Bitcoin Forum

Maybe you know me from my other thread where i launched a new XPM pool + GPU miner:
https://bitcointalk.org/index.php?topic=598542.0

If yes, you also know that it failed because of DOS attacks and other general server problems.

I've now decided to go open source, with the pool server and the GPU miner.
Now everyone with the expertise can create an XPM pool and also create new miners that work with the same protocol.
Also anyone with basic bitcoin code knowledge can make a solo-miner out of it. EDIT: It's now available.

The pool server code is located at: https://github.com/madMAx43v3r/xpmpool
And the GPU miner code is at: https://github.com/madMAx43v3r/xpmclient
UPDATE: The solo-mining server code is at: https://github.com/madMAx43v3r/xpmserver

The solo-mining server binaries can be downloaded at:
For Windows: https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI (https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI)
For Linux: https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k (https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k)
Note: The windows version crashes on shutdown, have not been able to fix this.

EDIT: If you have dynamic ip: There is a bug somewhere in the bitcoin network code that can cause the server to loose all connections to the network after your external ip has changed. It will still show 8 or more connections but if you look in debug.log you will no longer see new blocks or transactions. In this case you need to restart the server.
The chance of this bug to happen is around 50%. So be careful, if your external ip changes and you don't check the server and the bug has happend you will produce 100% orphan blocks.

Please use this thread to discuss and/or ask questions about the code.

If you like my work you can donate XPM to: ATHHPkcWjhUSqXpgJhTut7jzWo3JEbNj9J

I think you just revolutionised XPM mining... Congrats!   :D

It is really very dramatic and unbelievable for your move. I asked myself and I believe I couldn't do the same if I spent months for making a software and let go to open source so easily.

I am a cheapskate and never believe in the donation model, but to show my appreciation I will donate all the coins that were mined during the testing of your server yesterday and 10% of the coins mined in the first month (at least).

+1 :)

Wow mate you are really something else you know that. Good move you definitely belong in this space in my opinion, good luck and i'm donating 280xpm cause that was a bloddy good move.
You are a real man! and you are wise because this is how you breath life into xpm possibly...wait till mike gets a hold of this code :)

Thanks man, have received the 280 XPM.

Who is mike?

Mad props madmax  ;D

He optimized the first yacoin gpu miner kernal and moved on to seriously improve the xpm client making all the hp clients So he will no doubt be tinkering with this code (is my hopes) if and when it needs to be optimized for the diff rises

now we need a miner setup like cgminer so we can solo or pool mine and set backups

would be cool if he made a solo release so we can use it now like cgminer with a donation address attached to pay for improvements

If in 2 days nobody has done so (with open source), i will do it.

great would be cool if u made it cgminer like where we can set temps clocks etc

Well i'm using cgminer ADL code anyways, so shouldn't be that hard to enable those features.

nice thanks

Some peeps are gonna be mad to this, end of theft xpm mafia

cmon guys don't be cheap. Give the guy a donation and encourage more people to do this kind of thing thats how it works!

You sir are a gentleman and a scholar.Seriously,this move deserves a standing ovation

http://gifatron.com/wp-content/uploads/2013/05/applauding.gif

You can be proud of your self making the best miner, even more for this decision. 30 XPM donated.

Thanks, have received it.

If you still produce a miner, you can leave a donation % feature in it.

Great that you embody what cryto currency envisioned in decentralization. Truly respectful and greatly appreciated.

xpmpool repo has just received a fix, the Makefile for leveldb was incorrectly ignored and thus was missing in the repo.

EDIT: I've also updated the README.md with more info.

i would donate if the pool was up and all these ypool coin wreckers weren't harassing him for the code

another coin killed by ypool

he just made it open sourse i don't get what you are saying mate

Congratulation for releasing the open source.
Are you planning to implement the usual protocols: getwork, xpt and BlockTemplate ?

Congrats on releasing the code! Looking forward to the changes this brings to primecoin.

Just realized how important this is.
Will start working on solo miner first thing tomorrow. It will be a stripped down version of the pool server, without database, without website, but with Qt GUI.
And you can connect the existing pool miner to it just like it was a pool server.
I'll try to push for a same day release, but no gurantees.

