Bitcoin Forum
May 26, 2019, 04:34:59 PM *
News: Latest Bitcoin Core release: 0.18.0 [Torrent] (New!)
 
   Home   Help Search Login Register More  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 [36] 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 »
  Print  
Author Topic: Large Bitcoin Collider (Collision Finders Pool)  (Read 173913 times)
SlarkBoy
Member
**
Offline Offline

Activity: 108
Merit: 10


View Profile
March 28, 2017, 03:02:39 PM
 #701

Nice speed. new GPU generator  Grin

Code:
root@420vs69:~/LBC# ./LBC --id xxx --secret xxx --no_update --gpu --cpus 8
GPU authorized: yes
Ask for work... got blocks [1505284048-1505313999] (31406 Mkeys)
ooo..skip...ooo (25.37 Mkeys/s)
1558888499
Hero Member
*
Offline Offline

Posts: 1558888499

View Profile Personal Message (Offline)

Ignore
1558888499
Reply with quote  #2

1558888499
Report to moderator
GET 25 FREE SPINS AT REGISTRATION
GET 100% BONUS ON FIRST DEPOSIT
PLAY NOW
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
unknownhostname
Member
**
Offline Offline

Activity: 62
Merit: 10


View Profile
March 28, 2017, 04:28:47 PM
 #702

Will use 32 CPUs.
Testing mode. Using page 0, turning off looping.
Benchmark info not found - benchmarking... done.
Your maximum speed is 719889 keys/s per CPU core.


need GPU GPU GPU GPU  ^^
icanscript
Hero Member
*****
Offline Offline

Activity: 686
Merit: 502



View Profile
March 28, 2017, 08:13:47 PM
 #703

Hi rico666, Sent you a PM re bootable iso image,

I can do some testing on NVidia and AMD (I have an Radeon R9 270 also on another machine). Would I need different ids to test a different hardware setup?

Also, may as well add a little customisation to it. Grub screen, hostname etc Tongue

rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 28, 2017, 09:04:10 PM
 #704

Nice speed. new GPU generator  Grin

Code:
root@420vs69:~/LBC# ./LBC --id xxx --secret xxx --no_update --gpu --cpus 8
GPU authorized: yes
Ask for work... got blocks [1505284048-1505313999] (31406 Mkeys)
ooo..skip...ooo (25.37 Mkeys/s)

I'd be interested what keyrate you get from oclvanitygen on that configuration. It should be less than 25 Mkeys/s with the GPU 100% load. But I could be wrong, so I'm interested in comparative numbers.


Rico

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
SlarkBoy
Member
**
Offline Offline

Activity: 108
Merit: 10


View Profile
March 29, 2017, 03:11:05 AM
 #705


I'd be interested what keyrate you get from oclvanitygen on that configuration. It should be less than 25 Mkeys/s with the GPU 100% load. But I could be wrong, so I'm interested in comparative numbers.


Rico

I get 80 Mkeys/s with oclvanitygen. both avx2 and skylake.

CPU: Intel® Core™ i7-6700 or Intel® Core™ i7-4790K
GPU: GTX 1080


on another machine with skylake CPU, speed increased 5 Mkeys/s.

--time 40 --delay 900
Code:
Ask for work... got blocks [1529011424-1529084767] (76906 Mkeys)
oooo......oooo (30.22 Mkeys/s)
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 29, 2017, 07:22:11 AM
 #706

I get 80 Mkeys/s with oclvanitygen. both avx2 and skylake.

It's to be expected that you get same keyrate on oclvanitygen independent of the CPU, as oclvanitygen (if you have no pcre matching) is GPU-only.

I get on my GPU only between 15-16 Mkeys/s from oclvanitygen, so your GPU should be about 5 times as fast, which seems consistent with the data in the comparison chart http://www.videocardbenchmark.net/compare.php?cmp%5B%5D=3502&cmp%5B%5D=3373

However I get 20 Mkeys/s on my notebook (skylake+M2000M) - which translates to 40 MAddresses/s (where oclvanitygen only does the compressed ones, thus for oclvanitygen 15Mkeys = 15Maddresses). Making LBC on my notebook about 2,7 times faster than oclvanitygen. As of now.

Not true for your system. :-(  For you, the CPUs are still the bottleneck.

So you get 50-60 MAddresses/s from LBC, but 80 MAddresses from oclvanitygen. Your GPU load with LBC is around 50% I assume, where oclvanitygen shows 100%?


Rico

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 29, 2017, 08:40:13 AM
 #707

Highest block number for #51 is 2147483648
We are currently at 1541746271

=> 605737377 blocks to go
=> 635161675825152 keys to go

current keyrate 564310000 keys/s

=> 1125554 seconds = 13 days


Rico

PS: Of course, the numbers will be outdated when I press "Post"  Smiley

PPS: As a reminder for why there is a difference: We "jumped" forward in search space by about 370 Mblocks, which we'll have to work after #51 is found.

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
unknownhostname
Member
**
Offline Offline

Activity: 62
Merit: 10


View Profile
March 29, 2017, 09:30:10 AM
 #708

with oclvanitygen I was getting 150Mkeys with a 4x K80 GPU's
SlarkBoy
Member
**
Offline Offline

Activity: 108
Merit: 10


View Profile
March 29, 2017, 11:28:09 AM
 #709

It's to be expected that you get same keyrate on oclvanitygen independent of the CPU, as oclvanitygen (if you have no pcre matching) is GPU-only.

I get on my GPU only between 15-16 Mkeys/s from oclvanitygen, so your GPU should be about 5 times as fast, which seems consistent with the data in the comparison chart http://www.videocardbenchmark.net/compare.php?cmp%5B%5D=3502&cmp%5B%5D=3373

However I get 20 Mkeys/s on my notebook (skylake+M2000M) - which translates to 40 MAddresses/s (where oclvanitygen only does the compressed ones, thus for oclvanitygen 15Mkeys = 15Maddresses). Making LBC on my notebook about 2,7 times faster than oclvanitygen. As of now.

Not true for your system. :-(  For you, the CPUs are still the bottleneck.

So you get 50-60 MAddresses/s from LBC, but 80 MAddresses from oclvanitygen. Your GPU load with LBC is around 50% I assume, where oclvanitygen shows 100%?


Rico


yeah.
avx2+gpu = 65% - 70%
skylake+gpu = around 77%

But I like this generator right now. Not really disappointed with the speed and not taking 100% usage. Because I still can watch youtube video and gambling. to pay this dedicated servers.  Grin
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 30, 2017, 07:57:39 AM
 #710

Could the user operating currently(!) under the Id "johnsnow" contact me ASAP?

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
Jude Austin
Legendary
*
Offline Offline

Activity: 1139
Merit: 1000


The Real Jude Austin


View Profile WWW
March 30, 2017, 09:57:37 AM
 #711

Could the user operating currently(!) under the Id "johnsnow" contact me ASAP?

#51??

Get paid BTC to sign up for free tokens: http://earn.com/judeaustin/referral/?a=d3euriwoffdrlv4b
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 30, 2017, 10:25:17 AM
 #712

Could the user operating currently(!) under the Id "johnsnow" contact me ASAP?

#51??

Better. :-)

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
Jude Austin
Legendary
*
Offline Offline

Activity: 1139
Merit: 1000


The Real Jude Austin


View Profile WWW
March 30, 2017, 10:30:52 AM
 #713

Could the user operating currently(!) under the Id "johnsnow" contact me ASAP?

#51??

Better. :-)

He tripped over a gold mine didn't he? lol

Get paid BTC to sign up for free tokens: http://earn.com/judeaustin/referral/?a=d3euriwoffdrlv4b
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 30, 2017, 10:39:47 AM
 #714

Better. :-)

He tripped over a gold mine didn't he? lol

I think even better. I'm waiting for confirmation from another user, then I will make an announcement.

*** This is not related to #51 ***

(however, we're over 50% search space, so this too any moment now - I think.)

Rico

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
Jude Austin
Legendary
*
Offline Offline

Activity: 1139
Merit: 1000


The Real Jude Austin


View Profile WWW
March 30, 2017, 10:41:51 AM
 #715


I think even better. I'm waiting for confirmation from another user, then I will make an announcement.

*** This is not related to #51 ***

(however, we're over 50% search space, so this too any moment now - I think.)

Rico

Nice, excited to see what it is!

Jude

Get paid BTC to sign up for free tokens: http://earn.com/judeaustin/referral/?a=d3euriwoffdrlv4b
unknownhostname
Member
**
Offline Offline

Activity: 62
Merit: 10


View Profile
March 30, 2017, 11:48:07 AM
 #716

Could the user operating currently(!) under the Id "johnsnow" contact me ASAP?

#51??

Better. :-)

#52?? ^^

-------------------------------------------------------------------

23    johnsnow    3822    4m23    johnsnow    3822    4m
1    Unknownhostname    661507    31s ---- around 175 times more ... still I never find anything Smiley)
Real-Duke
Legendary
*
Offline Offline

Activity: 1568
Merit: 1048



View Profile
March 30, 2017, 12:07:12 PM
 #717

23    johnsnow    3822    4m23    johnsnow    3822    4m
1    Unknownhostname    661507    31s ---- around 175 times more ... still I never find anything Smiley)

You're not alone! Me too...even with less power  Wink
I wish rico has been searched for the owner of the ID RealDuke instead of johnsnow  Cool

......
.L I V E C O I N . N E T.
.
..PROFITBOX..
██  █████████████████████████
  █████████▄      ▄██████████
█████████████▄  ▄████████████
    █████████████████████████
  ██████████▀    ▀█ ▀████████
████  █████▀  ▄▄  ▀█  ▀██████
  ████████▀  ▄██▄  ▀█   ▀████
    ██████   ▀██▀   ██   ████
  █████████▄      ▄██████████
██  █████████▄  ▄████████████
  ███████████████████████████
██  █████████████████████████
  █████████████████████▀ ███
█████████████████████▀   ███
    █████████████▀     ████
  █████████████▀   ██    ████
████  █████▀     ██    ████
  ███████▀   ██    ██    ████
    █████    ██    ██    ████
  ███████    ██    ██    ████
██  █████    ██    ██    ████
  ███████████████████████████
.....
johnsnow1
Newbie
*
Offline Offline

Activity: 1
Merit: 0


View Profile
March 30, 2017, 12:12:36 PM
 #718

23    johnsnow    3822    4m23    johnsnow    3822    4m
1    Unknownhostname    661507    31s ---- around 175 times more ... still I never find anything Smiley)

You're not alone! Me too...even with less power  Wink
I wish rico has been searched for the owner of the ID RealDuke instead of johnsnow  Cool

I don't know it yet either. But supposedly some other user has also found it. We'll have to wait and see what Rico tells us.
unknownhostname
Member
**
Offline Offline

Activity: 62
Merit: 10


View Profile
March 30, 2017, 12:29:16 PM
 #719

I don't know it yet either. But supposedly some other user has also found it. We'll have to wait and see what Rico tells us.

Double found Smiley is better ^^
rico666
Legendary
*
Offline Offline

Activity: 1064
Merit: 1022


฿ → ∞


View Profile WWW
March 30, 2017, 01:00:50 PM
Last edit: March 30, 2017, 06:42:42 PM by rico666
 #720

1    Unknownhostname    661507    31s ---- around 175 times more ... still I never find anything Smiley)

A quantum of solace may be that you are co-responsible for this find.  Smiley



Ladies, Gentlemen and becoin!


We are proud to announce a new genuine find of the LBC pool. Please stand by for the technical details while we move the funds to a custodial address.

https://lbc.cryptoguru.org/trophies

2017-03-30 01:18:00 UTC

The pool found a private key to 7d89ad89cd10a3867b8f6bfc803838fa101b598b (1CSnQ1LnY37rwz8ezJn5xQrCrifZxExpWV) as 0x5e1667c899783. At the time of the find, there were 0.00001 BTC on that address.The funds were transferred to custody at 1Dg1XnH9BLKFf4XrWioYsxDJjSxr996Miq.



Finder is id Janussss (here user Janu$$ - currently at position 46 with 881 delivered Gkeys, operating 3(?) LBC CPU-clients in the vmware appliance. (I may not be exactly right - just so you get an impression about the luck factor in this project.)

Best story: How did I find out? Because at 1:18 a.m. UTC I was snoring peacefully and still I knew sooner than him.


For this, we have to go back 24 hours: I am trying to make a GPU client for unknownhostname so it runs on these fucking Google/Amazon cloudshizz servers. I know that by providing unknownhostname with a decent GPU client, time and space will fold into a singularity, but still I want to see how the pool goes beyond 1Gkeys/s - so I try. Hard. Unfortunately no success, because the shizm is strong in the cloud.

I have a GPU client that does not print @ GPU, but it starts crashing elsewhere. Weird errors and I'm about to give up. I write a PM to arulbero and whine a little bit about the fact

Good news is, that the keyrate raise you see on https://lbc.cryptoguru.org/stats is mostly due to your ECC arithmetics. It's users deploying the new clients. It's not more users.

I've been fighting for a day with a new GPU client for "unknownhostname" and his Tesla GPUs. While I do have a client that should not crash, libGMP is driving me crazy.

I see this error

Code:
root@LBC:~/collider/beta# time ./hrd-core -I 0000000000000000000000000000001111111111111111111111111111111111 -c 10000 -L 8 -p
Key given: 1111111111111111111111111111111111
but the first key must be in the range:
1  -  fffffffffffffffffffffffffffffffebaaedce6af48a03bbfd25e8ccf364141
Please retry with another key

real    0m0.046s
user    0m0.000s
sys     0m0.033s
root@LBC:~/collider/beta# time ./hrd-core -I 0000000000000000000000000000000000000000000000000000000000000001 -c 10000 -L 8 -p
Key given: 1
but the first key must be in the range:
1  -  fffffffffffffffffffffffffffffffebaaedce6af48a03bbfd25e8ccf364141
Please retry with another key

real    0m0.044s
user    0m0.004s
sys     0m0.030s

where I already have modified the printout and as you can see the key given is correct. So I assume there is something wrong with this code

Code:
if(mpn_cmp(kr, klim1r, 4)<0 || mpn_cmp(kr, klim2r, 4)>0       ){

but I'm not sure what???

One of the last things I do is to write a debug message in the code gmp_printf("Set privkey to: %Zx\n", k); and change the code to a
Code:
if(mpz_cmp(kr, klim1r)<0 || mpz_cmp(kr, klim2r)>0
. It works! I'm happy and thrilled, make a avx2+gpu client and write that to arulbero.

Then I try it on a Amazon AWS:

there is something wrong with this code

Code:
if(mpn_cmp(kr, klim1r, 4)<0 || mpn_cmp(kr, klim2r, 4)>0       ){

I replaced it with what seemed to be the intuitive solution

Code:
 if (mpz_cmp(k, klim1) < 0 || mpz_cmp(k, klim2) > 0) {
    gmp_printf("Key given: %Zx\nbut the first key must be in the range:\n%Zx  -  %Zx\n", k, klim1, klim2);
    printf("Please retry with another key\n");
    return 0;
  }

And now it's crashing elsewhere... *sigh*


Code:
root@LBC:~/collider/beta# time ./hrd-core -I 0000000000000000000000000000000000000000000000000000000000000001 -c 10000 -L 8 -p
hrd-core: malloc.c:3700: _int_malloc: Assertion `victim->fd_nextsize->bk_nextsize == victim' failed.

Roll Eyes No! I am the victim here.  Cheesy


With that I give up and go to sleep.

Meanwhile...  the LBC user with the Id johnsnow comes along and because he got a gpuauth=1 recently he downloads/updates (I guess, he may clarify exactly what happened) the GPU avx2+gpu client which - unfortunately - has still the debug message in it.

While he does LBC -x, he is probably not yet familiar with LBC and so when it starts to print "Set privkey to: 0x......" he is not alarmed and operates the client.

The night passes in Europe, I wake up and one of my 1st things is to check the logs. And I see in the logs for a particular client (that of johnsnow) <timestamp> <ip> invalid challenge response "Set privkey to 0x....." instead of a correct challenge response. I have long forgotten that it was actually me who had this debug message in the GPU generator(!). I thought WTF? and grepped for the string in the client. Nothing. In the server: Nothing.

I thought maybe the NSA was hacking/mocking me. (Not fully awake probably) So I tried one of the privkeys:

Code:
$ time gen-hrdcore-skylake+gpu-linux64  -I 0000000000000000000000000000000000000000000000000005e1667b699001 -c 10000 -L 8 -d 1
response: 30-54-0
7d89ad89cd10a3867b8f6bfc803838fa101b598b:u:priv:0000000000000000000000000000000000000000000000000005e1667c899001+0x782
response: 30-48-0
response: 30-14-0
response: 30-61-0
response: 30-26-0
response: 30-12-0
response: 30-2-34"3DUfw��
response: 30-73-0

real    0m23.895s
user    0m22.687s
sys     0m1.160s


*BAMM*  Shocked

Now I was definitely thinking NSA was mocking me and I wrote the "may johnsnow contact me ASAP" message.
He did and was very helpful, sending screenshots what happened on his end, reassuring me he had not tampered with the code, so slowly I began to realize what had happened.

I had a look who got the block interval that should have found this key: Id Janussss

I wrote to Janu$$ asking if he was actually Janussss (which I assumed) and wanted him to please confirm he actually had something in his FOUND.txt.

After a while he answered confirming he was Id Janussss, but having no FOUND.txt. That was a shock! Was the generator not working again? (some may remember the episode with the win clients) And while I was struggling with life, Janu$$ wrote a PM titled  ____TREFFER IN FOUND.txt____ (HIT IN FOUND.txt) - so he found FOUND.txt  Cheesy

The balance to the universe was restored.



Well - not entirely.


The pool found a 3rd genuine address with unspent outputs. Again uncompressed, again a minuscule amount. I seriously doubt though, that this address is an uncompressed address originally and I also doubt this is an experimental address. But I might be wrong and I would like to ask some independent party if he could perform some forensic analysis on the 3 addresses found.

edit: Also, I really am not comfortable with the fact, that we found out only by massive luck. While I do believe that Janu$$ would have found "FOUND.txt", I certainly would wish for something more failproof. Probably installing some kind of hook-find automatically or such.

Are they connected somehow?
How probable is it these from 2016 were originally actually uncompressed addresses?
Are they changes or are they the main transaction?
etc. Huh

Because especially for this address I have a - call it feeling - that it was created as a compressed address and we found a privkey which resolves to this hash160 hashing the uncompressed pubkey. This would be a school book collision!

IMHO - if we could track down for https://blockchain.info/de/address/1JwxZKTZiCqzBdYS38QM7ukKYqsSWQFRpT at least the info if that is a compressed or uncompressed address, then for a "compressed" answer we would have extremely strong indication of a collision.


Rico

all non self-referential signatures except mine are lame ... oh wait ...   ·  LBC Thread (News)  ·  BURST Activities
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 [36] 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 »
  Print  
 
Jump to:  

Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!