So are these good speeds for an I5-3570k running at stock? What measures should I be using to compare to see if I am doing a bad job or not? Anywhere I can look to compare various processors, for example mine vs an FX8350?
Topic: [XPM]unofficial jhPrimeminer thread - page 10. (Read 180178 times)
So are these good speeds for an I5-3570k running at stock? What measures should I be using to compare to see if I am doing a bad job or not? Anywhere I can look to compare various processors, for example mine vs an FX8350?
Code:
jhPrimeMiner.exe -o http://ypool.net:10034 -u workername -p workerpass -m 61 -target 9 -bttarget 9
Thanks I am trying it.
OK 13 hours later not only the miner is still working but also that the 7-ch/h number has come back to about 70% of peak performance, the difference could well be mostly variance.
Mining on ypool is an option for me again. Thanks. Sending some xpm your way as a small token.
Awesome! Glad we got to the bottom of it in the end! Thanks for the awesome donation anything and everything always welcome!
Is there any chance that versions from rdebourbon and tandyuk get merged so both windows and linux version get updated at the same time?
Yes, actually.. This is on the cards.. tandy now has write access to my repository on github, so I expect that as we go forward the two forks will merge into one..
Is there any chance that versions from rdebourbon and tandyuk get merged so both windows and linux version get updated at the same time?
Code:
jhPrimeMiner.exe -o http://ypool.net:10034 -u workername -p workerpass -m 61 -target 9 -bttarget 9
Thanks I am trying it.
OK 13 hours later not only the miner is still working but also that the 7-ch/h number has come back to about 70% of peak performance, the difference could well be mostly variance.
Mining on ypool is an option for me again. Thanks. Sending some xpm your way as a small token.
As with previous releases of my miner, tandyuk is working on the port. Based on progress he made yesterday I'm hoping there will be some news later today..
Any news on Tandy's linux build?
So are these good speeds for an I5-3570k running at stock? What measures should I be using to compare to see if I am doing a bad job or not? Anywhere I can look to compare various processors, for example mine vs an FX8350?
Code:
jhPrimeMiner.exe -o http://ypool.net:10034 -u workername -p workerpass -m 61 -target 9 -bttarget 9
Thanks I am trying it.
OK it's been 9 hours and the 4.0-F2 is still alive.
However I am getting less than 30% of 7-chains per hour. I used to get 3.8-4.5 7-chains/h on this i5 3210M laptop with 4 cores. Now I am getting 1.0. This performance problem started since early October on v3.3 and v3.2 (without any upgrading on the miner side). I assumed it was cause by intermitten disconnction. I have sent you the FoundShares.log. Maybe there is some clue in it.
Great news that it stays connected now..
Advice for increasing shares:
1. Try another primorial: The default is 43 - which works for some. I reduced it from my initial default of 61 due to some older CPUs struggling with the Fermat Testing for the larger primorial - but as you're on an i5 - I'd set it back to 61 via -m. Your MNPS will drop somewhat, and it will need to retune the sievesize/prime count for the new primorial, so dont start with a -s or -d as its likely to be less than optimal.
2. The default at the moment is to retarget to 10, as we're above a predetermined threshold where some early maths suggested the chance of finding a suitable 9ch was less than a 10ch.. Personally I seem to be finding more blocks on target 10 - but it does hurt share counts. If you want a better val.hr reset target & bttarget back to 9. Mumus was experimenting with different target/bttarget values and did get some interesting results - so a split approach may be worth trying: target 10 bttarget 9 or vice versa.
(NOTE: for those reading this that may be solo mining, I'd still recommend 10 over 9.. but YMMV!)
So in summary, try launching the miner with this syntax for better val/hr on current difficulty:
Code:
jhPrimeMiner.exe -o http://ypool.net:10034 -u workername -p workerpass -m 61 -target 9 -bttarget 9
(Edited for typos)
What do all those numbers mean?
When I run jhprimeminer, I get something like this.
11/07/13-05:42:39 - New Block: 248520 - Diff: 9.965746 / 9.965746
Best/Max diff: [ 8.946473 / 6.731931 ]
6ch/h: 39.46 - 463 [ 151 / 137 / 175 ]
7ch/h: 4.35 - 51 [ 15 / 22 / 14 ]
8ch/h: 0.26 - 3 [ 1 / 1 / 1 ]
MNPS:669.75 PPS:22581 WPS:35.732 ACC:631
Current Primorial: 43 - Sieve Size: 1171500 - Prime Count: 13649
What numbers matter the most?
Which number would I be focusing to improve when changing some values?
OK it's been 9 hours and the 4.0-F2 is still alive.
However I am getting less than 30% of 7-chains per hour. I used to get 3.8-4.5 7-chains/h on this i5 3210M laptop with 4 cores. Now I am getting 1.0. This performance problem started since early October on v3.3 and v3.2 (without any upgrading on the miner side). I assumed it was cause by intermitten disconnction. I have sent you the FoundShares.log. Maybe there is some clue in it.
Great news that it stays connected now..
Advice for increasing shares:
1. Try another primorial: The default is 43 - which works for some. I reduced it from my initial default of 61 due to some older CPUs struggling with the Fermat Testing for the larger primorial - but as you're on an i5 - I'd set it back to 61 via -m. Your MNPS will drop somewhat, and it will need to retune the sievesize/prime count for the new primorial, so dont start with a -s or -d as its likely to be less than optimal.
2. The default at the moment is to retarget to 10, as we're above a predetermined threshold where some early maths suggested the chance of finding a suitable 9ch was less than a 10ch.. Personally I seem to be finding more blocks on target 10 - but it does hurt share counts. If you want a better val.hr reset target & bttarget back to 9. Mumus was experimenting with different target/bttarget values and did get some interesting results - so a split approach may be worth trying: target 10 bttarget 9 or vice versa.
(NOTE: for those reading this that may be solo mining, I'd still recommend 10 over 9.. but YMMV!)
So in summary, try launching the miner with this syntax for better val/hr on current difficulty:
Code:
jhPrimeMiner.exe -o http://ypool.net:10034 -u workername -p workerpass -m 61 -target 9 -bttarget 9
(Edited for typos)
As with previous releases of my miner, tandyuk is working on the port. Based on progress he made yesterday I'm hoping there will be some news later today..
Great news. Go Tandy! I can't wait for 4.0 linux build.
I'm still using 3.3 on my linux machines. If somebody wants to use it until 4.0 is ready please search this thread for instructions.
I upgraded my Windows machines to 4.0 and it feels like it's noticably faster. I didn't compare stats, but my impression is that 4.0 finds definitely more coins.
Thank you rde and Tandy for great work!
I'll try post a notification here when the build is available, but in case I don't get a chance to do so please keep an eye on the dropbox.. I'm hopeful it will provide the fix to your networking issue..
Version 4.0F-2 of my miner is being uploaded to the dropbox now. The only difference is the addition of the new XPT ping implementation that should help resolve NAT issues.
Dropbox address is: https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM
Wow that was quick! Thank you. I am starting it (the non-avx version). I will come back in 8-9 hours and see how it goes.
OK it's been 9 hours and the 4.0-F2 is still alive.
However I am getting less than 30% of 7-chains per hour. I used to get 3.8-4.5 7-chains/h on this i5 3210M laptop with 4 cores. Now I am getting 1.0. This performance problem started since early October on v3.3 and v3.2 (without any upgrading on the miner side). I assumed it was cause by intermitten disconnction. I have sent you the FoundShares.log. Maybe there is some clue in it.
That's awesome! I assume tandyuk's release will also have -xpm option to solo as well? I can't wait to try it!
Yes it will have the -xpm option.
Can jhprimeminer just solo mining without the server generating? Do I need to have gen=1 in primecoin.conf on my server in order to solo mine with jhprimmeminer?
Yes it means you can solo mine without the server generating. No need for gen=1, but you will need server=1 & rpc options.
Lastly, if I have few computers mining for my server, is there way for me to look at the data as a group as well as individually? How fast they're generating as a group, and so on?
Thanks again! I'm so excited to be able to run jhprimeminer with -xpm feature on ubuntu!
Thanks again! I'm so excited to be able to run jhprimeminer with -xpm feature on ubuntu!
Currently no way to get collective stats on the group. There is potential for this being added though via tandyuks ongoing stats reporting work.
I'll try post a notification here when the build is available, but in case I don't get a chance to do so please keep an eye on the dropbox.. I'm hopeful it will provide the fix to your networking issue..
Version 4.0F-2 of my miner is being uploaded to the dropbox now. The only difference is the addition of the new XPT ping implementation that should help resolve NAT issues.
Dropbox address is: https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM
Wow that was quick! Thank you. I am starting it (the non-avx version). I will come back in 8-9 hours and see how it goes.
Can this run on Linux? Also, can -xpm "" deposit to the default "" address?
Currently there is no linux build of the beta, but the port is close to completion..
The -xpm argument requires an address at the moment.
The parameter for -xpm is coin address, not account name right?
Also, any estimate on at least beta release date for ubuntu?
Thanks,
That is correct, it's the XPM wallet address you wish to mine into.
As with previous releases of my miner, tandyuk is working on the port. Based on progress he made yesterday I'm hoping there will be some news later today..
That's awesome! I assume tandyuk's release will also have -xpm option to solo as well? I can't wait to try it!
Can jhprimeminer just solo mining without the server generating? Do I need to have gen=1 in primecoin.conf on my server in order to solo mine with jhprimmeminer?
Lastly, if I have few computers mining for my server, is there way for me to look at the data as a group as well as individually? How fast they're generating as a group, and so on?
Thanks again! I'm so excited to be able to run jhprimeminer with -xpm feature on ubuntu!
I'll try post a notification here when the build is available, but in case I don't get a chance to do so please keep an eye on the dropbox.. I'm hopeful it will provide the fix to your networking issue..
Version 4.0F-2 of my miner is being uploaded to the dropbox now. The only difference is the addition of the new XPT ping implementation that should help resolve NAT issues.
Dropbox address is: https://www.dropbox.com/sh/sq24hzo993afy9c/l7icP0KiuM
Can this run on Linux? Also, can -xpm "" deposit to the default "" address?
Currently there is no linux build of the beta, but the port is close to completion..
The -xpm argument requires an address at the moment.
The parameter for -xpm is coin address, not account name right?
Also, any estimate on at least beta release date for ubuntu?
Thanks,
That is correct, it's the XPM wallet address you wish to mine into.
As with previous releases of my miner, tandyuk is working on the port. Based on progress he made yesterday I'm hoping there will be some news later today..
Can this run on Linux? Also, can -xpm "" deposit to the default "" address?
Currently there is no linux build of the beta, but the port is close to completion..
The -xpm argument requires an address at the moment.
The parameter for -xpm is coin address, not account name right?
Also, any estimate on at least beta release date for ubuntu?
Thanks,
I am trying an old version of your miner modified by cabin (installed on Aug 22). It appears to work like before. I will run it for some hours and see the result.
OK about 12 hours since I started the miner .. Just now I find the miner printing out many identical
Code:
Sieve/Test ratio: 0.0 / 0.0% - New PrimorialMultiplier: 2
lines.From ypool block history the miner went into that state about three hours ago. I am impressed that it kept alive for 9 hours.
Your strange problem came up in the yPool dev chat last night, and someone behind multiple layers of NAT (even ISP side) identified the issue as the same one they recently encountered! The fix for him was to implement the XPT ping function - which he kindly provided a source change for.. I'm in the process of reviewing the update, and once complete I will post a new build to my dropbox.
I'll try post a notification here when the build is available, but in case I don't get a chance to do so please keep an eye on the dropbox.. I'm hopeful it will provide the fix to your networking issue..
I am trying an old version of your miner modified by cabin (installed on Aug 22). It appears to work like before. I will run it for some hours and see the result.
OK about 12 hours since I started the miner .. Just now I find the miner printing out many identical
Code:
Sieve/Test ratio: 0.0 / 0.0% - New PrimorialMultiplier: 2
lines.From ypool block history the miner went into that state about three hours ago. I am impressed that it kept alive for 9 hours.
Jump to: