BitShares Forum
Other => Graveyard => BitShares PTS => Topic started by: goldbishop on November 07, 2013, 08:13:47 pm
-
Hi i want to ask.. is possible start more instance of miner on linux? (acutally i followed this guide :http://bitsharestalk.org/index.php?topic=55.0 (http://bitsharestalk.org/index.php?topic=55.0)) if it's possible..how to?
Thanks
-
1) Install ProtoShares
2) Run PS for the first time and let it download the blockchain (otherwise you're going to have a ton of wasted electricity) - waiting for all the other block chains to download
3) Create new directories up to the amount of instances you want to run on the server. Rule of thumb is 1 instance per core and 2GB ram (create a swap file if you don;t have enough ram)
cp .protoshares PS/1
cp .protoshares PS/2
... N
Put the following protoshares.conf in each of the directories (making changes to the port and genlimit as necessary)
addnode=162.243.67.4
addnode=162.243.54.126
addnode=37.139.29.236
addnode=64.90.183.137
addnode=180.183.205.118
addnode=162.243.61.72
addnode=64.90.183.137
genproclimit=x
gen=1
port=833x
rpcport=843x
Replace the X with the instance number, and genproclimit with 1-3
4) Create a startPS.sh file with the following inside:
#!/bin/sh
/root/ProtoShares/src/bitcoind -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
...
Up to your server load
5) chmod +x startPS.sh
6) Profit
Why the sleep in there. I found that trying to start all the instances at once doesn't allow the instances enough time to start hashing. Running it without the sleep takes about 45 minutes before hashing starts, running it with the sleep starts the hashing after about 6 minutes for the first instance, 12 for 1+2, and another hash every minute after that.
If you find that useful you can always donate :P
PbwQHoJi7PKt9KCujbscNWfH43J4aCpPmN
-
1) Install ProtoShares
2) Run PS for the first time and let it download the blockchain (otherwise you're going to have a ton of wasted electricity) - waiting for all the other block chains to download
3) Create new directories up to the amount of instances you want to run on the server. Rule of thumb is 1 instance per core and 2GB ram (create a swap file if you don;t have enough ram)
cp .protoshares PS/1
cp .protoshares PS/2
... N
Put the following protoshares.conf in each of the directories (making changes to the port and genlimit as necessary)
addnode=162.243.67.4
addnode=162.243.54.126
addnode=37.139.29.236
addnode=64.90.183.137
addnode=180.183.205.118
addnode=162.243.61.72
addnode=64.90.183.137
genproclimit=x
gen=1
port=833x
rpcport=843x
Replace the X with the instance number, and genproclimit with 1-3
4) Create a startPS.sh file with the following inside:
#!/bin/sh
/root/ProtoShares/src/bitcoind -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
...
Up to your server load
5) chmod +x startPS.sh
6) Profit
Why the sleep in there. I found that trying to start all the instances at once doesn't allow the instances enough time to start hashing. Running it without the sleep takes about 45 minutes before hashing starts, running it with the sleep starts the hashing after about 6 minutes for the first instance, 12 for 1+2, and another hash every minute after that.
If you find that useful you can always donate :P
PbwQHoJi7PKt9KCujbscNWfH43J4aCpPmN
Wonderful, thank you.
-
once i'll get a block i will send you a donation! <3
(http://i.imgur.com/MEeIQsx.png)
just a quest : the balance address is only one for all istance or every istance has own address?
-
Is there a typo here? Should the datadir be /root/PS/2 for the second one?
4) Create a startPS.sh file with the following inside:
#!/bin/sh
/root/ProtoShares/src/bitcoind -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
-
Is there a typo here? Should the datadir be /root/PS/2 for the second one?
4) Create a startPS.sh file with the following inside:
#!/bin/sh
/root/ProtoShares/src/bitcoind -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
sleep 360
/root/ProtoShares/src/bitcoind -datadir=/root/PS/1 -daemon
yes it's root/ name of your directory -daemon
-
Compile a 64-bit version on Linux, and you don't need multiple instances! You can just setgenerate true 8, or whatever you want!
PpRyWVBJ2xnw1dium4qdcQPnX3YfZCqZas
-
Compile a 64-bit version on Linux, and you don't need multiple instances! You can just setgenerate true 8, or whatever you want!
PpRyWVBJ2xnw1dium4qdcQPnX3YfZCqZas
Have you compared the hashpermin on a 4CPU vs 8CPU? When I tested it using 64 bit linux, it showed the same hashrate, even when you let it decide the # of processes (-1). I was under the impression that running multiple instances may allow you to use more groups of 4 cores if your machine has more. I could be wrong though.
-
The wallet is the same for all instances
I tried the 64bit on a 128core machine. Here is the results
genproclimit: 96
hashespermin: 10.45023078
Separating them out across multiple directories (genproclimit 3) 11.253 x 24
-
kenpodragon,
Can you explain please, how is the wallet the same for all instances? I figured there would be a new wallet in each data directory. Or is it because you duplicated the initial one, that they are still the same?
Does that mean you can use commands like "./protoshared getbalance" without specifying a datadir and it will show the results for anything mined from all instances?