[XPM]unofficial jhPrimeminer thread - page 7.

dbasql

full member

Activity: 219

Merit: 100

Ethics and Science need to shake hands

Quote from: Tamis on December 11, 2013, 09:36:53 AM

What's the best miner for ypool to run on ubuntu 64bit servers right now ?

For 12.04 from my personal notes..

apt-get update

apt-get dist-upgrade - (answer yes)

apt-get install build-essential libssl-dev openssl git libgmp-dev

git clone https://github.com/tandyuk/jhPrimeminer.git

cd jhPrimeminer

make

Tamis

sr. member

Activity: 476

Merit: 250

What's the best miner for ypool to run on ubuntu 64bit servers right now ?

LeftToeCut

full member

Activity: 154

Merit: 100

Quote from: magnificat_mafia on December 10, 2013, 07:48:57 AM

Quote from: LeftToeCut on December 10, 2013, 07:43:41 AM

Does anyone have a working OSX binary and could upload it somewhere? I've got a quite a few macs here but am unable to compile.

What is the problem you are having? I was able to compile on osx

time mostly, I got xcode and a macports but got lots of errors about something missing. I am not entirely sure what went wrong but if you could upload a binary I would appreciate it.

magnificat_mafia

newbie

Activity: 36

Merit: 0

Quote from: LeftToeCut on December 10, 2013, 07:43:41 AM

Does anyone have a working OSX binary and could upload it somewhere? I've got a quite a few macs here but am unable to compile.

What is the problem you are having? I was able to compile on osx

LeftToeCut

full member

Activity: 154

Merit: 100

Does anyone have a working OSX binary and could upload it somewhere? I've got a quite a few macs here but am unable to compile.

ssl303

newbie

Activity: 46

Merit: 0

I have the following output. Is there way to predict how long it takes for me to mine a block looking at the output? How do you calculate and what number do you use?

New Block: 303100 - Diff: 9.992898 / 9.992898
Best/Max diff: [ 9.913507 / 0.000000 ]
6ch/h: 79.24 - 2077 [ 604 / 636 / 837 ]
7ch/h: 8.81 - 231 [ 70 / 79 / 82 ]
8ch/h: 0.88 - 23 [ 6 / 7 / 10 ]
9ch/h: 0.11 - 3 [ 1 / 0 / 2 ]
MNPS:423.47 PPS:35823 WPS:13.837 ACC:2588
Current Primorial: 61 - Sieve Size: 1974500 - Prime Count: 42389

dbasql

full member

Activity: 219

Merit: 100

Ethics and Science need to shake hands

Quote from: mhps on December 09, 2013, 01:32:14 AM

Quote from: dbasql on December 08, 2013, 10:58:35 PM

Quote from: John122 on December 07, 2013, 12:57:38 PM

Quote from: dbasql on December 06, 2013, 11:35:14 PM

Anyone with any input? I have access to many different types of processors and will gladly post the results.

You could try adding "-target 9 -bttarget 9" to your config, I've seen an increase with it (well, I think so, it's kinda hard to tell with this coin since there's no "hashrate" per se).
I think it focuses your miner on finding 9-chains instead of 10 since you find a much greater deal of those than 10-chains, and they are still worth a lot.
As for the multiple -m, not sure about this, I've not seen many people use that config before.

Thanks a lot for the reply.
To be clear are you saying there is no real way to measure how fast a processor is compared to another processor?
I was looking here and it lists some different CPU's but I don't know how to see the output.
http://xpmwiki.com/index.php?title=Hardware_comparison
In other words is it possible that I am using two XEON 5639 with 24 threads and it could actual not perform as well as a new generation XEON with less threads. I have access to all kinds of processors and can post info I just know how to look for that info and what part of it is useful.
I can of course look at this information with the standard XPM wallet but not with JHPrimeminer. Once I start the batch file with these parameters I don't know how to gauge performance.

chain/day makes assumption and different miners use slightly different difinition of cpd. Even with the same cpd you would get different return on different pools because their reward policies are different. The best parameter is how many XPMs you make perdat on average.

Some of the pools reward heavily the block finder. So in order to get a fair comparison for long term mining speed you would need to wait your miner to find a few blocks before drawing conclusion. That takes time unless you have several of the same machines to bundle together. Then you have parameter space to explore for each miner software.

You need to be methodical and careful to do it right.

First of thanks for the reply
Ok this makes sense I am using Ypool. I can segment an account I guess and see how long it takes to get 1 XPM for each type of CPU as a baseline. It would at least provide a rough estimate.
Right now I can say one server with Dual 5639 I get one XPM per 24 hour period. I think this is pretty decent but I don't have anything else to compare it to.

mhps

hero member

Activity: 516

Merit: 500

CAT.EX Exchange

Quote from: dbasql on December 08, 2013, 10:58:35 PM

Quote from: John122 on December 07, 2013, 12:57:38 PM

Quote from: dbasql on December 06, 2013, 11:35:14 PM

Anyone with any input? I have access to many different types of processors and will gladly post the results.

You could try adding "-target 9 -bttarget 9" to your config, I've seen an increase with it (well, I think so, it's kinda hard to tell with this coin since there's no "hashrate" per se).
I think it focuses your miner on finding 9-chains instead of 10 since you find a much greater deal of those than 10-chains, and they are still worth a lot.
As for the multiple -m, not sure about this, I've not seen many people use that config before.

Thanks a lot for the reply.
To be clear are you saying there is no real way to measure how fast a processor is compared to another processor?
I was looking here and it lists some different CPU's but I don't know how to see the output.
http://xpmwiki.com/index.php?title=Hardware_comparison
In other words is it possible that I am using two XEON 5639 with 24 threads and it could actual not perform as well as a new generation XEON with less threads. I have access to all kinds of processors and can post info I just know how to look for that info and what part of it is useful.
I can of course look at this information with the standard XPM wallet but not with JHPrimeminer. Once I start the batch file with these parameters I don't know how to gauge performance.

chain/day makes assumption and different miners use slightly different difinition of cpd. Even with the same cpd you would get different return on different pools because their reward policies are different. The best parameter is how many XPMs you make perdat on average.

Some of the pools reward heavily the block finder. So in order to get a fair comparison for long term mining speed you would need to wait your miner to find a few blocks before drawing conclusion. That takes time unless you have several of the same machines to bundle together. Then you have parameter space to explore for each miner software.

You need to be methodical and careful to do it right.

dbasql

full member

Activity: 219

Merit: 100

Ethics and Science need to shake hands

Quote from: John122 on December 07, 2013, 12:57:38 PM

Quote from: dbasql on December 06, 2013, 11:35:14 PM

Anyone with any input? I have access to many different types of processors and will gladly post the results.

You could try adding "-target 9 -bttarget 9" to your config, I've seen an increase with it (well, I think so, it's kinda hard to tell with this coin since there's no "hashrate" per se).
I think it focuses your miner on finding 9-chains instead of 10 since you find a much greater deal of those than 10-chains, and they are still worth a lot.
As for the multiple -m, not sure about this, I've not seen many people use that config before.

Thanks a lot for the reply.
To be clear are you saying there is no real way to measure how fast a processor is compared to another processor?
I was looking here and it lists some different CPU's but I don't know how to see the output.
http://xpmwiki.com/index.php?title=Hardware_comparison
In other words is it possible that I am using two XEON 5639 with 24 threads and it could actual not perform as well as a new generation XEON with less threads. I have access to all kinds of processors and can post info I just know how to look for that info and what part of it is useful.
I can of course look at this information with the standard XPM wallet but not with JHPrimeminer. Once I start the batch file with these parameters I don't know how to gauge performance.

John122

full member

Activity: 129

Merit: 100

Quote from: PenAndPaper on December 07, 2013, 02:04:58 PM

The main problem is that you can't tell by the live stats when your miner does better. Undecided

It seems that primes/ sec aren't a reliable measure.

Indeed, "share value per hour" is better, but you find a few 9-chain or a 10-chain and it jumps quite higher, so while it's better it's still unreliable for performance measurement.

PenAndPaper

sr. member

Activity: 252

Merit: 250

The main problem is that you can't tell by the live stats when your miner does better. Undecided

It seems that primes/ sec aren't a reliable measure.

John122

full member

Activity: 129

Merit: 100

Quote from: dbasql on December 06, 2013, 11:35:14 PM

Anyone with any input? I have access to many different types of processors and will gladly post the results.

You could try adding "-target 9 -bttarget 9" to your config, I've seen an increase with it (well, I think so, it's kinda hard to tell with this coin since there's no "hashrate" per se).
I think it focuses your miner on finding 9-chains instead of 10 since you find a much greater deal of those than 10-chains, and they are still worth a lot.
As for the multiple -m, not sure about this, I've not seen many people use that config before.

dbasql

full member

Activity: 219

Merit: 100

Ethics and Science need to shake hands

Quote from: dbasql on December 05, 2013, 08:25:44 PM

I know this is a dumb question but I have to ask. Huh

How do you guys see the stats such as PPS etc when you have started JH from a batch file? Are you writing this out to a log file?
Also anyone know if this is a strong config for Xeon 24 thread CPU? I will post the rate I am mining at if I can figure out how to do it.

jhPrimeminer-T16.exe -u worker -p pass -m 53 -m2 59 -m3 61 -m4 67 -o ypool.net:8083

Thanks DBA

Anyone with any input? I have access to many different types of processors and will gladly post the results.

dbasql

full member

Activity: 219

Merit: 100

Ethics and Science need to shake hands

I know this is a dumb question but I have to ask. Huh

How do you guys see the stats such as PPS etc when you have started JH from a batch file? Are you writing this out to a log file?
Also anyone know if this is a strong config for Xeon 24 thread CPU? I will post the rate I am mining at if I can figure out how to do it.

jhPrimeminer-T16.exe -u worker -p pass -m 53 -m2 59 -m3 61 -m4 67 -o ypool.net:8083

Thanks DBA

PenAndPaper

sr. member

Activity: 252

Merit: 250

Quote from: dustofdeath on December 03, 2013, 06:55:09 PM

any ideas abotu the jhprimeminer "has stopped working" error?

trying to mine at ypool but it just instantly crashes.

"jhPrimeMiner.exe -o http://ypool.net -xpt -u condemned.XPM_1 -p x"

The string above should be correct?

Check the details of the error message and see if a file mpir.dll is mentioned.
Probably you have the wrong mpir.dll install for your cpu.

jet1460

newbie

Activity: 5

Merit: 0

The Mumus Links are not working....

dustofdeath

member

Activity: 98

Merit: 10

any ideas abotu the jhprimeminer "has stopped working" error?

trying to mine at ypool but it just instantly crashes.

"jhPrimeMiner.exe -o http://ypool.net -xpt -u condemned.XPM_1 -p x"

The string above should be correct?

PenAndPaper

sr. member

Activity: 252

Merit: 250

Quote from: mhps on November 30, 2013, 07:43:29 PM

Quote from: PenAndPaper on November 30, 2013, 06:45:20 AM

Quote from: mhps on November 30, 2013, 03:06:36 AM

Those number aren't proven very useful. they are indicators of different steps of the mining process.
The Chain/day is more useful in telling how quick you could find 9-chains, a fraction (i.e. 1-difficulty) of which will be blocks. The value/h tells how this miner has been doing given the current reward policy -- higher means more reward from the pool.

Isn't the pps the main speed indicator? I remember a hardware comparison list and i think it was sorted based on pps. Not sure though.

No. It was a long time ago most people used pps, when no one understood the mining process much except Sunny King. these days cpd is used but different miners use slightly different definition of cpd.
the only reliable indicator is how many XPM you mine perday, after mining for some days when many a few blocks are found. It can take weeks to do so. this number is tied to the pool performance if you are not soloing.

I mean is it possible for a miner that mines with 5k pps to actually mine more coins from a miner that mines with 10k pps?

mhps

hero member

Activity: 516

Merit: 500

CAT.EX Exchange

Quote from: John122 on December 02, 2013, 09:58:30 PM

Actually, it seems to do that everytime you find a block (10-chain). Well, it does for me.
If you look at your pool account, you should see it as a "Found block" and receive the income accordingly.

I didn''t see a new block found on the pool side.
20 9-chains and no 10-chain. Is it just bad luch or the command line makes the miner only look for 9 chains?

muto

sr. member

Activity: 457

Merit: 250

I found several 10 chains and one 11 chain and it showed up fine. So no problem at my machines

Topic: [XPM]unofficial jhPrimeminer thread - page 7. (Read 180178 times)