> On 1 Nov 2019, at 12:02, Philip Thrift <[email protected]> wrote: > > > > On Friday, November 1, 2019 at 5:51:26 AM UTC-5, Bruno Marchal wrote: > >> On 31 Oct 2019, at 12:00, Philip Thrift <[email protected] <javascript:>> >> wrote: >> >> >> >> On Thursday, October 31, 2019 at 5:47:14 AM UTC-5, John Clark wrote: >> On Wed, Oct 30, 2019 at 6:54 PM Philip Thrift <[email protected] <>> wrote: >> >> >> The computation is the same independently of the substrate of its >> >> implementation. For example, you could run the same program on a computer >> >> based on vacuum tubes or transistors, with the same output. >> Stathis Papaioannou >> >> > That's the case for the conventional-Platonistic definition of computing. >> > Not the case for computing with a material-intrinsic semantics. >> >> So according to "material-intrinsic semantics" the 4 that a vacuum tube >> computer produces when it adds 2+2 is not the same 4 that a transistor >> computer produces when it adds 2+2; and the 4 a white man gets when he adds >> 2+2 does not mean the same thing as the 4 a black man gets when he adds 2+2, >> and there is a male 4 when a man makes the addition and a female 4 when a >> woman does. So how can a serious person consider anything as monumentally >> silly as a computational theory involving "material-intrinsic semantics"? >> >> John K Clark >> >> >> If there is a program in C vs. a program in Python (vs. Java, etc.) that >> produce the same I/O, which uses the least energy? > > It all depends on the algorithm, and not of the language (but still on the > way that language is implemented). > > The only thing which requires energy is in the erasure of the information > (Landauer). Yet, it has been shown (by Hao Wang) that we can get Turing > universality with elementary operations which never erase anything. This is > of course reflected in quantum computations, which have to reversible, and > never dissipate energy. > > Computations does not require energy, except for read and write, and > interaction with the users. > > For example, instead of using the combinators K (which erase information, as > Kxy = x, implies that the information in y has vanished), we can use the base > I, B, C, W: > > Ix = x (identity) > Bxyz = x(yz) (composer, applicator) > Cxyz = xzy (permuter) > Wxy = xyy, (duplicator) > > In that case we can avoid using energy. > > So, your question is that it depends on the number of erasing done in your > algorithm, and in the universal machine implementing your algorithm. > > The presence of the quantum axioms ([]p->p, + p-> [<>p, for p sigma_1) in the > self-referential “observable” modes suggests that this remains true in the > physics extracted from arithmetic (or from any universal machinery). > > Bruno > > > > > >> >> Or a man vs. a woman that adds. :) >> >> >> Material-intrinsic semantics are a UCNC conferences topic. Check them out. >> >> @philipthrift >> > > > I am including with the program (C, Python, etc. the whole system (compiler, > interpreter) that takes the program (implementing a common "algorithm") and > ultimately produces machine code for different machines. > > If there was a universal compiler that could take a program in any (of the > top 10 languages, say) and produce the lowest energy and fastest version > transformation of the program for the target machine, that would be quite a > compiler.
That cannot exist, but some approximation of this could make sense. By the Blum speed-up theorem (or a more general version due ti Blum & Al.) there is no fastest universal machine. We can diabolise to bring a more faster one, in theory. There are nice theories which explains why they cannot be used in practice, but still play a role if we are concerned with the truth about the machine, and the relation between machine and truth. Bruno > > @philipthrift > > -- > You received this message because you are subscribed to the Google Groups > "Everything List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected] > <mailto:[email protected]>. > To view this discussion on the web visit > https://groups.google.com/d/msgid/everything-list/5edc388e-6c40-42a7-bf22-ca37cf369ae5%40googlegroups.com > > <https://groups.google.com/d/msgid/everything-list/5edc388e-6c40-42a7-bf22-ca37cf369ae5%40googlegroups.com?utm_medium=email&utm_source=footer>. -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/F33ED9D3-7992-4C0D-8ABF-82EB6147A62B%40ulb.ac.be.
