[BBR] Boolberry: Privacy and Security - Guaranteed Since 2014 - page 375.

doldgigger

full member

Activity: 170

Merit: 100

Quote from: crypto_zoidberg on April 19, 2014, 08:34:37 PM

Transaction outs with guaranteed anonymity. An additional attribute for out is added: it can ONLY be used in the transaction with mixin level in ring signature not less than a given number. The feature is added to ensure the user's anonymity of transactions ring signature over time.

I might be missing something, but wouldn't it be more effective to enforce the mixin level on a more global level? Mixing in a key which appears somewhere else in an unmixed fashion will be less effective against smart blockchain analysis, after all.

clintar

full member

Activity: 212

Merit: 100

Quote from: otila on July 12, 2014, 04:37:01 AM

I have some optimizations made for boolberry, I hope you like them.

boolberry-reciprocal-divide.patch 25-33% speedup. Some header files could need some cleaning up.
If your compiler complains about typeof in helper.h , rename it to __typeof__ or use CMAKE_CXX_FLAGS -std=gnu++11
https://docs.google.com/file/d/0B9h02pRWKaODN29mUDdqOXZFbDA/edit

faster (Linux) and simpler keccak
https://docs.google.com/file/d/0B9h02pRWKaODZXBSclBITmxIWDQ/edit

Linux only: use jemalloc instead of glibc for memory allocation, 6-23% speedup.
If you have a poor Linux distribution which does not have jemalloc package, get it from https://github.com/jemalloc/jemalloc.git

Code:

--- CMakeLists.txt~	2014-07-12 11:18:09.698295127 +0300
+++ CMakeLists.txt	2014-07-12 11:18:11.921305454 +0300
@@ -117,7 +117,7 @@
   set(Boost_LIBRARIES "${Boost_LIBRARIES};ws2_32;mswsock")
 elseif(NOT MSVC)
 if(NOT APPLE)
-  set(Boost_LIBRARIES "${Boost_LIBRARIES};rt;pthread")
+  set(Boost_LIBRARIES "${Boost_LIBRARIES};rt;pthread;jemalloc")
 endif()
 endif()

I have also patch for cpuminer , but it has that bloat keccak code which I would first try to get rid of and use that simpler one above..

EDIT: patch is GPL

oooh boy was that fun to compile

otila

sr. member

Activity: 336

Merit: 250

I have some optimizations made for boolberry, I hope you like them.

boolberry-reciprocal-divide.patch 25-33% speedup. Some header files could need some cleaning up.
If your compiler complains about typeof in helper.h , rename it to __typeof__ or use CMAKE_CXX_FLAGS -std=gnu++11
https://docs.google.com/file/d/0B9h02pRWKaODN29mUDdqOXZFbDA/edit

faster (Linux) and simpler keccak
https://docs.google.com/file/d/0B9h02pRWKaODZXBSclBITmxIWDQ/edit

Linux only: use jemalloc instead of glibc for memory allocation, 6-23% speedup.
If you have a poor Linux distribution which does not have jemalloc package, get it from https://github.com/jemalloc/jemalloc.git

Code:

--- CMakeLists.txt~	2014-07-12 11:18:09.698295127 +0300
+++ CMakeLists.txt	2014-07-12 11:18:11.921305454 +0300
@@ -117,7 +117,7 @@
   set(Boost_LIBRARIES "${Boost_LIBRARIES};ws2_32;mswsock")
 elseif(NOT MSVC)
 if(NOT APPLE)
-  set(Boost_LIBRARIES "${Boost_LIBRARIES};rt;pthread")
+  set(Boost_LIBRARIES "${Boost_LIBRARIES};rt;pthread;jemalloc")
 endif()
 endif()

I have also patch for cpuminer , but it has that bloat keccak code which I would first try to get rid of and use that simpler one above..

