[XPM]unofficial jhPrimeminer thread

[ikilled]

Sorry,
the one in the background is mining slower than the one in the foreground.

So, for example if i run jhprimeminer in foreground in the terminal:
./jhprimeminer -o http://... -u user.1 -p pass
it will mine faster than if I run it in the background (with the & at the end):
./jhprimeminer -o http://... -u user.1 -p pass >> jhprimeminer.log &

Maybe I am wrong...
I'm basing my findings by looking at the jhprimeminer's output when running in the foreground (in the ssh client window while logged in)
and  running in the background (I redirect the output to a file >> jhprimeminer.log) and logging out of shell.

One weird thing I notices is that this log file "jhprimeminer.log" is updating very slowly.

[cabin]

can someone message me with a list of all the miner / locations where to get it so i can update the front page 

I updated my build with the latest share values:
11ch   100
10ch   80
9ch   16
8ch   1
7ch   0.03125
6ch   0.000976

The builds I know about are:

Mumus:
V7.1 64bit - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v7.1.zip
v7.1 64bit AVX - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v7.1-AVX.zip
v6.0 32bit - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v6-32bit.zip

Cabin
It is quite similar to the Mumus build but does not do the large divisor test anymore since others have said this part is buggy and wastes a bit of time. It also has the -h parameter which specifies what length of chain to search for (default 9 which is the case for all current miners) and it removes some unused fermat test branches and other calculations.
64 bit - https://www.dropbox.com/s/szayx15skhevv7x/jhPrimeminer.zip

rdebourbon
This is an optimized version of Mumus 7.1 build with my sieve updates and a few other performance increases. Block rate should be as good as / better than Mumus. Use same as you would Mumus V7.1
64 bit - https://www.dropbox.com/sh/zij7j205ucsmj3n/s3NKCVo8mL/jhPrimeMiner_RdBBeta3.zip

[jh00]

I updated my build with the latest share values:
11ch   100
10ch   80
9ch   16
8ch   1
7ch   0.03125
6ch   0.000976

I am using the opportunity to ask everyone working on an unofficial build of jhPrimeminer to avoid having hardcoded share values. The server does send the share value (when a share is accepted) to the miner for a few weeks now. Take a look at this commit: https://github.com/jh000/jhPrimeminer/commit/3c93c1e68d05ebf41893306d179e0ccb869f8741
The share value can be retrieved in this function:
xptClient_processPacket_shareAck(xptClient_t* xptClient);

I might need to do a few more changes to the share values, maybe even make them dynamic. Having them hardcoded only adds confusion on the user side whenever the values are changed.

[x3maniac]

can someone message me with a list of all the miner / locations where to get it so i can update the front page 

I updated my build with the latest share values:
11ch   100
10ch   80
9ch   16
8ch   1
7ch   0.03125
6ch   0.000976

The builds I know about are:

Mumus:
V7.1 64bit - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v7.1.zip
v7.1 64bit AVX - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v7.1-AVX.zip
v6.0 32bit - https://dl.dropboxusercontent.com/u/1189851/jhPrimeminer-GMP-v6-32bit.zip

Cabin
It is quite similar to the Mumus build but does not do the large divisor test anymore since others have said this part is buggy and wastes a bit of time. It also has the -h parameter which specifies what length of chain to search for (default 9 which is the case for all current miners) and it removes some unused fermat test branches and other calculations.
64 bit - https://www.dropbox.com/s/szayx15skhevv7x/jhPrimeminer.zip

rdebourbon
This is an optimized version of Mumus 7.1 build with my sieve updates and a few other performance increases. Block rate should be as good as / better than Mumus. Use same as you would Mumus V7.1
64 bit - https://www.dropbox.com/sh/zij7j205ucsmj3n/s3NKCVo8mL/jhPrimeMiner_RdBBeta3.zip

updated thanks for the info!

[mhps]

Cabin
It is quite similar to the Mumus build but does not do the large divisor test anymore since others have said this part is buggy and wastes a bit of time. It also has the -h parameter which specifies what length of chain to search for (default 9 which is the case for all current miners) and it removes some unused fermat test branches and other calculations.
64 bit - https://www.dropbox.com/s/szayx15skhevv7x/jhPrimeminer.zip

rdebourbon
This is an optimized version of Mumus 7.1 build with my sieve updates and a few other performance increases. Block rate should be as good as / better than Mumus. Use same as you would Mumus V7.1
64 bit - https://www.dropbox.com/sh/zij7j205ucsmj3n/s3NKCVo8mL/jhPrimeMiner_RdBBeta3.zip

Thanks. I am trying your jhPrimeMiner_RdBBeta3.zip above. It still uses the old values for the chains found (i.e. 0.1 for 6chain, 1 for 7chain etc.).

[samos123]

@Cabin: Your build doesn't work under wine for me.

[clintar]

Sorry,
the one in the background is mining slower than the one in the foreground.

So, for example if i run jhprimeminer in foreground in the terminal:
./jhprimeminer -o http://... -u user.1 -p pass
it will mine faster than if I run it in the background (with the & at the end):
./jhprimeminer -o http://... -u user.1 -p pass >> jhprimeminer.log &

Maybe I am wrong...
I'm basing my findings by looking at the jhprimeminer's output when running in the foreground (in the ssh client window while logged in)
and  running in the background (I redirect the output to a file >> jhprimeminer.log) and logging out of shell.

One weird thing I notices is that this log file "jhprimeminer.log" is updating very slowly.

It probably has something to do with the way it's outputting. I get weird characters on screen in my client, so it's probably having a hard time interacting with the terminal. You can try using screen instead of the &

[clintar]

@Cabin: Your build doesn't work under wine for me.

yeah, not working for me, either. Sounds like the one I'd want to try. Wish you guys were compiling on linux. I might have to break down and try my hand at trying to get some of your changes into the repo that's compiling on linux now. (that will take a long time I'm sure)

[TheXev]

I started using rdebourbon's beta 3 yesterday under Windows, and the performance on my A10-4600m is threw the roof now:

Code:

Val/h:21.572 - PPS:10778 - SPS:22.929 - ACC:473
 Chain/Hr:  4:  2228.14  5:   191.39  6:    18.50  7:     1.64  8:     0.00 9:     0.18 10:     0.00 11:     0.00 12:     0.00

Code:

Total/Valid shares: [ 207 / 201 ]  -  Max diff: 9.84485
 6ch/h:    18.55 - 203 [ 48 / 39 / 116 ] - Val: 20.300
 7ch/h:     1.64 - 18 [ 3 / 2 / 13 ] - Val: 18.000
 8ch/h:     0.00 - 0 [ 0 / 0 / 0 ] - Val: 0.000
 9ch/h:     0.18 - 2 [ 1 / 1 / 0 ] - Val: 200.000
Share Value submitted - Last Block/Total: 0.000000 / 32.708984

Code:

Number of mining threads: 4
Sieve Size: 1000000
Sieve Percentage: 20
Round Sieve Percentage: 70
Prime Limit: 400000
Primorial Multiplier: 53
L1CacheElements: 128000
Total Runtime: 0 Days, 10 Hours, 57 minutes, 19 seconds
Total Share Value submitted to the Pool: 32.70898

Taking to rdebourbon and AeroCloud was a huge help.  rde's build is great because once you know how to tune, you know what to look for, and those new values of SPS and ACC is a huge help.

[clintar]

I started using rdebourbon's beta 3 yesterday under Windows, and the performance on my A10-4600m is threw the roof now:

Code:

Val/h:21.572 - PPS:10778 - SPS:22.929 - ACC:473
 Chain/Hr:  4:  2228.14  5:   191.39  6:    18.50  7:     1.64  8:     0.00 9:     0.18 10:     0.00 11:     0.00 12:     0.00

Code:

Total/Valid shares: [ 207 / 201 ]  -  Max diff: 9.84485
 6ch/h:    18.55 - 203 [ 48 / 39 / 116 ] - Val: 20.300
 7ch/h:     1.64 - 18 [ 3 / 2 / 13 ] - Val: 18.000
 8ch/h:     0.00 - 0 [ 0 / 0 / 0 ] - Val: 0.000
 9ch/h:     0.18 - 2 [ 1 / 1 / 0 ] - Val: 200.000
Share Value submitted - Last Block/Total: 0.000000 / 32.708984

Code:

Number of mining threads: 4
Sieve Size: 1000000
Sieve Percentage: 20
Round Sieve Percentage: 70
Prime Limit: 400000
Primorial Multiplier: 53
L1CacheElements: 128000
Total Runtime: 0 Days, 10 Hours, 57 minutes, 19 seconds
Total Share Value submitted to the Pool: 32.70898

Taking to rdebourbon and AeroCloud was a huge help.  rde's build is great because once you know how to tune, you know what to look for, and those new values of SPS and ACC is a huge help.

I saw you got a block earlier. Wonder if there is a guide anywhere, yet for targetting higher chain shares.

[CoinBuzz]

how could i run it under linux with wine?

please make 32bit version too.

[rdebourbon]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

[tadakaluri]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

Thank you very much for your work.

[tadakaluri]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

Thank you very much for your work.

WARNING - Avast Says this file have VIRUS

[rdebourbon]

WARNING - Avast Says this file have VIRUS

Not sure why - but both Chrome & Avast keep flagging my builds.. Can you send the virus report so I can analyse.

[TheXev]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

Thank you very much for your work.

WARNING - Avast Says this file have VIRUS

Avast flags everything as a virus.  I really have no clue why anyone uses it.  flags cgminer, now rde's build... prob cause some goofball ran it on a corporate network and another admin picked it out.

[CoinBuzz]

please someone tell me how can i run these miners on 32bit linux?

[ikilled]

Sorry,
the one in the background is mining slower than the one in the foreground.

So, for example if i run jhprimeminer in foreground in the terminal:
./jhprimeminer -o http://... -u user.1 -p pass
it will mine faster than if I run it in the background (with the & at the end):
./jhprimeminer -o http://... -u user.1 -p pass >> jhprimeminer.log &

Maybe I am wrong...
I'm basing my findings by looking at the jhprimeminer's output when running in the foreground (in the ssh client window while logged in)
and  running in the background (I redirect the output to a file >> jhprimeminer.log) and logging out of shell.

One weird thing I notices is that this log file "jhprimeminer.log" is updating very slowly.

It probably has something to do with the way it's outputting. I get weird characters on screen in my client, so it's probably having a hard time interacting with the terminal. You can try using screen instead of the &

Thanks, yep probably is because of the way how I am redirecting the output.

Now i forcibly disconnected the ssh sessinon (without quitting jhprimeminer and exit-ing) , leaving the jhprimeminer running.
I found out how to read output from any process running (from another ssh session) - with this command:
http://unix.stackexchange.com/questions/58550/how-to-view-the-output-of-a-running-process-in-another-bash-session
Now I can see that the pps is normal - as it I would run it in the foreground.

Thanks everybody for the help :-)

[FoldingTime]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

I would really like to get this running under native linux. Can you provide the source so we can try to compile it under linux?

Thanks.

[rdebourbon]

I've posted a new beta v3.1 that removes the hard coded share values. Share value now only reported on block summary.

I've also created a single dropbox folder for all my releases now - so only a single link to check:
https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM

No performance difference to beta v3 only the display update.

I would really like to get this running under native linux. Can you provide the source so we can try to compile it under linux?

Thanks.

Source is forked on github:
https://github.com/rdebourbon/jhPrimeminer