Bitcoin Forum
July 25, 2021, 05:06:21 PM *
News: Latest Bitcoin Core release: 0.21.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 2 3 [4]  All
  Print  
Author Topic: XEL - The Codable Decentralized Supercomputer - XEL.ORG  (Read 1996 times)
syclone
Hero Member
*****
Offline Offline

Activity: 752
Merit: 500



View Profile
March 07, 2021, 12:38:15 AM
 #61

EK WE NEED YOU!   Please come back and revive this beauty.  I'll buy you a tesla.
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
1627232781
Hero Member
*
Offline Offline

Posts: 1627232781

View Profile Personal Message (Offline)

Ignore
1627232781
Reply with quote  #2

1627232781
Report to moderator
drays
Legendary
*
Offline Offline

Activity: 2282
Merit: 1069


View Profile
March 08, 2021, 10:38:57 PM
Last edit: April 11, 2021, 11:16:06 PM by drays
Merited by josegines (1)
 #62

Of course, it would be great if EK suddenly returns, and syclone buys him a tesla or two...

Apart of this... Some unrelated thoughts below.

Finished reading some of the ePL docs. They are pretty good btw.

In general, it looks like not many people (if any) have read those docs, as though the doc (for example https://docs.xel.org/learn-epl.html) is quite good and is written pretty well, it clearly lacks proofreading. It contains big number of typos, some mistakes, and some (though small) parts are not understandable, probably they are edited in a hurry. All this shows the real interest in the project was never high among followers... There were people who spent their time writing a good doc, but no people to spend some time to read it and provide feedback :/

Anyway, after spending some more time thinking on the topic, I'll try to address some of the issues Ismail123456 raised.

In my opinion there is basically no chance that this project will be revived.

From the point of view of computer science (CS), there is no use for this approach. In theoretical CS we learn how to classify problems in terms of their computational complexity (P, NP, EXP, NEXP,...). We think that some problems in NP are intractable (but we don't know for sure).
Tractable problems (the ones in P) can be solved in reasonable time, while intractable problems cannot be solved in reasonable time in general, no matter how much computational power you throw at those problems (not considering quantum computing). In general there is no hope in trying to solve an intractable problem by guessing, which is what this project is aiming at.
In short: From the point of view of computer science, this is a hopeless approach.

(1) First of all, the expression "In general there is no hope in trying to solve an intractable problem by guessing" is not correct. There is definitely a hope to solve intractable problem by guessing, but there is no guarantee it will be solved in polynomial (reasonable) time. So the conclusion "From the point of view of computer science, this is a hopeless approach." is not correct as well. It is not hopeless approach at all, it is just an approach which doesn't provide guarantee of solution.

The practice shows that though even if in general case the solution cannot be guaranteed, in many (and very often - in most of) practical cases solution can be quite easily found either by so called 'guessing', by brute force, or exhaustive search, or by other selective search methods.

One example from practice Smiley Long time ago when still learning in university, I was offered a task to develop a program to automate the timetable generation for the several university departments. Usually the process of compiling the schedule takes a lot of time from department management, so the higher management decided automating it could save valuable resources of department staff. This is a typical JSP (Job shop scheduling problem), and of course I was told by our computer science experts, who were proficient on JSP theory (and in general were much better in theory than practice) that task to generate optimal timetable is intractable and in general case has no solution. Well, but who cares about general case? I was well aware that this task was successfully completed by university staff two times per year for many years, and though it was a bit time consuming, it was perfectly solvable. So long story short, I developed that program, and it worked very well. It was generating a timetable by 'directed' search, starting from the seemingly the most promising variants. To get a more or less acceptable solution it needed to work just under one minute, but the longer it worked, the better timetable it generated. Now I recall that this task is an ideal task for solving with XEL (with some needed enhancements in storage capacity, sure).

(2) Secondly, some study of the documentation show this project is not solely aiming at solution of a problem by guessing. Yes, the example in the doc finds prime numbers by guessing and testing, but this is not the only use. Except random guessing, this approach is good for brute-force search, and for some other search strategies too.

(3) Thirdly, though I didn't work in a scientific research for very long time, and so I do not know all the cases where the XEL's approach could be applicable, I see at least one case where XEL's capabilities fit very well: those are the problems solvable by a whole class of methods - Monte Carlo ones (https://en.wikipedia.org/wiki/Monte_Carlo_method). In my opinion (as a person, who personally used this method for problem solution in computational physics), it is a perfect match. You can perfectly parallelize such tasks, run them on computers of very different capacities, and bring the results together then. The more samples you generate and analyze, the more precise your solution will be.

In short, in my humble opinion - though intractable problems are not solvable for general case, in most practical cases solution can be found. And random guessing is by far not the only way this platform could be used.
I find this approach far from being hopeless, and platform far from being unusable. It just needs to be used correctly; for right set of tasks. And of course, it needs people willing to work on it, not purely speculate...

... this space is not for rent ...
josegines
Hero Member
*****
Offline Offline

Activity: 740
Merit: 528



View Profile
March 23, 2021, 09:13:32 AM
 #63

It is true that there has been no movement or development in recent months, but the network continues to be maintained thanks to 9-10 peers. I suppose they will have a cost, unless for some reason, it is free to maintain them. This gives me a modicum of hope. XEL left Bittrex 4 months ago and there they are still on without any apparent task.

PI NETWORK Descarga la APP para empezar a ganar PI! Join the beta Código invitación: karenabril
Ismail123456
Newbie
*
Offline Offline

Activity: 4
Merit: 0


View Profile
April 11, 2021, 02:36:02 AM
 #64

...

Ad 2) It can be used for search strategies (like brute force) that can be solved by random guessing.
Ad 3) It works for some special cases of Monte Carlo Methods, namely those that can be solved by simple random guessing  Wink

Explanation: You can't trust any computation node in the network. Therefore you must verify each (partial) result from every single computation node. Furthermore the XEL network needs to determine who of the workers gets rewards. If you can't verify the (partial) results the workers best strategy would be to cheat (because it is less work).

For that reason the only meaningful application that remains is solving problems by random guessing, for which you can easily verify if the result is correct.
Pages: « 1 2 3 [4]  All
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!