EDIT: patch is GPL

dga

hero member

Activity: 737

Merit: 511

Quote from: crypto_zoidberg on July 11, 2014, 08:18:04 PM

Quote from: jl777 on July 11, 2014, 08:05:17 PM

Quote from: crypto_zoidberg on July 11, 2014, 07:51:55 AM

Quote from: bigj on July 11, 2014, 02:32:23 AM

Quote from: dga on July 10, 2014, 02:30:11 PM

Quote from: MagicSnow on July 10, 2014, 02:14:41 PM

WTF... Network hashrate 3.28 GH/s, i found only two pools, one with around 5.45 MH/s the second with around 2.25 MH/s... someone is solo mining really hard or there is other bigger pools?

It's a mix of a few big solo miners and some private pools. I still solo mine at home on 5 machines, for example - I get a block every other day or so. You don't have to be huge to solo BBR still.

Not sure if it's still happening, but for a while, some of the big EC2 miners would run their own pool inside EC2 to keep the bandwidth charges down and latency low.

If you've got more than about 1-2MH/s, I'd solo, but that's just me.

I'm kind of new to the BBR business, but I get suspicious when I see this tremendously skewed distribution of mining power.

Given that cpuminer comes (roughly) with the same hash power as the XMR miner (in the 100-600 kilohash range on standard quad-cores), a total network power of 3.71 GH/s seems really odd to me. How is that even possible without gpu miners? Ok, it's probably not possible without them.. so we have some gpu barons out there. That's fine, the coin evolves. But even then, a network power in the gigahash(!) range is hardly explained by cpu+gpu miners. So, I suspect either (big!) botnets or fpgas in the game. From the technical point of view, the latter would be great for BBR. From a practical point of view, both alternatives are bad, as it means that very few miners are actually generating most of the coins -- and thus, are in almost perfectly control of distribution and price.

Am I mistaken here, or do we have a seruious problem with BBR and the miners here?

Me and also David (dga) already commented such questions here https://bitcointalksearch.org/topic/m.7479014, please review this post.

There is clearly someone that keeps selling large amounts of BBR daily, thus clearing the buy side of the orderbooks. Price does not seem to matter much, indicating his cost basis is significantly below current prices. So, either one of the original miners or some GPU enhanced mining. The amount of sales are around 10,000 per day, which is pretty close to the new BBR created...

So, is there a plan to make BBR mining much more difficult for GPU? It seems pretty evident that the cost of CPU mining is nowhere near as good as this mythical GPU guy. I just dont see people selling CPU mined coins below cost.
If there is a way to make the scratchpad 1 gig? That would sure put a dent into any GPU efforts. We can then see if the daily 10000 sales stops.

Is anybody against this?

James

I am, against, because:
1. 1GB of scratchpad gonna kill possibility of SPV client.
2. GPU miner will help coin to withstand botnets.
3. "mythical GPU guy" is cbuchner1&Co, according to his posts he is going to release his GPU miner for BBR.

Re #3: I'm curious about wolf0's comment. The hashrate does seem in excess of what cbuchner is claiming, but who knows. It seems rational that there's someone else with a GPU miner out there, and probably has been for a bit.

But I concur - mostly because of the potential for unintended consequences. I think we should (politely) ask cbuchner1 the release schedule, and in the meantime, I'll get back to thinking about an OpenCL version even though we now have stratum.

Notably, 1 gig wouldn't hurt a GPU much, and might be even worse for CPUs. We've already seen the hashrate on CPU drop as the scratchpad exceeds L3 cache, and remember that GPUs excel first in computation, but second in memory bandwidth. They're not super-duper for random accesses, but the 256 bit line fetch for Wild Keccak isn't extremely unfriendly, either. Most GPUs could do 1GB today. 8 would cut out nearly all of the commodity GPUs, but then again, it would start eliminating a lot of people's CPU-based systems too!

kingscrown

hero member

Activity: 672

Merit: 500

http://fuk.io - check it out!

nevermind - just had ot change pool toget -k thinge

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: nexxan on July 11, 2014, 08:30:36 PM

@cryptozoidberg
Do reward blocks give out the transaction fee?

A dice site that can use the blockchain to start off with would be nice to try and increase txn counts and fees given to miners.

Yes, sure, block reward includes fee.

nexxan

newbie

Activity: 7

Merit: 0

I have started a pool for a few of my machines. Working on the getting started guides and will hopefully have some binaries put together shortly.

I will have a proper thread announcement soon but would like to see if anyone can donate some hash time so I can try to ensure stability and resource usage.

http://www.nassive.com/

The pool was just built from the newest boolbd, and pool-boolberry-node with support for auto-updated scratchpad.bin

I am also wondering how people would feel with a dice site for BBR?

@cryptozoidberg
Do reward blocks give out the transaction fee?

A dice site that can use the blockchain to start off with would be nice to try and increase txn counts and fees given to miners.

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: kingscrown on July 11, 2014, 08:23:08 PM

can you gfive step by step to run cpuminer for windows? im gettign some scratchpad error.

Which error amigo ? Can you post it here ?
We also have HOWTO here: http://boolberry.com/howto.html#winquick

kingscrown

hero member

Activity: 672

Merit: 500

http://fuk.io - check it out!

can you gfive step by step to run cpuminer for windows? im gettign some scratchpad error.

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: jl777 on July 11, 2014, 08:05:17 PM

Quote from: crypto_zoidberg on July 11, 2014, 07:51:55 AM

Quote from: bigj on July 11, 2014, 02:32:23 AM

Quote from: dga on July 10, 2014, 02:30:11 PM

Quote from: MagicSnow on July 10, 2014, 02:14:41 PM

WTF... Network hashrate 3.28 GH/s, i found only two pools, one with around 5.45 MH/s the second with around 2.25 MH/s... someone is solo mining really hard or there is other bigger pools?

It's a mix of a few big solo miners and some private pools. I still solo mine at home on 5 machines, for example - I get a block every other day or so. You don't have to be huge to solo BBR still.

Not sure if it's still happening, but for a while, some of the big EC2 miners would run their own pool inside EC2 to keep the bandwidth charges down and latency low.

If you've got more than about 1-2MH/s, I'd solo, but that's just me.

I'm kind of new to the BBR business, but I get suspicious when I see this tremendously skewed distribution of mining power.

Given that cpuminer comes (roughly) with the same hash power as the XMR miner (in the 100-600 kilohash range on standard quad-cores), a total network power of 3.71 GH/s seems really odd to me. How is that even possible without gpu miners? Ok, it's probably not possible without them.. so we have some gpu barons out there. That's fine, the coin evolves. But even then, a network power in the gigahash(!) range is hardly explained by cpu+gpu miners. So, I suspect either (big!) botnets or fpgas in the game. From the technical point of view, the latter would be great for BBR. From a practical point of view, both alternatives are bad, as it means that very few miners are actually generating most of the coins -- and thus, are in almost perfectly control of distribution and price.

Am I mistaken here, or do we have a seruious problem with BBR and the miners here?

Me and also David (dga) already commented such questions here https://bitcointalksearch.org/topic/m.7479014, please review this post.

There is clearly someone that keeps selling large amounts of BBR daily, thus clearing the buy side of the orderbooks. Price does not seem to matter much, indicating his cost basis is significantly below current prices. So, either one of the original miners or some GPU enhanced mining. The amount of sales are around 10,000 per day, which is pretty close to the new BBR created...

So, is there a plan to make BBR mining much more difficult for GPU? It seems pretty evident that the cost of CPU mining is nowhere near as good as this mythical GPU guy. I just dont see people selling CPU mined coins below cost.
If there is a way to make the scratchpad 1 gig? That would sure put a dent into any GPU efforts. We can then see if the daily 10000 sales stops.

Is anybody against this?

James

I am, against, because:
1. 1GB of scratchpad gonna kill possibility of SPV client.
2. GPU miner will help coin to withstand botnets.
3. "mythical GPU guy" is cbuchner1&Co, according to his posts he is going to release his GPU miner for BBR.

jl777

legendary

Activity: 1176

Merit: 1134

Quote from: crypto_zoidberg on July 11, 2014, 07:51:55 AM

Quote from: bigj on July 11, 2014, 02:32:23 AM

Quote from: dga on July 10, 2014, 02:30:11 PM

Quote from: MagicSnow on July 10, 2014, 02:14:41 PM

WTF... Network hashrate 3.28 GH/s, i found only two pools, one with around 5.45 MH/s the second with around 2.25 MH/s... someone is solo mining really hard or there is other bigger pools?

It's a mix of a few big solo miners and some private pools. I still solo mine at home on 5 machines, for example - I get a block every other day or so. You don't have to be huge to solo BBR still.

Not sure if it's still happening, but for a while, some of the big EC2 miners would run their own pool inside EC2 to keep the bandwidth charges down and latency low.

If you've got more than about 1-2MH/s, I'd solo, but that's just me.

I'm kind of new to the BBR business, but I get suspicious when I see this tremendously skewed distribution of mining power.

Given that cpuminer comes (roughly) with the same hash power as the XMR miner (in the 100-600 kilohash range on standard quad-cores), a total network power of 3.71 GH/s seems really odd to me. How is that even possible without gpu miners? Ok, it's probably not possible without them.. so we have some gpu barons out there. That's fine, the coin evolves. But even then, a network power in the gigahash(!) range is hardly explained by cpu+gpu miners. So, I suspect either (big!) botnets or fpgas in the game. From the technical point of view, the latter would be great for BBR. From a practical point of view, both alternatives are bad, as it means that very few miners are actually generating most of the coins -- and thus, are in almost perfectly control of distribution and price.

Am I mistaken here, or do we have a seruious problem with BBR and the miners here?

Me and also David (dga) already commented such questions here https://bitcointalksearch.org/topic/m.7479014, please review this post.

There is clearly someone that keeps selling large amounts of BBR daily, thus clearing the buy side of the orderbooks. Price does not seem to matter much, indicating his cost basis is significantly below current prices. So, either one of the original miners or some GPU enhanced mining. The amount of sales are around 10,000 per day, which is pretty close to the new BBR created...

So, is there a plan to make BBR mining much more difficult for GPU? It seems pretty evident that the cost of CPU mining is nowhere near as good as this mythical GPU guy. I just dont see people selling CPU mined coins below cost.
If there is a way to make the scratchpad 1 gig? That would sure put a dent into any GPU efforts. We can then see if the daily 10000 sales stops.

Is anybody against this?

James

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Dear friends ( especially pool operators )!

Last commit enabled automatic scratchfile updating from boolbd daemon.
Now you just have to configure your pool with target folder/scratchpad.bin file and time interval for saving scratchpad, and pool via daemon rpc will be regularly update shared scratchpad file.

!!!!! This update will work only with last version of daemon!!!!!!

Instructions:
1. Pull daemon sources and rebuild it.
2. Pull pool seurces.
3. Update your config.json with following lines:

Quote

"clusterForks": "auto",
"poolAddress": "1L1ZPC9XodC6g5BX8j8m3vcdkXPiZrVF7RcERWE879coQDWiztUbkkVZ86o43P27Udb3qxL4B41gbaG pvj3nS7DgFZauAZE",
"scratchpadFilePath": "/usr/share/nginx/html/scratchpad.bin",
"scratchpadFileUpdateInterval": 14400000,
"blockRefreshInterval": 1000,

scratchpadFileUpdateInterval - interval for scratchpad updating(milliseconds), 14400000 by default (4 hours).

4. Turn your http server to share this file and update mining instructions to make command line point to this file via URL.

Good "luck"!

booletic

newbie

Activity: 13

Merit: 0

bbr.farm: is updated to the latest code,
we also upgraded our server to welcome more miners.

Please use stratum, and happy mining

Code:

./minerd -a wildkeccak -o stratum+tcp://bbr.farm:7777 -u YOUR_WALLET_ADDRESS -p X -P -D -t NUMBER_OF_THREADS -k http://bbr.farm/download/scratchpad.bin

You can find me on #boolberry channel on FreeNode

btc-mike

hero member

Activity: 938

Merit: 1001

Quote from: MagicSnow on July 11, 2014, 01:55:29 PM

What is happening with the bbr.farm pool?

Not sure. I notified pool operator.

Extremepool.org is operating normally.

MagicSnow

newbie

Activity: 55

Merit: 0

What is happening with the bbr.farm pool?

maxcan

member

Activity: 117

Merit: 100

Quote from: crypto_zoidberg on July 11, 2014, 11:59:05 AM

Quote from: 33zer0w0lf on July 11, 2014, 11:53:05 AM

Quote from: crypto_zoidberg on July 11, 2014, 07:54:58 AM

.......

sorry I was away for a little while.
http://boolberry.extremepool.org has been updated to the latest code.

Great! Thank you!

PS: Yet another update is coming soon

Can't wait

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: 33zer0w0lf on July 11, 2014, 11:53:05 AM

Quote from: crypto_zoidberg on July 11, 2014, 07:54:58 AM

.......

sorry I was away for a little while.
http://boolberry.extremepool.org has been updated to the latest code.

Great! Thank you!

PS: Yet another update is coming soon

33zer0w0lf

full member

Activity: 224

Merit: 100

Quote from: crypto_zoidberg on July 11, 2014, 07:54:58 AM

Quote from: Sy on July 11, 2014, 04:11:38 AM

Is extremepool broken?

Just pulled cpuminer-multi, added the -k and all my linux miner spam alot of mining stuff (good) interuppted (bad) by

Quote

[2014-07-11 14:10:58] Re-login, disconnecting...
[2014-07-11 14:10:58] Re-connec... and relogin...

and the pool is rejecting all shares.

Quote

-o stratum+tcp://boolberry.extremepool.org:7777 -k http://www.extremepool.org/download/scratchpad.bin

Yesterday i've pushed major update for pool, probably boolberry.extremepool.org was not updated yet when you had tryed.

sorry I was away for a little while.
http://boolberry.extremepool.org has been updated to the latest code.

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: Sy on July 11, 2014, 04:11:38 AM

Is extremepool broken?

Just pulled cpuminer-multi, added the -k and all my linux miner spam alot of mining stuff (good) interuppted (bad) by

Quote

[2014-07-11 14:10:58] Re-login, disconnecting...
[2014-07-11 14:10:58] Re-connec... and relogin...

and the pool is rejecting all shares.

Quote

-o stratum+tcp://boolberry.extremepool.org:7777 -k http://www.extremepool.org/download/scratchpad.bin

Yesterday i've pushed major update for pool, probably boolberry.extremepool.org was not updated yet when you had tryed.

crypto_zoidberg

hero member

Activity: 976

Merit: 646

Quote from: bigj on July 11, 2014, 02:32:23 AM

Quote from: dga on July 10, 2014, 02:30:11 PM

Quote from: MagicSnow on July 10, 2014, 02:14:41 PM

WTF... Network hashrate 3.28 GH/s, i found only two pools, one with around 5.45 MH/s the second with around 2.25 MH/s... someone is solo mining really hard or there is other bigger pools?

It's a mix of a few big solo miners and some private pools. I still solo mine at home on 5 machines, for example - I get a block every other day or so. You don't have to be huge to solo BBR still.

Not sure if it's still happening, but for a while, some of the big EC2 miners would run their own pool inside EC2 to keep the bandwidth charges down and latency low.

If you've got more than about 1-2MH/s, I'd solo, but that's just me.

I'm kind of new to the BBR business, but I get suspicious when I see this tremendously skewed distribution of mining power.

Given that cpuminer comes (roughly) with the same hash power as the XMR miner (in the 100-600 kilohash range on standard quad-cores), a total network power of 3.71 GH/s seems really odd to me. How is that even possible without gpu miners? Ok, it's probably not possible without them.. so we have some gpu barons out there. That's fine, the coin evolves. But even then, a network power in the gigahash(!) range is hardly explained by cpu+gpu miners. So, I suspect either (big!) botnets or fpgas in the game. From the technical point of view, the latter would be great for BBR. From a practical point of view, both alternatives are bad, as it means that very few miners are actually generating most of the coins -- and thus, are in almost perfectly control of distribution and price.

Am I mistaken here, or do we have a seruious problem with BBR and the miners here?

Me and also David (dga) already commented such questions here https://bitcointalksearch.org/topic/m.7479014, please review this post.

Topic: [BBR] Boolberry: Privacy and Security - Guaranteed Since 2014 - page 375. (Read 1210752 times)