makes a difference.

Also if a civilization has the capacity to build such a swarm, I think it wouldn't skip a star just because it is too compact, or not compact enough. Other factors such as its position in trade routes and civilization borders should be more important.

Improve this answer
$\endgroup$
2
$\begingroup$

Star Size doesn't matter for the delay

If the computational systems of your story are vaguely similar to our physics, their computational power will likely be proportional to the volume and power supply of a "computing block". Let's say all our computing blocks are the same size and the ones closer to the star operate a higher power, while the outer layers operate at lower power.

So if I have an AI which needs 50 Zeta-Super-Flops to operate, this will translate to let's say 50 Million Compute Blocks. To create minimal delays between these blocks working together we should pack them as close together as possible. Optimal would be a solid sphere, but in your story we're using a Matrioshka architecture - so power supply through heat from a star must be the most important matter for building our computer.

So our compute blocks need to be arranged on one or more layers of a spherical shell. The curvature of this shell will not make a huge difference, since the signals need to propagate on the shell between the compute blocks (assuming we cannot send signals as a shortcut right through the star) And the shells will also most likely have a pre-defined distance between each other (so the waste heat can be propagated to the next shell)

Concerning propagation distance - a bigger star will only mean the same amount of compute blocks will occupy a smaller section of the overall shell, the curvature of the section will be less. But it won't result in a relevant difference in propagation delay. A bigger star will simply mean you have more compute blocks and can run additional computation on the other compute blocks.

Surface heat may make a difference

Depending on your technology the difference in surface heat may be a lot more relevant, if you can leverage all the heat and your computational power is proportional to the input energy. So the deciding factor for choosing a star would be it's overall total energy output (surface heat x surface area). And maybe the life expectancy of the star might be relevant on the scales we talk about.

Distance to civilization

Since the internal delays of the mega computer are not that different for different star sizes I thin the most important factor would be proximity. Because astronomical scales are huge. Having a supercomputer in your home star system, vs. having one a lightyear away makes a huge difference. Even if the computer would be slower, having to wait two years just to send your question and receive your answer will likely dwarf the computation times.

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ Thank you, this is a great answer. In terms of surface heat, if we were to use an even hotter white dwarf, say 50-75,000 k, would this ultra-compact, ultra-hot Brain compete favourably with a less compact/dense computational swarm around a larger star? As for distance, I'm currently working under an assumption of 100 AU between the white dwarf and the primary, a light-lag of a little under 14 hours. I suppose if the capital has to wait 14 hours for the brain to receive a command, and 14 hours for the brain's reply to reach them, a tiny signal delay within the brain makes little difference. $\endgroup$
    – Nemactual
    Commented Jul 3, 2023 at 23:46
  • 1
    $\begingroup$ I think they would use a combination of both. A swarm around the primary for instant access and "live" computations, where the answer is needed faster than 14 hours. And gigantic computations which would need many hours on the swarm are off-loaded to the Matryoshka brain around the white dwarf, which is e.g. 1.000 times more powerful. - If a computation would take several days or even longer on the swarm, you would rather send it to the Matryoshka brain, where it will compute in minutes. - Also capacity of the swarm would be in high demand, so less urgent computations would also be offloaded. $\endgroup$
    – Falco
    Commented Jul 4, 2023 at 9:13
2
$\begingroup$

The Good, the Bad, and the Ugly

The good, as others have mentioned, is the delay isn't really an issue — but I've not liked the reasons provided.

You can't think of a Matrioshka Brain like a single CPU. It's going to have breathtakingly massive parallelization. There will be CPUs, memory, etc. all over the place. In other words, the problem won't be a whole lot different than workstation designers deal with today when trying to coordinate multiple processing cards with a much (much) slower backplane. In reality, good programming rids oneself of 99.9% of those kinds of delay problems. Any program large enough or complex enough to encounter backplane (or in your case, light) delays would be expected to be slower. So, no problems. Life is good.

