Open Source XPM Pool + GPU Miner (aka. madPrimeMiner)

[ap1]

I found out how to compile the cl when I checked the code too
But maybe Linux driver issue or so, I cannot compile the kernel.

Using device 0 as GPU 0
Using device 1 as GPU 1
Compiling ...
Source: 65657 bytes
binsize = 427808 bytes
OpenCL error: -30 at ../xpmclient.cpp:852

Actually I tried eXtremal's original open source miner too and never compile right neither.
He said AMD SDK is buggy and I used ATI driver compiler but still not successful.

I've found subtle solution if you have more than 1 card.

Disable all BUT 1 card in config.txt. Run xpmminer. kernel.bin is created. Shut down miner. Undo changes in config. Run with normal config.
Works for me.

Thanks! So it is what the forum is for, exchange solutions

[MarcusDe]

I found out how to compile the cl when I checked the code too
But maybe Linux driver issue or so, I cannot compile the kernel.

Using device 0 as GPU 0
Using device 1 as GPU 1
Compiling ...
Source: 65657 bytes
binsize = 427808 bytes
OpenCL error: -30 at ../xpmclient.cpp:852

Actually I tried eXtremal's original open source miner too and never compile right neither.
He said AMD SDK is buggy and I used ATI driver compiler but still not successful.

I've found subtle solution if you have more than 1 card.

Disable all BUT 1 card in config.txt. Run xpmminer. kernel.bin is created. Shut down miner. Undo changes in config. Run with normal config.
Works for me.

Thanks! So it is what the forum is for, exchange solutions

Agree :-)
Now, maybe your eagle eye would find why there is no CPD showing for multiple cards on pool? :-P

[djm34]

can someone release a dtc poolserver for solomining (for windows) ? Thanks.

[ap1]

Agree :-)
Now, maybe your eagle eye would find why there is no CPD showing for multiple cards on pool? :-P

I do usually debug pretty quickly. I don't have the source code for DTC pool... the original madmax xpm one seems working fine I think?

[MarcusDe]

Agree :-)
Now, maybe your eagle eye would find why there is no CPD showing for multiple cards on pool? :-P

I do usually debug pretty quickly. I don't have the source code for DTC pool... the original madmax xpm one seems working fine I think?

@ap1:
Yes, I started with xpm pool source, it didnt work there too.

@djm34:
I'll try in time... but for me priority is to get full working code with working cpd and other stats.
Besides, gpool.net has monopoly with datacoin, check blocks found by my small pool (6 workers + mine 1) and that pool.
I mine DTC from 1st day with some (long) breaks solo with CPU, now finally there is possibility for healthy competition and maybe price will go up to the moon when XPM reaches diff 11.

[ap1]

@ap1:
Yes, I started with xpm pool source, it didnt work there too.

@djm34:
I'll try in time... but for me priority is to get full working code with working cpd and other stats.
Besides, gpool.net has monopoly with datacoin, check blocks found by my small pool (6 workers + mine 1) and that pool.
I mine DTC from 1st day with some (long) breaks solo with CPU, now finally there is possibility for healthy competition and maybe price will go up to the moon when XPM reaches diff 11.

One thing I don't want to mine XPM anymore is the price of XPM keep dropping... basically halved within a month or two?
Seems must be something to do with GPU mining...

[madMAx510510]

If you change TARGET in .cl files, make new kernel.bin then your CPD in stats also will change.

edit:

270X settings for DTC:

Code:

#define SIZE 4096			// Size of local sieve array (must be multiple of 1024)
#define LSIZE 256			// Local worksize for sieve (don't change this)
#define STRIPES 400			// Number of sieve segments (you may tune this)
#define WIDTH 20			// Sieve width (includes extensions, you may tune this)
#define PCOUNT 38400		// Number of sieve primes (must be multiple of 256, you should tune this, 40960 is for 280X)
#define SCOUNT PCOUNT		// Don't change it
#define TARGET 9			// Target for s_sieve

If you change target to 9 you need to use the latest *.cl files from git, 6 days ago. Otherwise you have bugs...

Also you may have to reduce STRIPES to maybe 150, otherwise the miner crashes because the sieve now returns 3 times more candidates.

[Primorial]

11-ch share values on http://89.179.126.88:8081/ appear to be disproportionately low!

Last 12 blocks found by the pool are all 11+ chains, still 11-shares are less than 2x more valuable than 10-shares, while they should be at least 20x more valuable?

Code:

Current Share Values
 9  0.04601 XPM
10  0.53773 XPM
11  0.93356 XPM

24h Share Count
 9   765       (~ 35.19765 XPM)
10    75       (~ 40.32975 XPM)
11     9       (~  8.40204 XPM)

Diff. 10.96499747 

[Remember remember the 5th of November]

So let me get this straight? You integrate the pool into the XPM source code thus exposing the pool to pretty much direct access?  I know one can disable access other from localhost, but this pretty much requires the XPM node to run alongside the pool, rather than on separate servers. Meaning if you DDoS the pool, you DDoS the node or vice-versa.

[LongAndShort]

So let me get this straight? You integrate the pool into the XPM source code thus exposing the pool to pretty much direct access?  I know one can disable access other from localhost, but this pretty much requires the XPM node to run alongside the pool, rather than on separate servers. Meaning if you DDoS the pool, you DDoS the node or vice-versa.

troll? (yes i'm asking) you should ask the same on every pool code thread! Name me one pool software that does not run alongside the node

[MarcusDe]

There is possibility to set web frontend on 1st server, pool software on 2nd server and node on 3rd server.
But who cares if even such big pools with big money going thru them as BTCGuild weren't ddos resistant.

[SpeedDemon13]

There is possibility to set web frontend on 1st server, pool software on 2nd server and node on 3rd server.
But who cares if even such big pools with big money going thru them as BTCGuild weren't ddos resistant.

The DTC pool doesn't show the miner stats, but it does show the balance.

[MarcusDe]

Agree :-)
Now, maybe your eagle eye would find why there is no CPD showing for multiple cards on pool? :-P

I do usually debug pretty quickly. I don't have the source code for DTC pool... the original madmax xpm one seems working fine I think?

That ADL trick doesnt work for me, still error -3, compiled under linux.

[MarcusDe]

For your information...
I made new thread for datacoin pool, https://bitcointalk.org/index.php?topic=636017.new#new

[djm34]

I had a go to madpriminer on dtc pool (didn't check with xpm) but the temp of the card are pretty high
(the card is in a well ventilated room). It is right now at 89°C (it is a R9 290x) getting 177cpd

Is this normal ? (powertune at 0). It never went that high even in scrypt/scrypt-n and neither with the other xpm miner.

[eXtremal]

eXtremal, any updates to the miner?

I remember about miner.. and have one idea about Fermat test optimization, but no enough time to implement it I will post some results at this week.

[csa1234]

is this miner faster than Claymore? I mean not by the stats, but for real, how many blocks per day an R9 280x can produce with this miner? Cause with Claymore an R9 280x can only mine 3 blocks per day average

PS: Solomining

[MarcusDe]

is this miner faster than Claymore? I mean not by the stats, but for real, how many blocks per day an R9 280x can produce with this miner? Cause with Claymore an R9 280x can only mine 3 blocks per day average

PS: Solomining

I dont know about blocks, but its about 4.0-4.5 10-ch shares on my 280X.

[primer10]

eXtremal, any updates to the miner?

I remember about miner.. and have one idea about Fermat test optimization, but no enough time to implement it I will post some results at this week.

+1 

[kev7112001]

is this miner faster than Claymore? I mean not by the stats, but for real, how many blocks per day an R9 280x can produce with this miner? Cause with Claymore an R9 280x can only mine 3 blocks per day average

PS: Solomining

claymores can get 6.6 up to 7.2 cpd on a 7970/280 depending on cpu and clocks