[XPM]unofficial jhPrimeminer thread

[nlmocvn]

Good evening guys, well lol i'm a newbie and I need some help with this.

I'm using Intel 3930K @ 4.4GHZ for XMP mining and I wonder if my following results using jhPrimeminer are fine or need to be tweaked?

http://i1293.photobucket.com/albums/b595/Kozmos81/FADY-PC/2014-02-11_19-52-09_zps3bc583b1.png



Plus, how to know my daily XPM production rate from that image?

Hey kozmos,

as I already stated in my previous post, the production rate of the next higher chain level is reduced by factor 10-15 - based on personal observations. I'd expect you to have an average output rate of ~3 Shares/h => 72 shares/day => ~0,6 XPM per day atm, when running your miner continously.

Tips on tweaking I can't provide, to be honest I wasn't able to establish a general pattern for setting up mining machines til now. Usually the default setup using primoral 61 and Sievesize 1M-1,5M always worked best for me.

Hello admiral and thx a lot for your answer I will read your comments in this topic so I can understand the way of how XPM is working

can you help me with my scheenshot? How much XPM i can get per day?
http://i1364.photobucket.com/albums/r734/babykiss1707/primecoin_zpsfd234f11.jpg

[rdebourbon]

I'm porting to Linux at the moment and should have a release soon.  Source for v4f is on my github.

Just a quick update I've got an initial port of v4H running on a Linux virtual machine. . Obviously this is not an ideal benchmarking environment but the speed per thread seems to be at least on a par with my native windows miner on same physical machine. . Chain rates are as expected and I've even been very lucky and found a couple of 9ch and a 10ch.. For some this may seem ordinary,  but my dev pc has a normal 9ch rate of 1/day. .

Any way I've got a tiny bit of cosmetic clean up to do and will be putting out binaries for more public testing soon. I'm currently using gmp 5, so there may be some easy  performance gains to be realised by switching to gmp 6..That's next after the resolution of the auto tuner. .

Happy mining!

[primer10]

I'm porting to Linux at the moment and should have a release soon.  Source for v4f is on my github.

Just a quick update I've got an initial port of v4H running on a Linux virtual machine. . Obviously this is not an ideal benchmarking environment but the speed per thread seems to be at least on a par with my native windows miner on same physical machine. . Chain rates are as expected and I've even been very lucky and found a couple of 9ch and a 10ch.. For some this may seem ordinary,  but my dev pc has a normal 9ch rate of 1/day. .

Any way I've got a tiny bit of cosmetic clean up to do and will be putting out binaries for more public testing soon. I'm currently using gmp 5, so there may be some easy  performance gains to be realised by switching to gmp 6..That's next after the resolution of the auto tuner. .

Happy mining!

That's great! Can't wait to try it out, thanks 

[rdebourbon]

I've posted jhPrimeMiner_RdBBeta4.0I_Linux.tar.gz to my mega.co.nz folder https://mega.co.nz/#F!FA9jVJoC!avbe9C3WMdlroM8gbJG7Ew

This tar.gz archive has a few different versions of the miner compiled and ready for testing - use the one appropriate for your CPU.. As with V4H for Windows, this is still a BETA binary release only. Source code will be posted to Github in the next week or so depending on my schedule..

Please let me know if you have any issues with the miner!

Happy Mining!

[rdebourbon]

Progress report 31March2014.
1 bug found and fixed in Linux port.  New build v4Ib available in mega folder.

Bug only affected pool mining and was an issue with reconnecting when pool servers restarted.

Thanks to all the beta testers so far.

Happy mining!

[rdebourbon]

New BETA v4.0J (Win & Linux)
======================

I've posted a new build of my miner this morning to the mega.co.nz folder. This is mainly an update for linux compatibility issues, but does include a cache optimisation that yields a neat little boost in raw sieve speed (~5-10%).

I will be posting a trial build of v4J with GMP6 statically linked later today, but this is very much a trial build.. I have not run extensive tests with the GMP6 - although on my test PC it does appear to work as expected.

Happy Mining!

[antonio8]

@rdebourbon

I am new to your miner and don't know if this has been posted but do you have a sample bat file to help me configure?

I don't see a readme in your miner.

[rdebourbon]

@rdebourbon

I am new to your miner and don't know if this has been posted but do you have a sample bat file to help me configure?

I don't see a readme in your miner.

At a minimum you'd have:
JhPrimeMiner.exe -o http://ypool.net:10034  -u yourworker -p yourpassword

If you're using the computer for other things you may want to reduce the thread count. To do this add -t x where x is less than total cores in computer.

If Wps is very high > 100 consider a bigger sieve size.  Default size is approximately 2 Mill.  If Wps < 30 consider smaller sieve size.  Try keep sieve Size a multiple of 16384. Change sieve size with -s xxxxx

Finally use as high a  primorial as gives a decent mnps. I recommend 61 its been very lucky for me.  Change primorial with -m xx

That's it really. .

[antonio8]

@rdebourbon

I am new to your miner and don't know if this has been posted but do you have a sample bat file to help me configure?

I don't see a readme in your miner.

At a minimum you'd have:
JhPrimeMiner.exe -o http://ypool.net:10034  -u yourworker -p yourpassword

If you're using the computer for other things you may want to reduce the thread count. To do this add -t x where x is less than total cores in computer.

If Wps is very high > 100 consider a bigger sieve size.  Default size is approximately 2 Mill.  If Wps < 30 consider smaller sieve size.  Try keep sieve Size a multiple of 16384. Change sieve size with -s xxxxx

Finally use as high a  primorial as gives a decent mnps. I recommend 61 its been very lucky for me.  Change primorial with -m xx

That's it really. .

Thanks

[24core]

Hi there,

I have set the sieve to 16 Million (16384000) to get my WPS down and this works for the first stat.

However, it seems to reset and my WPS jumps back to 177.

Has anyone else seen this problem ?

[rudn]

Hi there,

I've seen nearly the same problem.

with sieve size 16384000 the smallest WPS i can reach is 211.
(Dual Prozessor Xenon E5 with HT enabled, 32 Logical Prozessors!)

[COOLBGT]

Hi there,

I've seen nearly the same problem.

with sieve size 16384000 the smallest WPS i can reach is 211.
(Dual Prozessor Xenon E5 with HT enabled, 32 Logical Prozessors!)

With the new updates you should no longer aim for WPS 20. Just use the default sieve and increase it if you wish. There is also an auto-tune option enabled by default which will run for like 4 hours

[berbip]

Hi guys,

New xpm  miner here

And a bit confused !

I use a Core i5 2500K processor for mining, do i use jhPrimeminer_RdBBeta4.0J_Win_SSE_Test2 or jhPrimeminer_RdBBeta4.0J_Win_AVX_Test2 for my processor ?

Is there anything specific i need to add tot my batch file ?

Thanks,

[COOLBGT]

Hi guys,

New xpm  miner here

And a bit confused !

I use a Core i5 2500K processor for mining, do i use jhPrimeminer_RdBBeta4.0J_Win_SSE_Test2 or jhPrimeminer_RdBBeta4.0J_Win_AVX_Test2 for my processor ?

Is there anything specific i need to add tot my batch file ?

Thanks,

You should definitely use the AVX version. As for the .bat file, after the .exe you should have something like -o http://relay.ypool.net -u yourworkernae -p x -m 61 -t numbersofthreads

[Megacoins]

Hello. I have i7 4930k. What command to disconnect one core from mining?

[UltraShock97]

Link don't work!! Help me!

[kyndik]

Hey All,

Just a quick update to let you know I've uploaded a new beta version 4.0G to the new mega.co.nz home for my miner (after dropbox suspended my public folder).

New mega.co.nz home:
https://mega.co.nz/#F!FA9jVJoC!avbe9C3WMdlroM8gbJG7Ew

This new version should be between 10-15% faster than 4.0F and adds support for the upcoming mining version 2 (-miningVersion argument). In simple terms the new mining version requires the header hash to be a probable prime..

I'll be posting updated source to GitHub later.

Cheers!

rdebourbon, Hi!
Could you please explain why AMD FX-4300 @ 4.3GHz (-t 4, MNPS=280) slower then AMD Athlon II 280 @ 3.6Ghz (-t 2, MNPS=297). Tried on jhPrimeminer_RdBBeta4.0J_x64 for Athlon;
jhPrimeminer_RdBBeta4.0J_x64-AVX and jhPrimeminer_RdBBeta4.0J_x64 for FX-4300
OS: Fresh Windows 8.1 x64

Thank you in advance.

[rdebourbon]

rdebourbon, Hi!
Could you please explain why AMD FX-4300 @ 4.3GHz (-t 4, MNPS=280) slower then AMD Athlon II 280 @ 3.6Ghz (-t 2, MNPS=297). Tried on jhPrimeminer_RdBBeta4.0J_x64 for Athlon;
jhPrimeminer_RdBBeta4.0J_x64-AVX and jhPrimeminer_RdBBeta4.0J_x64 for FX-4300
OS: Fresh Windows 8.1 x64

Thank you in advance.

Without doing more exhaustive tests I couldn't say for certain, but I suspect the issue will be on-die cache differences within the CPUs..

A quick look at CPU-World (I didn't go any further so data could be wrong) indicates that despite the higher core count, the L1 data cache is smaller (half the size!) on the FX-4300 compared to the Athlon II..

FX-4300:
2 x 64 KB 2-way set associative shared instruction caches
4 x 16 KB 4-way set associative data caches

Athlon II 280:
2 x 64 KB 2-way set associative instruction caches
2 x 64 KB 2-way set associative data caches

If this is true, it will result in a higher L1 miss ratio - which could account for the performance difference. Having said all that, I'd be a bit surprised if that was the sole reason for a such a drastic difference in performance..

[ngdias]

Hi,

I'm testing this using my laptop with an Intel CORE i7 vPro, whatever that means. I'm using jhPrimeminer_RdB v4.0J (Linux), the AVX version.
I tried a few settings, and I read the optimization instructions, but I'm not sure I understood everything correctly. These were some of my results:

MNPS:998.96   PPS:14106        WPS:143.802  ACC:98     
Current Primorial: 43 - Sieve Size: 2097152 - Prime Count: Variable

MNPS:899.89   PPS:12706        WPS:129.540  ACC:98     
Current Primorial: 43 - Sieve Size: 262144 - Prime Count: Variable

MNPS:899.89   PPS:12706        WPS:129.540  ACC:98     
Current Primorial: 43 - Sieve Size: 262144 - Prime Count: Variable

MNPS:917.90   PPS:21042        WPS:132.134  ACC:159     
Current Primorial: 61 - Sieve Size: 2597152 - Prime Count: Variable

MNPS:947.95   PPS:13448        WPS:136.458  ACC:98     
Current Primorial: 43 - Sieve Size: 597152 - Prime Count: Variable

MNPS:784.21   PPS:18014        WPS:112.889  ACC:159     
Current Primorial: 61 - Sieve Size: 297152 - Prime Count: Variable

MNPS:776.80   PPS:18027        WPS:111.821  ACC:161     
Current Primorial: 61 - Sieve Size: 262144 - Prime Count: Variable

MNPS:1011.78  PPS:14306        WPS:145.648  ACC:98     
Current Primorial: 43 - Sieve Size: 2097152 - Prime Count: Variable

This last setting is just running the program on defaults. The idea is to maximize the MNPS, right? So far it seems that no tweaking yields the best results. Is this correct?
I'd appreciate suggestions for improvement

How to calculate potential average yield from solo mining using the client stats?
Also, how do you compare the performance of this client to the official wallet mining ability?

Thanks.

[CampBxSucks]

i have some coins on the site, can anyone help me? how to contact the ypool.net owners? I only have support@ypool.net