On the bad side is the limited solar power available from a white dwarf compared to most other star types. But is this a problem? Not really. You just have more panels, or the civilization lives with a bit less capability. Or they don't even notice because what they have is "normal" for their experience. Frankly, I wouldn't sweat this one at all unless you want a subplot about some mad scientist who needs a bigger computer and keeps arguing that the lack of solar power won't matter....

Gratefully, I can't think of anything ugly about this. The good thing about dealing with a fictional technology in a fictional setting is that there really isn't sufficient science to say that it can't be done. And since you haven't asked if it can be done (thank you!), you're good to go. Rock and roll my friend.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thank you, this is another great answer. I envision this particular brain as one of the oldest still in existence, with a possible subplot being how this society (which is so incredibly integrated/dependent on the brain) may be impacted if anything were to ever happen to the brain, ie damage or sabotage. You also raise a good point about capability, and how something like this would be normal for them. I suppose if they ever decided additional capability/capacity was needed, they could just construct something like additional J-brains in orbit of the dwarf to act as additional capacity. $\endgroup$
    – Nemactual
    Commented Jul 3, 2023 at 23:55
1
$\begingroup$

The universe itself seems to be stable enough to work around the speed of light limitations. Nowhere does any information get passed faster than the speed of light, and this includes celestial bodies where the lightdistance between its extremities are considerably different.

Your concerns about the distances needed to build around a white dwarf star immediately raise the question why the white dwarf star's own body does not suffer from these delays.

Why wouldn't the Matrioshka Brain be capable of handling this delay that the entire universe is built upon? If anything, it's proven to be a very stable universal rule (pun very much intended).

Secondly, I'm no physicist but I'm pretty sure that the laws of thermodynamics state that the computation itself must take a non-zero amount of time (and therefore delay). Comparatively, the speed of light delay is going to pale in comparison to these computational delays.

Thirdly, the acceptability of any delay is one of conventional acceptance, not of technical (im)possibility. We accept that the mail (i.e. physical envelopes) takes a day to get to you due to human and vehicle travel times. However, we do not accept the same standard of an email, because we understand that there isn't a physical travel process involved that inevitably delays the delivery time.

Any system that would be built would inherently be built in the understanding that the inevitable delays are inevitable.

Or, to use an analogy, no one is going to buy an IKEA flatpack, transport it home go through the laborious process of constructing it, and then be surprised that there's IKEA furniture in their house. That what they knew they were signing up for.

Improve this answer
$\endgroup$
4
  • 1
    $\begingroup$ "Comparatively, the speed of light delay is going to pale in comparison to these computational delays." - no, not really. The speed of light delay already started limiting us, while being far lower on the technological scale than the civilization presented in the question. Out fastest processors cannot become must faster or bigger because the signals from one side to the other side would take more time than a clock cycle. At 3 GHz light can only travel 10 cm in a single clock cycle, and the signals in a processor usually have to be routed around... $\endgroup$
    – vsz
    Commented Jul 3, 2023 at 6:29
  • 1
    $\begingroup$ "why the white dwarf star's own body does not suffer from these delays" but it does! Any information or influence - such as gravity, electrostatic charges between particles, heat energy, etc. - can only propagate through the star's body at or below the speed of light. (Many things change quite a bit slower - look at how long it takes for heat/light generated in the core to be radiated from a star.) It's just that stars tend to change very slowly, so it's not really a big deal. $\endgroup$
    – Cadence
    Commented Jul 3, 2023 at 8:00
  • $\begingroup$ @vsz: The subsequent conclusion is that anyone who builds a >10cm processing unit has therefore already needed to account for the travel time of information being longer than a clock cycle. I'm not saying the limit doesn't exist, I'm saying it's inherently a solved (or mitigated, or accepted) problem by the time you're talking about having constructed a machine of the size OP is thinking about. It's nonsensical to assume that the builders of such a machine did not have the intelligence/forethought to account for the most universal limit that everything abides by. $\endgroup$
    – Flater
    Commented Jul 3, 2023 at 22:41
  • $\begingroup$ @Cadence: The key word there being "suffer", not "can be observed". The goal of my question was to point out that white stars still exist and "work" even with the existence of these limits. $\endgroup$
    – Flater
    Commented Jul 3, 2023 at 22:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .