Large Bitcoin Collider (Collision Finders Pool)

[rico666]

Who's trembling in despair now, mortal?

 

Indeed. In my defence, the reason why this lapse happened is me programming my ass off for the past two weeks.
Dropping off the keyboard - barely awake, I just started my LBC client to have it some searching done while I was snoring away.
I didn't realize it was still pointed to my development server. Well...

So what am I programming anyway? This is what a "cold start" looks like now:

Code:

# ./LBC -c 2
Generator: gen-hrdcore-sse42-linux64
New generator found. (DL-size: 0.41MB)
New BLF data found. (DL-size: 166.22MB)
Benchmark info not found - benchmarking... done.
Benchmarked factor 3.581787
Ask for work... got blocks [31833874-31834225] (369 Mkeys)
ooooooo
etc.

You basically download ~40kB of code and it figures out the rest. In theory.
If there is a new generator available, it updates and benchmarks.
If there is new balances data available it updates. If there is only a diff, it downloads that
and patches to minimize data traffic.
If the server is unresponsive, it retries.
If you query the server, it answers:

Code:

$ LBC -q
Server answer to 'query' is:
{
"done" : 999204,
"lastsee" : 1475960421,
"ips" : {
"<some IP>" : 2028,
"<some other IP>" : 234
}
}
'done' means we have delivered   1047.741 valid Gkeys.

So I will - once again - need 2-3 diehards to test a "Jarvis" pre-release next week. You should have a Linux 64bit and a machine capable of AVX2 or SSE4.2



Rico

[1714105979]

1714105979

[1714105979]

1714105979

[1714105979]

1714105979

[1714105979]

1714105979

[Jude Austin]

Who's trembling in despair now, mortal?

 

Indeed. In my defence, the reason why this lapse happened is me programming my ass off for the past two weeks.
Dropping off the keyboard - barely awake, I just started my LBC client to have it do some searching done while I was snoring away.
I didn't realize it was still pointed to my development server. Well...

So what am I programming anyway? This is what a "cold start" looks like now:

Code:

# ./LBC -c 2
Generator: gen-hrdcore-sse42-linux64
New generator found. (DL-size: 0.41MB)
New BLF data found. (DL-size: 166.22MB)
Benchmark info not found - benchmarking... done.
Benchmarked factor 3.581787
Ask for work... got blocks [31833874-31834225] (369 Mkeys)
ooooooo
etc.

You basically download ~40kB of code and it figures out the rest. In theory.
If there is a new generator available, it updates and benchmarks.
If there is new balances data available it updates. If there is only a diff, it downloads that
and patches to minimize data traffic.
If the server is unresponsive, it retries.
If you query the server, it answers:

Code:

$ LBC -q
Server answer to 'query' is:
{
"done" : 999204,
"lastsee" : 1475960421,
"ips" : {
"<some IP>" : 2028,
"<some other IP>" : 234
}
}
'done' means we have delivered   1047.741 valid Gkeys.

So I will - once again - need 2-3 diehards to test a "Jarvis" pre-release next week. You should have a Linux 64bit and a machine capable of AVX2 or SSE4.2



Rico

Just when I think you couldn't get any dumber you do something like this....

AND TOTALLY REDEEM YOURSELF!!!

Haha, keep up the good work man!

[scooterking]

edit: if you download the HRD-client after I posted this, I have already repacked
http://62.146.128.45/download/LBChrd-0.837_l64.tbz2
so it contains the newest blf file (also md5sums is updated, so you can check).
In that case reloading the blf file is not necessary.

So what's up with this latest build ?  I was getting a benchmark of 2.5 with 2 cores, went up to 16 cores and my benchmark went DOWN to 2.49  ??


sky

[rico666]

So what's up with this latest build ?  I was getting a benchmark of 2.5 with 2 cores, went up to 16 cores and my benchmark went DOWN to 2.49  ??

Lower numbers are better?
Although 2.5 -> 2.49 is probably a measurement tolerance.
The benchmark is 1 mio keys per <time shown in seconds> I will add a translation to keys/s.
If you get a factor of 2.5 it means you're doing like 400000 keys/s per core.
2.49 -> 401606  

edit: Also please be aware, that hyperthreaded cores yield only about 30% of the performance of physical cores.
So let's say you have a 8-core CPU with hyperthreading enabled (16 logical cores) and benchmark shows 2.5

That means you can do about 3.2 mio keys/s with LBC -c 8, but only about 4.16 mio keys/s with LBC -c 16
Modern CPUs may perform a little better.

Personally, I leave the hyperthreaded cores alone, so while my CPU has 4 physical and 8 logical cores, I drive LBC with -c 4
The hyperthreaded cores are left for the OS and are more than enough (for me) to do my programming, browsing, even compiling.


Rico

[scooterking]

Got it, thanks..  I was mining last night and I woke up to see "Server requires minimum client version 0.841.".  Where is .841 ?

Thanks!

[rico666]

Got it, thanks..  I was mining last night and I woke up to see "Server requires minimum client version 0.841.".  Where is .841 ?

Glitch during the server update (new shinier stats) - unfortunately. Apologies.
For now 0.832 is still good enough a.k.a. minimum requirement.
I ditched the "factor" info in favor of the keys/s already:

Code:

# LBC -v
0.857
# rm benchmark.pst
# LBC -c 1 -t 1 -l 1
Best generator chosen: gen-hrdcore-avx2-linux64
Benchmark info not found - benchmarking... done.
Your maximum speed is 562959 keys/s per CPU core.
Ask for work... got blocks [33206546-33206593] (50 Mkeys)

I hope it's more intuitive this way.


Rico

[scooterking]

And.... how to get .857 ?

[rico666]

And.... how to get .857 ?

Try to steal my notebook.   
Or wait a couple of days and get 0.860 - or so.

Working on a new release as fast as I can.

Rico

[rico666]

Hi I stay gift on the secret page directory.io for check
I think it will be found this about one week  

You wrote this October 5th and although a week is almost over,
I assume you are not talking about
https://blockchain.info/address/1PVwqUXrD5phy6gWrqJUrhpsPiBkTnftGg
because this has been deposited in May 2014.

I swear I have nothing to do with this address - except my client found its private key tonight!

I must admit, I am still quite puzzled, because if things are as they seem (to me), this would be the 1st real (like in real real) PK found by the pool. Is this for real or am I still not fully awake?

Damned be the TODO - I thought we would have more time to sort that out.

Ok, how to proceed if I find something?

The most important thing is to handle your knowledge about the found private key as if it was yours. I.e. under no circumstances is that key to be announced anywhere publicly. You may have good intentions, but you never know who's listening. See if you can contact the person behind the address you have found the key to. TODO: Reference to a best practice approach of contacting the owner has to be elaborated/added.

I know - I wrote this text, but it is bad. So what will be the process? I looked up various European national regulations for finders-keepers.
There is a deadline of 6 months after which the finder gains the ownership if the original owner does not show up or make a claim.
If the value of the found is less than 10EUR (Germany), this deadline runs from the day found which would be today.
Still, the original owner can claim  it back for a period of 3 years. Here's what I think - so far - should be done:

• I transferred the funds to 16cFBcM27DGriyvz5i8SLmd8n5ai8hCTEE
• When the confirmations are through, I may or may not (should I? What are the pros/cons?) publish the PK
• I have of course no intention to keep these funds (like $5 dollar atm)
• If the rightful owner does show up within the 6-month period, I will even round up that sum to 0.01 BTC and give that back.
• If no one will show up within the 6 months, the funds will go as pool bounty to its members.
• If - after that - the owner should appear within 3 years, I will transfer these 0.007899 BTC back to him regardless

Phew. I'm glad this has happened with such a small sum. Honestly, I pulled this process out of my nose today morning, so I'd really really appreciate if the Bitcoin community could work out some "standard" to this. I've looked up the forums and this case seems so exotic, that most of the takes on it is really just nonsense blah blah.

So:

How can one make clear he is actually the rightful owner? Just stating "It's me" obviously won't cut it.
It may very well be, that he used an entirely different PK (of the 2⁹⁶) for this address (because this one is a little bit shitty), so I assume providing ANOTHER PK to this address would count as good proof - wouldn't it?


Rico

[becoin]

Hi I stay gift on the secret page directory.io for check
I think it will be found this about one week  

You wrote this October 5th and although a week is almost over,
I assume you are not talking about
https://blockchain.info/address/1PVwqUXrD5phy6gWrqJUrhpsPiBkTnftGg
because this has been deposited in May 2014.

I swear I have nothing to do with this address - except my client found its private key tonight!

I must admit, I am still quite puzzled, because if things are as they seem (to me), this would be the 1st real (like in real real) PK found by the pool. Is this for real or am I still not fully awake?

Damned be the TODO - I thought we would have more time to sort that out.

Ok, how to proceed if I find something?