日期: 2014/5/10 11:44
到: ATHHPkcWjhUSqXpgJhTut7jzWo3JEbNj9J
支出: -30.00 XPM
交易费: -0.01 XPM
净额: -30.01 XPM
交易ID: 87e2a09092ead9adb16a5a5f541fdd7b724e8ad88e7d6fd5c17194e6e7f37cf1


donated!

Well done madMax!!

cool, now i can try to make a better xpm gpu pool miner


Suck it primeGPU and Claymore! Xpm mafia

Simply BRAVO! You made my faith in crypto foes up again! :-)

Thanks a lot for this open source XPM GPU Miner! This is what the crypto scene needs !

Fermat test 4 times (about 400%) slower than open source xpmclminer implementation, I will try merge miners code, 8-9 CPD on Radeon R9 290X is really :)

Oh so it is faster than madmax? which meant the following: claymore > primeGPU ? xpmclminer > madPrimeMiner ?

any eta when we start test something?

Mining algorithm consists of two parts - sieve and Fermat test. madPrimeMiner have very fast sieve and slow non-optimized Fermat test, xpmclminer have too poor sieve, but fast Fermat test. I will merge two miners code for create fastest XPM miner.

Bravo, cant wait! Suck it, claymore!

That's a very piece of good news!!

We would need this once diff crosses into 11..

Just no ypool please :)

Claymore probably could make use of it to improve his miner. Good and bad for xpm. Bad since his is only ypool only. Since Beeeeer now has dev fee, he should consider producing one for Beeeeer

Awesome work madMAx, thx for you efforts

Built with system libwt version 3.3.1, or 3.3.2 is minimal required?
Client built sucessfully.

Thanks for making it open source and sharing it with the community . Has anybody compiled binaries for windows ?

+1 need for win7 :)

o.O its already compiled for windows. Click the link to the other thread and read it..Aint no one got time for that, right?

Doesn't really answer the question..
How do you compile it ?

the question was has anyone compiled windows version and i assume they think its just a miner so like i said its already compiled for windows

Built success on linux (pool and client):
Now try replace a Fermat test function

Well done.
I'm excited to see what the performance will be with the replaced fermat. :)

Original:

With replaced Fermat:

My code uses 384-bit operands, need to create 352-bit version and tune sieve parameters.

This is on a 290X correct?

Yes.
I changed weave depth and got this results:
Fermat tests number increased, CPD - no.. need more time to found right configuration. I will release something after 7CPD reaching on R9 290X.

Great job man. Will your work also increase results on a 280X or are you specifically tuning it for 290 cards?

/usr/local/include/czmq_prelude.h:498:13: error: conflicting declaration ‘typede
f int SOCKET’
 typedef int SOCKET;
             ^
In file included from netbase.h:11:0,
                 from util.h:33,
                 from bignum.h:12,
                 from main.h:10,
                 from main.hpp:13,
                 from primeserver.h:13,
                 from pool.h:13,
                 from pool.cpp:10:
compat.h:29:15: error: ‘SOCKET’ has a previous declaration as ‘typedef u_int SOC
KET’
 typedef u_int SOCKET;
Any ideas?

+1 to the qn.

Btw, I like your signature :)

I make temporary patch:

working on a windows compilation for miner and pool

PS: My pool fee will be 50%  ;)

the question was has anyone compiled windows version and i assume they think its just a miner so like i said its already compiled for windows
[/quote]

working on a windows compilation for miner and pool

PS: My pool fee will be 50%  ;)
[/quote]

better 100%)

Compilation done but i have segfault on :


Someone can help me ?

im just kidding, of course it will be free

selfish bastard...

ok, i wont compile, get another guy to do it, have a nice day. Bye

man please dont go! we need you  ;D

ohh :(

you can do it :) gogogadget O compil :)

Guys, this is not a trollbox please stay close to the topic.

Patch: https://www.dropbox.com/s/ukfszdh4y17ndkr/xpmminer_patch.tar.gz
Before (one gpu crashed, it's unstable):
after:
5xRadeon R9 290 (non-X), primorial 12. It's not final result :)

is this in usable form for ypool or beeeer or solo?  or am I in a hurry?

Only private pool mining available, you must build pool and client from source.

There will be a cold day in hell before I manage to do that.....   :D

Or another variant, I can send address of my pool and you join it :)

Oh I thought you were gonna change xpmclminer instead of madMax's miner... looks like I got it wrong ..

In this case, is it not good enough to use madmax's compiled miner since only .cl file is changed? In other words, no need to build client?

Interested.
Do you have a windows build for the miner?

Miner is here https://bitcointalk.org/index.php?topic=598542.0
Patch: https://www.dropbox.com/s/ukfszdh4y17ndkr/xpmminer_patch.tar.gz (replace kernel.bin)
For private pool address - send PM (payment after 11XPM mined, if I right).

Claymore already take code of sieve algorithm and release new version with 7.0CPD on Radeon R9 290X :) This miner with patch from xpmclminer shows only 6.6CPD now.

I will change xpmclminer for speedup and supporting madmax's pool. It's more good because xpmclminer support Radeon HD5xxx/6xxx and NVidia GPUs.

Is this only linux or windows too?

:)
Both..

Replacing the old kernel.bin to your version does not work for 280X card :(

Yeah, looks like my fermat test wasn't top of the line  ;D

Was working on the solo miner, but since Claymore is now faster, don't need to bother really.

Let's see if eXtremal can match claymores speed.

Output ?

Pool stats:
107 workers? Hmm...

Don't stop working on solo miner, please. We need solution for solo mining.
I personally can't stand ypool. And claymore miner uses 5x more cpu than yours.

OpenCL error: -11 at ..\xpmclient.cpp:888

I don't think claymore's faster, they measure the chains differently probably.

Also, fug ypool, mafiapool, they started being shady from day one.

Try with latest drivers, I use catalyst 14.3 for create this binary.

+1

Waiting for solo miner...

Drivers should be 14.4 or 14.3?
How to aply the patch?

thanks for the xpm but u didnt have to give it back

is this project dead like the pool ?

lol the miner isnt dead or the pool

+ 100

Peeps, if you want to make use of existing madmax's client with the improved fermat.cl, you can do the following:

1) mkdir ../gpu (while in the madmax dir where xpmclient is located)
2) copy the new fermat.cl, together with the sha256.cl and sieve.cl into the gpu dir (sha256 and sieve .cl inside xpmclient gpu src)
3) if you are using Catalyst 13.12, you need Catalyst 14's libamdocl64.so. replace or load it using LD_LIBRARY_PATH

Btw, once .cl is compiled into .bin, you do not need to stick with the above steps, including the need to install new Catalyst drivers

I have kernel.bin for 280x and 7950, if anyone is interested.

CREDIT: to eXtremal for improving the fermat.cl, the use of Catalyst 14's libamdocl64.so, among other things :) Thanks!

Thanks, but how to use this?

@eXtremal

Would it be possible to make changes to xpmclminer instead? I find it easier to mod xpmclminer for pool mining than madmax's.

Is it just coincidental, or did Claymore just grab this open source miner, recompile it, and post it on his version 8.0 miner? The CPD is right about what madMax promised(4.8-4.9 CPD at 1150); Caymore's miner was previously running about 2.33 CPD (version 7.0).

I love the fact that madMax open sourced his miner, and I do plan to donate, but if someone else is going to be charging for his work, I think he should give it an extra days worth of work and post the miner up for a pool (with or without a fee). Just a thought.

As much as I hate to see Windows-only, ypool-only, 10% dev fee (dev fee is fine, 10% is too much. Also, Claymore and primeGPU, have not kept to their words to releasing their source when 2 GPU miners were open sourced. Now why should ppl be held ransom to these 2 buggers?) xpm mafia devs, I do not think it is easy to just grab the code, recompile and post as his 8.0 miner. My suspicion is that Claymore did not release his fastest miner outright at the start.

EDIT: Now remember this impt point, Claymore (so did primeGPU) protected his miner from casual reverse engineering using vmprotect, it is trivial to just "dumb down" his code and you would be non the wiser ..

He might not have released his fastest miner, but Claymore specialized in optimizing his miner for 290s. To get double the CPD from a 280x using his miner, just a day after this open source miner was released is a little fishy. I wouldn't mind if he improved on it and charged 5% -- but 10%?

I have a hunch that he releases the latest version after the open source version of madMAx, then took what ever code in that and merged it with his. That is just my speculation on it.

Damn ...

If you look at v7 and below, Claymore has always used K<num>.dat where the .dat file is the compiled OpenCL bin. If you see v8, it is replaced by a Data.bin that looks encrypted... Something to hide ?? What do you guys think??

Claymore and ypool, shady stuff.

The only way to compete is to make the compilation of the source more easily by preparing ready-to-go various makefile/project files for VC++, mingw, cmake, etc.

It would be necessary to have compiled binaries (better with MD5 checksum) making available most to non-programmer miners.

May be some low-rate pools are needed too. Otherwise, people still go to ypool after successful compilation.

You are right. Claymore has replied in his thread. There's gonna be shady ppl and shady business anywhere. Ppl just have to decide what's best for them.

Very interested if it's compiled on 13.12 :)
Can you please post a link.

Here you go: http://we.tl/oLAkmn6ss8

ok so someone tell me how to use this miner with w7x64 ?

I'm with you. I'd like to test this miner but I don't know anything about coding / compiling for win7x64. GPU: r9 280x and r9 270x Catalyst 13.12 preferred

anyway lots of thanks to everyone for all the efforts to get mad miner running!! awesome :)

I think its appalling that just about 4 of us donated to this guy yet he brought the faster miner and open sourced it with pool software cmon..Most of you complain about the thugs but then you are tightnesses when its actually the right thing to do and donate...tsk tsk  you get robbed everyday by your governments and people around you then your're dubious and lash out at the wrong people here in forums and your homes for petty stuff yet do nothing about the real causes. Its messed up cause many of you don't get how this works even though its not rocket science...You don't donate and support people open sourcing then there is no incentive to do it then you have to pay through the nose for everything!!!!!!! then you cry again whining about rubbish rinse and repeat!

Thanks, that worked :)
What's your donation address ?

give me the same stuff you are smoking

Nah, its fine. Donate to either MadMax or eXtremal.

I encourage you to throw up an addy mate..we really do need to create or bring back this tipping culture again imo


This may look pretty deep but its the "them against us" mentality that divides us in my opinion

Tyvm good sir for that kernel file.
I'm trying to run miner on 13.12 from yesterday and it fails :-P

I will surely do.

Can eXtremal post his donation address please.
EDIT: Just realised I'm mining at his private pool and I have his E-mail, duh...
Still it would be wise for him to post his donation address so other people can send him some credits for his work.

(I have Max in my PM so no problems there)

you are absolutely right. I donated once to eXtremal (I am not sure if this eXtremal is the same guy as the one from peercointalk but very likely he is :) ) though it wasn't much. I have yet to donate to madMax, which I am going to now.

pool down?

Yeah thats great i mean every little bit counts as we all know as mienrs :) .1 .001 xpm seems like nothing but if everyone starts this culture of tipping again for contributions then it all adds up and is highly encouraging anyhow i don't wanna hijack the thread i'm just appalled all the same

Pool not stable (you must know :) ), today autorestarting script would be created.

Any pools / wallets I can connect with this miner?

btw, 1st coin earned will be donated to creator or miner :-D

4.62 CPD on 7970HD on 14.4 !  :P

Good job =)

i'm getting 4.16CPD on HD7970 (1100/1500) with catalyst 13.12. Which frequencies has your cards?

obj/pool.o: In function `PrimeWorker::Work(_zctx_t*, void*)':
/home/miner5/Downloads/xpmpool-master/primeserver/src/pool.cpp:100: undefined reference to `zsocket_set_sndhwm'
obj/backend.o: In function `PoolBackend::ThreadLoop(_zctx_t*, void*)':
/home/miner5/Downloads/xpmpool-master/primeserver/src/backend.cpp:115: undefined reference to `zsocket_set_rcvhwm'
collect2: error: ld returned 1 exit status
make: *** [primecoind] Error 1


another error, please help,thanks.

Hopefully someone can tune this miner to make the older AMD/ATI cards run better, ie. 5800's and 6900's....

Compile zeromq & czmq.

./webserver --docroot /usr/local/share/Wt --http-address 0.0.0.0 --http-port 80
terminate called after throwing an instance of 'Wt::Dbo::backend::PostgresException'
  what():  Could not connect to: FATAL:  password authentication failed for user "webserver"
FATAL:  password authentication failed for user "webserver"

help on this ,thanks.

Hi,

I am new to Unix. Could somebody post guide how to compile client on Unix (e.g. Ubuntu). I would really like to mine xpm. Standard autogen/configure/make does not work here.

+1

Yes, can someone compile it and throw it up here? Also some instructions wont hurt...
Many thanks!

why do u need to compile the client? 1st post already has the compiled binary. If you encountered error due to the OpenCL binary (kernel.bin), you just need to put the .cl in proper place, run the xpmclient to let it compile.

More likely you want to compile the pool? Wayla has been trying to compile, perhaps he can release some guide or compiled binary once he is done; I have not gone down that track yet.

I am trying to get the server figured out, but would prefer a solo version with no fees.  Is anyone working on this?

madmax was working on this .. not sure whether he is continuing his work, I hope he does!

if you get your own server running, you do not have to worry about the fees :)

all default compilation work if u have build all dependencies...

please anyone , can tell me how to compile and run this client , i'm pretty new , could you help me please ?

madmax pool version for Beeeeer.org is been developing... soon....

who is the dev? madmax?

+1

some issue,need help

It looks to me like the database or user rights for Postgres DB not set up properly?

Temporary patch for use with 16384 open files linit (ulimit -n 16384): https://www.dropbox.com/s/74etmfwx6h3ye8a/diff.txt
With this patch my private pool can handle more than 350 workers without crashes :)
I'll send PM to MadMax with more details.

How often pool pays ? I have allready 55 xpm in requested.

XPM needs 2 days to mature.

Pool can't pay now, all mined coins unconfirmed.. after 20-24 hours pool can access first mined coins.

Yes, the modular exponentiation code that was released with this miner is a good start for others to use, but very slow.

overclocking cards helps more cpd? also it says balance:1.6xpm  requested:1xpm. Whats that requested xpm mean?

Yes, my 3xRadeon R9 290X @ 1040/1250MHz (memory overclocking has no effect unlike scrypt mining):

1 xpm ready to payout.. payouts starts after two days mining.

I'm wrong about 400% (miner shows 500-600k Fermat tests per second, but miner do some other work, not only Fermat tests :) ), really 100-120%.

Yesterday was Mother's Day-very busy with family, but I was able to find about 15 minutes to convert the Pool Sever into a Solo Pool Server (with no webserver or database support).

Benchmark results (Solo Pool Server Mode)
Benchmark duration: 60 minutes

GPU used: SAPPHIRE DUAL-X R9 280X 3GB GDDR5 OC
http://www.sapphiretech.com/presentation/product/product_index.aspx?pid=2022&lid=1

Modular Exponentiation: Used eXtremal's patch

[GPU 0] T=-1C A=-1% E=0 primes=0.098621 fermat=499649/sec cpd=3.76/day
(ST/INV/DUP): 152x 7ch(0/0/0) 15x 8ch(0/0/0) 1x 9ch(0/0/0)
GPU 0 found share: 8-ch type 3
Share accepted.
Work received: height=538045 diff=10.918976 latency=0ms
GPU 0 found share: 7-ch type 1
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 0 found share: 8-ch type 3
Share accepted.
GPU 0 found share: 7-ch type 1
Share accepted.
GPU 0 found share: 7-ch type 1
Share accepted.
[GPU 0] T=-1C A=-1% E=0 primes=0.098627 fermat=500023/sec cpd=3.76/day
(ST/INV/DUP): 156x 7ch(0/0/0) 17x 8ch(0/0/0) 1x 9ch(0/0/0)

Conclusion: The sieve code is very fast, but the modular exponentiation code needs some improvement. Should be able to achieve an estimated 5 CPD with a faster modular exponentiation implementation on an R9 280X (GTX 580 - 4.2 CPD)

Happy belated Mother's Day and Happy Mining!

which driver should i use to get more cpd 13.12-14.3 or 14.4 with 290 cards?

right now the fastest solo version have Claymore which implemented part of madmax code onto his GPU miner. Achieving 5.28 with R9 280x

4 blocks per day with 6 x R9 280x

In the meanwhile the owner of beeeeer is working to allow madmax to mine on his pool, im still in conversation with him. Stay tune

so what is better?

claymores miner with ypool or mad miner with eXtremal?

This, much lower cpu load, almost the same speed, open source.

I think you are better off using MadMiner with eXtremal.
Even if Claymore is a bit faster (which I don't know for sure), he is taking a 10% fee and Ypool also takes their fee.
eXtremal charges 5% right now, so that makes quite a difference.

On top of that I think Claymore does not deserve his 10% when he is using Max his code, so it's better not to support him.

And dunno about you, but when I run the numbers, I lose an extra 10% of my coins somewhere. Either Claymore is taking more than 10% or yool just steals.
This miner gives me a higher net.

what about

claymore's solo vs mad eXtermal?

people are talking goods about solo v8 claymore

This is funny. A totally superior miner is released that makes all others look ridiculous. Next thing you know, it's open sourced, then the authors of the slow miners grab the code, make their miners as fast if not faster and keep themselves in business. They were supposed to go out of business...

Claymore's solo still takes 10% of your earnings.

Had to CREATE ROLE in postgresql...farout blocks are flying through on the network now and diff is climbing faster then i've seen it since 6-7

claymores miner is using jhprimeminer code plus whatever also u dont have to mine with ypool i have been mining solo just fine with claymores and primegpu hitting many blocks daily

im using claymore solo mining, good results 5.20 CPD per card R9 280x even with 10% discount which is appox. 4.68 CPD which is greater than 4.30 CPD from extreme version plus -5% which makes 4.08 CPD per card

So in conclusion:

Extreme version: 4.08 CPD (discount 5% fee included)
Claymore v8 solo: 4.68 CPD (discount 10% fee included)

what clocks are you running at?

Guys I've tried claymores solo with approximate power of 68 CPD. Still no blocks found after 5 hours. I have 8 active connections. Am I doing something wrong?

good job, I will give a try with my a few cards

You can not find a single block for 23 hours and then find 24 block for 1 hour. It all depends on your luck =)

You need more than 8 connections, its all in readme file, read it.


So, guys, how can i solo with this miner? Cant see any releases... Help appreciated.

u dont need more than 8 connections to find a block u can find a block with 1 i would check to makes sure your internet is ok first then restart router maker sure your primecoin conf is set to server and u have ports open for the client

2014-05-13 03:00:36 dumpaddr thread start
2014-05-13 03:00:36 GenerateBitcoins(1)
2014-05-13 03:00:36 PoolServer started.
2014-05-13 03:00:36 PoolBackend started.
2014-05-13 03:07:08


[2014-May-13 10:56:09.183632] 2518 - [info] "config: reading Wt config file: /etc/wt/wt_config.xml (location = './webserver')"
[2014-May-13 10:56:09.183999] 2518 - [info] "WServer/wthttp: initializing built-in wthttpd"
[2014-May-13 10:56:09.184168] 2518 - [info] "wthttp: started server: http://192.168.0.102:80"
[2014-May-13 10:56:20.851730] 2518 - [info] "Wt: session created (#sessions = 1)"



primecoin server and webserver run ok， but client can't connect, please help . thanks

Did you compile an improve version of the madMAx miner? If yes, is it solo and/or pool?

Pool can be unstable now, moving to shares 9ch-11ch from 8ch-10ch.
Yes, download standard MadMax miner and patch

Patch for miner with target 11ch: <deleted> (Fermat test patch included). Current results:

Is anybody working on improving MadMax his code even further to get the same performance as claymore?

How to use this patch?
And is there a different kernel or other for 290 compared to 280x?
Didnt get my 290`s running.

Replace kernel.bin, its same for all AMD GPU's.
Catalyst 14.x required.
I'll try to optimize miner in next 7 days (I already know two optimizations, need time to implement and debug).

pool now says, 1.19cpd with3x290 is that normal?

Hello!  what about Nvidia ? Is there a win7 miner ? I have 4 gtx 750 Ti. Can i mine som XPM with those?? Thanks!

Yes, now it's normal, 3xRadeon 290X give 1,33cpd.
Now, with current difficulty each 17-18 10-number chain is block, miner searches 11-number chains only 15 times slower.
NVidia not supported now, sorry.

Are you going to test Nvidia allso ?? If yes when?  :)  Thanks!  :)

Nvidia is a whole different animal, asking that in this thread makes no sense...

Go to peercoin forums, there's an nvidia miner there, but linux only. Optimized for 750 ti too.

Thanks!  :)

Are the 10-chains found by the new 11-chain way utilized?

New patch: https://www.dropbox.com/s/brco83jy9g7mnkx/patch_11_v2.tar.gz (More optimized Fermat test, used 352-bit operations instead of 384-bit)
10ch version, 3xRadeon R9 290X 1040/1250:
7 chains/day on R9 290X !

11ch version:

I don't predict what version is more profitable for pool mining (now - may be 10ch, it produces more shares), but difficulty grows very quickly...

Yes, 11ch version also produces 10ch shares, but less than original version.

Awesome nice work can you put a donation addy in your sig please :)

at which pool are you guys mining?

which folders from the patch do I copy on madprimeminer? 10ch and GPU?

copy the kernel.bin file from either the 10ch or 11ch folder over the one in your miner folder.
You can send eXtremal a PM to ask for access to his private pool (5% fee) or you can compile the pool yourself from the opensource code.

There will be a cold day in hell before I compile this...  ;D


so how much is the minimum payout in external's pool?

First payout is at 1XPM so you can quickly verify that payouts work (after 2 days of maturing the blocks).
After that payouts will be every 11 XPM.

Extremal, can you please explain, which version of kernel.bin we should use till difficulty hit 11, 10ch or 11ch?

Mined only about 2.8 xpm in 13 hours with previous 11ch version ~1.16 cpd (~18.25cpd on new 10ch version) with previous 11ch version. this is normal?

I thought he was pretty clear he said "I'm not going to predict what is better" but for the pool the 10ch is because the 11ch will yeild less 10ch (meaning the pool wont pay as much cause you will mainly be going after 11ch meaning you will be going for blocks and not any kind of chain in between)
*holds you hand* use the 10ch

I can vouch cause i'm using the 11ch for a small pool of mine and it yeilds 8 and 9 chain almost like normal but goes after 11ch more then 10ch so there is very little 10ch but i found a block which counted as a 11ch
by the way im using 290x and the 10ch kernal proved very little improvement about .25cpd per card over the last kernal ex made

I collect statistics about 11ch version and I think 10ch remains more profitable.. until next optimized 11ch release and difficulty=10,96.
Can you post your CPD with previous and current version?

i have 68CPD power

the profit calculator says i suppose to mine 55xpm a day with current diff, and through the last day i mined about 40xpm.

11ch 32,27 times less then 10ch (must be 10-12, profit calculator also use same coefficient), original miner not ready for 11ch, I'm working on it.

I replaced kernel and tried to connect to your pool, however the program closes immediately after opening after an error code that I dont have time to read...

I tried with both kernels (ch10 & ch10) and program refuses to run

(also copied the cl files from the GPU patch folder, although i dont know if they were needed)

Is this project still open source ?

Are you using 13.12 drivers? Gotta use 14.4 drivers...

I spoke to eXtremal and he said he'll keep it open source.

madprimeminer worked well with 13.12 the new kernel does not?

do I also need to copy the cl files from the GPU files?

If someone can compile the new .cl files using 13.12 then he can create a kernel.bin that will work for 13.12 drivers.

card	10ch-old	10ch-new	11ch-old	11ch-new
7970(1100/1500)	4.17	4.76	0.31-0.32	0.31-0.32
7950(1085/1500)	3.68	4.13	0.26-0.27	0.26-0.27

May be i doing something wrong, but accoring to 10-15 min running, new and old 11ch versions have same speed.

To see the error you need to make a bat file and put this in it exactly the way it is then run it and  show us the error here also upgrade your drivers to 14.3 or 14.4 from here i guess http://www.techpowerup.com/downloads/2352/amd-catalyst-14-3-beta-1-0/mirrors


no just move the .bin file over if its windows the .cl is for compiling on liniux

11ch version can found blocks :)


It's last two blocks on pool (534730, 534747).

Same with Radeon R9 290X.. wait a next version
Claymore also prepares 11ch version :)

OpenCL compiler from 13.x has a bug, 14.x required for patched kernel.

Hi guys,

I am very confused now... So it seems maxmad will not be making solominer.

Which miner is recommended to use now? xpmminer or maxmad's xpmclient?

