[ANN][XEL] Elastic Project - The Decentralized Supercomputer - page 225.

coralreefer

sr. member

Activity: 464

Merit: 260

Quote from: Evil-Knievel on November 05, 2016, 04:28:17 PM

Quote from: coralreefer on November 04, 2016, 01:38:46 PM

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Hi EK, if you post your updates to a branch I'll test the retargetting this weekend when I'm working on the miner.

Sorry, I'm afraid I can push the changes not earlier than tomorrow evening. My internet connection was (and is) pretty limited as I am on the go right now. Taking the train tomorrow morning back home, and then I will sync everything. Until then, I am trying out several other ideas for the retargeting that I had ... I have written a small python "retargeting simulator" which will test the retargeting in various different scenarios. I will make that one public too, tomorrow!

Hope the Donna64 extension that I submitted to your miner turned out to be working fine Wink

Take your time on the retargetting updates. In my opinion this is still the number one open issue we need to get nailed down right now so no need to rush. The Donna64 extension worked great...thx for the tip! Also, I redid my crypto logic so now I have a single build for the miner and crypto lib. Unless you have any suggestions on an alternate approach for the alogs, I'll move forward incorporating them into this design. Either way, it won't be hard to change it later.

Evil-Knievel

legendary

Activity: 1260

Merit: 1168

Quote from: inlovewith on November 04, 2016, 08:35:52 PM

Can you play Crysis on this Super Computer?

If you mange to show, that the problem of "Playing Crysis" can be somehow reduced to a knowingly NP-complete problem, such as e.g. the Hamiltonian cycle problem, I am pretty sure you could then (somehow) express it as a brute force search over a 384 bit search space. If you succeed with that, chances are in your favor that you'll be able to play Crysis on this super computer!

Evil-Knievel

legendary

Activity: 1260

Merit: 1168

Quote from: coralreefer on November 04, 2016, 01:38:46 PM

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Hi EK, if you post your updates to a branch I'll test the retargetting this weekend when I'm working on the miner.

Sorry, I'm afraid I can push the changes not earlier than tomorrow evening. My internet connection was (and is) pretty limited as I am on the go right now. Taking the train tomorrow morning back home, and then I will sync everything. Until then, I am trying out several other ideas for the retargeting that I had ... I have written a small python "retargeting simulator" which will test the retargeting in various different scenarios. I will make that one public too, tomorrow!

Hope the Donna64 extension that I submitted to your miner turned out to be working fine Wink

Evil-Knievel

legendary

Activity: 1260

Merit: 1168

Quote from: daylox on November 04, 2016, 09:43:47 PM

What testing methods do you use? Function testing, automated testing or performance testing?

I think there is no need for performance testing here, since we can assume that the NXT guys already did that so we can simply assume that the code is already peer-reviewed in terms of performance bottlenecks. Our additions at most can increase the overhead by a constant factor which is limited by the maximum allowed WCET.

Function tests, yeah! There are basic unit tests that check for correct block organisation and stuff, but if you ask me ... in my eyes the whole test driven development is just an illusion. Your tests only test what you can think of, they are no super intelligent AI logic. So if you cannot imagine (or see) all the different kinds of ways a system might be attacked, than your unit tests will test for them either.

What I am trying to say is: you can write 10000 unit tests which will all succeed but your software might still be riddled with holes like a Swiss cheese.

IMHO the only correct way to "test" is excessively trying to break the system. This is nothing a function or unit test can ever achieve Wink

lda1000

sr. member

Activity: 581

Merit: 253

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Thanks for the summary - it's really handy for people that don't follow the forum every day.
Can't wait to run the released software.

Great work.

daylox

sr. member

Activity: 307

Merit: 250

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

What testing methods do you use? Function testing, automated testing or performance testing?

inlovewith

sr. member

Activity: 257

Merit: 252

Can you play Crysis on this Super Computer?

cryptodv

hero member

Activity: 1092

Merit: 507

btcstakes.com

Hello Lannister, I've sent you several messages via bitmessage. Is there any other way of contacting you?

m4nki

hero member

Activity: 1039

Merit: 510

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Yes, thanks for that summary!

chinadaye

member

Activity: 85

Merit: 10

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Great work

Whether to consider the development of smart phone client, the world's billions of smart phones, these phones together can provide a very powerful computing power, and most of the time in the night, these computing power is idle, if intelligent Mobile mining on the run, it will bring a huge number of users.

ImI

legendary

Activity: 1946

Merit: 1019

Quote from: ttookk on November 04, 2016, 06:25:33 AM

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

(…)

The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

I think this calls for bug bounties, doesn't it. Something like

"You break the client and it is not due to complete inaptitude: 0.3 BTC",
"You break the network: 0.7 BTC",
"You break it, show how and others can reproduce it: 1 BTC",
"You abuse the retargetting algorithm or other parts of Elastic network: 1 BTC",
"You come up with another solution: 2 BTC",
"Your solution turns out to be the best: 5 BTC".

The numbers are arbitrary, as are the exact conditions regarding how to win a bounty, but maybe it's a start.

sounds good to me

coralreefer

sr. member

Activity: 464

Merit: 260

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Hi EK, if you post your updates to a branch I'll test the retargetting this weekend when I'm working on the miner.

ttookk

hero member

Activity: 994

Merit: 513

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

(…)

The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

I think this calls for bug bounties, doesn't it. Something like

"You break the client and it is not due to complete inaptitude: 0.3 BTC",
"You break the network: 0.7 BTC",
"You break it, show how and others can reproduce it: 1 BTC",
"You abuse the retargetting algorithm or other parts of Elastic network: 1 BTC",
"You come up with another solution: 2 BTC",
"Your solution turns out to be the best: 5 BTC".

The numbers are arbitrary, as are the exact conditions regarding how to win a bounty, but maybe it's a start.

HmmMAA

hero member

Activity: 1111

Merit: 588

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Thanks for the update EK , keep up the good work !!!

Limx Dev

copper member

Activity: 2324

Merit: 1348

Quote from: Evil-Knievel on November 04, 2016, 04:29:42 AM

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

Great work....Here is a pic from my DK3 retargeting algorithm.

wizzardTim

legendary

Activity: 1708

Merit: 1000

Reality is stranger than fiction

Thank you for the summary!

You're doing a great work. Keep it up!

Evil-Knievel

legendary

Activity: 1260

Merit: 1168

Quote from: beyinsi on November 03, 2016, 10:26:14 PM

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

Hi beyinsi,

my personal point of view is that perhaps 90% are done. The remaining 10% are mostly details (e.g., the retargeting algorithm) which we all discussed on the last pages and still need to come to a consensus.

Website: Check
Blockexplorer: Check
Fast and efficient C based miner: Check
Wallet/Backend functionality: Check
Elastic PL Development Editor: Check
Elastic PL Parser (the core language): Check
Wallet Frontend: Missing parts (Claim from Genesis or Miner overview for example)
The retargeting algorithm: Done, but not sure if flawed. Discussion is still open here

And of course testing testing testing ;-)

beyinsi

hero member

Activity: 742

Merit: 501

too much tecnical discussions...

What percentages ~ of the XEL project completed? tell us for normal xel followers,

ek, pls make a summary for us

ttookk

hero member

Activity: 994

Merit: 513

Do we have to have this discussion here? I think there are more pressing questions regarding XEL, than whether or not it is considered non-profit or not.

jeffthebaker

legendary

Activity: 1526

Merit: 1034

Quote from: Riseman on November 02, 2016, 08:32:56 PM

Quote from: ImI on November 02, 2016, 05:29:26 PM

Quote from: .m. on November 02, 2016, 10:39:28 AM

That is interesting experiment, but most of the people are money hungry and it may be difficult for non-profit community driven project to attract attention even if it is deserved.

"non-profit"?

you are aware of that basically ALL cryptos are non-profit, apart from crap like ripple and onecoin and stuff like that?

or who "owns" BTC in your eyes? its CEO? Grin

there simply are no "for-profit" coins as nobody is interested in that stuff.

Well, scams are always for-profit, which among altcoins there're plenty of, if not the most part. But decent projects with an interesting tech always gain traction unless the developers make some bad mistakes.

Any coin with a premine or ICO is for-profit, which is essentially every coin out there. Most are for-profit nowadays. In the same vein, I'd argue most altcoins are scams, too.

Topic: [ANN][XEL] Elastic Project - The Decentralized Supercomputer - page 225. (Read 450523 times)