The most important thing is to handle your knowledge about the found private key as if it was yours. I.e. under no circumstances is that key to be announced anywhere publicly. You may have good intentions, but you never know who's listening. See if you can contact the person behind the address you have found the key to. TODO: Reference to a best practice approach of contacting the owner has to be elaborated/added.

I know - I wrote this text, but it is bad. So what will be the process? I looked up various European national regulations for finders-keepers.
There is a deadline of 6 months after which the finder gains the ownership if the original owner does not show up or make a claim.
If the value of the found is less than 10EUR (Germany), this deadline runs from the day found which would be today.
Still, the original owner can claim  it back for a period of 3 years. Here's what I think - so far - should be done:

• I transferred the funds to 16cFBcM27DGriyvz5i8SLmd8n5ai8hCTEE
• When the confirmations are through, I may or may not (should I? What are the pros/cons?) publish the PK
• I have of course no intention to keep these funds (like $5 dollar atm)
• If the rightful owner does show up within the 6-month period, I will even round up that sum to 0.01 BTC and give that back.
• If no one will show up within the 6 months, the funds will go as pool bounty to its members.
• If - after that - the owner should appear within 3 years, I will transfer these 0.007899 BTC back to him regardless

Phew. I'm glad this has happened with such a small sum. Honestly, I pulled this process out of my nose today morning, so I'd really really appreciate if the Bitcoin community could work out some "standard" to this. I've looked up the forums and this case seems so exotic, that most of the takes on it is really just nonsense blah blah.

So:

How can one make clear he is actually the rightful owner? Just stating "It's me" obviously won't cut it.
It may very well be, that he used an entirely different PK (of the 2⁹⁶) for this address (because this one is a little bit shitty), so I assume providing ANOTHER PK to this address would count as good proof - wouldn't it?


Rico

Firstly, what you claim you've found is not a collision. Hence the name of your 'project' is misleading. And secondly, how will you prove the 'rightful' owner of the PK you've found is actually not... you?

[nrg1zer]

• If the rightful owner does show up within the 6-month period, I will even round up that sum to 0.01 BTC and give that back.

The rightfull owner could send you a signed message, BUT if i would be the owner, i would have no idea who to contact. I dont think alot of people are reading your thread. It could be any wallet... maybe a cold storage or something. You could wait years before someone show up. And even if the one realizes, his coins are gone, how could he know you took them.
I think you are at a point where you have to decide if you want to stop this, because you have proven it works, or if you want to continue, but this could be called theft.
Which way you will go?

[rico666]

The rightfull owner could send you a signed message, BUT if i would be the owner, i would have no idea who to contact. I dont think alot of people are reading your thread. It could be any wallet... maybe a cold storage or something. You could wait years before someone show up. And even if the one realizes, his coins are gone, how could he know you took them.

I know. Some prominent place where to announce this?
Not like
 

“There’s no point in acting surprised about it. All the planning charts and demolition orders have been on display at your local planning department in Alpha Centauri for 50 of your Earth years, so you’ve had plenty of time to lodge any formal complaint and it’s far too late to start making a fuss about it now. … What do you mean you’ve never been to Alpha Centauri? Oh, for heaven’s sake, mankind, it’s only four light years away, you know. I’m sorry, but if you can’t be bothered to take an interest in local affairs, that’s your own lookout. Energize the demolition beams.”

:-)

I think you are at a point where you have to decide if you want to stop this, because you have proven it works, or if you want to continue, but this could be called theft. Which way you will go?

I am at a point, where I am shocked, because if it turns out I was right about the BTC PK search space, "stopping this", is like you have a patient with inner bleeding, you have proven he has inner bleeding, but you just turn away, because you think "no see no harm" is a good strategy.

Seriously?

I can stop it and tomorrow a chinaman (or whoever - just an example) decides to set up a clandestine GPU farm doing exactly that.




@becoin: Too bad I see just "this user is ignored". Perhaps you should have spilled out less shit. Maybe someone else can answer you - I'm not in the mood to unignore you right now.


Rico

[becoin]

@becoin: Too bad I see just "this user is ignored". Perhaps you should have spilled out less shit. Maybe someone else can answer you - I'm not in the mood to unignore you right now.

It is convenient but childish way to avoid reasonable questions. People that ignore me usually end up wasting their time and money.

Today bitcoin price is up and I'm in a good mood. I'm sorry you're in a depressed mood right now.

[yo-blin]

Hi I stay gift on the secret page directory.io for check
I think it will be found this about one week  

You wrote this October 5th and although a week is almost over,
I assume you are not talking about
https://blockchain.info/address/1PVwqUXrD5phy6gWrqJUrhpsPiBkTnftGg
because this has been deposited in May 2014.

I made two mistakes.
1. I did not take into account the fact that the LBC checks only those addresses that were not empty at the time the .pst file
2. The current work block interval [36245522-36245537]
I forgot to divide that number by 8192

I made a transaction of the open test Page 5 October
http://btcdirectory.azurewebsites.net/30000000
after October 6th, you have made a new file balances.pst
but the purchase transaction confirmation came only October 10.
Therefore, no one has found.

Now you can use these amounts for the test after the next update balances.pst
LBS -p 3661-3663

[rico666]

  Ok Gonzo, you beat me to it:

Firstly, what you claim you've found is not a collision. Hence the name of your 'project' is misleading.

Why is it not a collision? How do YOU prove the PK used when this address was generated was not different from the PK found now?
At a time, when you by all means most certainly don't know ANY of the PKs. What is it, you do not understand in this statement:

It may very well be, that he used an entirely different PK (of the 2⁹⁶) for this address (because this one is a little bit shitty), so I assume providing ANOTHER PK to this address would count as good proof - wouldn't it?

You see - to me, the fact that you are writing many, but short and shitty (= not well thought out) texts just proves you to be some kind of poor miserable existence. Usually I do not waste my time with your kind, because there are way too many like you. So - heaven help - try to understand what this is all about, what's being written/said and THINK about it. I know it's quite a lot of text and it may be hard for you. Just try. It is ok, if you give it more time, because it seems you need more. (In everyday pool operation, I see that missing processing power can be counterbalanced by throwing more time at it.)

And secondly, how will you prove the 'rightful' owner of the PK you've found is actually not... you?

Another "I just vomited something without even spending a microsecond thinking about it".

Or maybe you are heavily infested with Dunning–Kruger and do not even know what to think about. Let me help:

Proving the non-existence of something is impossible. That is BTW also the reason why there is the presumption of innocence. You cannot prove you are innocent, you can only be proven guilty. Understood? I sure hope so.

I cannot prove to not have been the "rightful" owner of the found PK, but there is evidence now and in the future to support the claim it is an original not-belonging-to-me-or-someone-I-know PK.

Let's get some pedantic semantics out of the way first:

Now that I found it, I may very well be considered "co-owner" already.
If the PK found is not the original PK used (remember? there are like 2⁹⁶ PKs to every address) I am at the moment the sole owner of the PK found.

Ok - here comes the evidence your honor:

The unspent output is from May 2014. If you think I staged this up 2 years before I even thought about starting the project, I feel flattered. You must certainly think I am some long-term planning mastermind genius. Thank you. Unfortunately, that doesn't change what I think about you - sorry.

Not only would I have to do this at a time when I was doing completely different things, I would also have to rely on a shitty 50bit 45.x bit PK to hold for over 2 years. So I would not only be a long-term planning mastermind genius, I would also have ballz of steel. Again, thank you, thank you. But no - I didn't.

Now would be/is a really a bad time to pull something like that off. I am in the middle of releasing the next LBC and in fact today morning, I 1st checked my private messages with beta tester reports before I realized that FOUND.txt file when I wanted to fix the bugs. I also write the new manual etc. etc. Lots of work, bad time. Bad timing.

First I even thought it was another hash160 found from the puzzle transaction, when I saw it had funds on it, I thought it might have been the bounty that yo-blin announced and only when I saw the date, I realized this is different. Sure. All this is no "proof", only describing the situation as it was.

If in the future really someone comes along and presents a PK that is different from what has been found, you will a) have collision proof b) very strong evidence this is genuine. Until then, shut the fuck up unless you have to say something of importance.

Rico

[rico666]

I thought I give the shitty Amazon AWS EC2 a try.  




Rico

[cisA]

Another point of view .. -> "bitcoin lotto"  


...

An idea to close this russian roulette bug/feature ?

[rico666]

Another point of view .. -> "bitcoin lotto"  

...

An idea to close this russian roulette bug/feature ?

You mean "What can one do to prevent the pool from finding his PK?"?

At the moment my only advice is: Move your funds to a P2SH address.


Rico

[scooterking]

I compiled a new allBalances.txt file, and was curious as to why this file is not converted to a new bloom filter file ?

[scooterking]

Oh, and when I started it up in auto it never said

"Reading balances... storable found - using that (faster)"

any other flags to add ?