And what pool to use? eXtremal's private pool (5%) or should I compile maxmad's xpmpool? Which miner works for eXtremal's pool?
I don't need a detailed guide but even after compiling xpmpool's successfully there are prestsql user/role to set and not successful to initDb...

Thanks!

P.S. To MaxMad, I've just sent the 4.87XPM I mined with your server last week to you, which are all I've got in the wallet :D
I need to get the miner working to donate more  ;D

Hi,

I am new to Unix. Could somebody post guide how to compile client on Unix (e.g. Ubuntu). I would really like to mine xpm. Standard autogen/configure/make does not work here.

There is an untested binary version here https://bitcointalk.org/index.php?topic=598542.0 download it and double click

Not quite true .. this was the version that most was using when mining with madmax's pool. Even now with eXtremal's pool, that is no reason to compile the client yourself..

You need the kernel.bin that works for your GPU, that's about it.

I hope madmax is working on this: https://bitcointalk.org/index.php?topic=602292.msg6689199#msg6689199

As for those who wants it now, you could either use eXtremal's pool, compile xpmpool or get SuperComputing to share his code/binary:

How-to calculate profit from CPD with MadMax's miner.

For example 68CPD with difficulty 10,95:
XpmPerDay = (68*0,05 + 68/32,27) * (1000 / 119,9025) = (3,4 + 2,1) * 8,34 = 45,87XPM
Pool fee: 2,2935XPM
Total: 43,5765, it's like about 40xpm :)

It seems, calculator uses different coefficient, not 32,27.

Radeon R9 290X with diffuculty 10,95 mines 0,558 block/day (using 10ch version).. 11ch shows 0,47cpd + some 10ch shares (I can say how much shares, no statistics). If shares more than 2, 11ch version more profitable since diffuculty reach 10,95 :)

Thanks for the quick calculator and break down.

How to convert the Pool Sever into a Solo Pool Server -  without a web/database server.

Please make the following changes in the pool.cpp file and recompile:
.../xpmpool-master/primeserver/src/pool.cpp

Comment out the following code:

Line 30: //mBackend = 0;
Line 94: //mBackend = zsocket_new(ctx, ZMQ_DEALER);
Line 100: //zsocket_set_sndhwm(mBackend, 1*1000*1000);
Line 113: //err = zsocket_connect(mBackend, "tcp://localhost:8888");
Line 150: //zsocket_destroy(ctx, mBackend);
Line 344: //SendData(iter->second, mBackend);
Line 367: //SendData(mData, mBackend);
Line 569: //SendData(mData, mBackend);
Line 698: //mBackend = new PoolBackend(pwallet);
Line 750: //delete mBackend;

Change the following code as shown below:

Line 709: mMinShare = 7;//mBackend->Settings().MinShare;
Line 710: mTarget = 10;//mBackend->Settings().Target;

Recompile the pool server and you can now solo mine. Configure xpmclient miner with your XPM payout address (not used - key pool from your wallet.dat file will be used as usual)

Happy Mining

@SuperComputing, thanks for sharing ! Pls post your address  :)

Supercomputing, that's amazing man!

Thanks...

where is this pool.cpp file?

edit i find https://github.com/madMAx43v3r/xpmpool/blob/master/primeserver/src/pool.cpp  so after changing these what should i do?

Are you trolling? It clearly says what to do

That's the guy who actively wanted to bring this pool down with DDOS and whatnot...or at least he claimed to be.

Yes. Rather unfortunate to have such individuals in this world.

He was cheering for madMAx's pool to go down, now he wants help with the solo miner. That's irony in it self, really ungrateful.

You can't even downvote MF.

lol thats funny here is the private pool adress     ;D  :  http://89.179.126.88:8081

You are a lot weird you know that right

why dont you just join ypool   ;D  ?

Because i have my own pool! why don't you just go get some help with your mind.

You are clearly unwell and by clearly, i mean, you think you are fine but everyone can see you are not

What's your interest to publish pool address?
Yesterday I post something mining statistics, now we have another statistics, 1 rat from ~20-25 users - 4-5% rats  ;D

To all: difficulty more than 10.95, move please to 11ch version, after 24 hours accounting will be adjusted. A big part of shared produces by 10ch version are empty now.

He is clearly unwell. He is asking ppl to join ypool yet ypool's jh00 said he tried to ddos them as well and jh00 has blocked his ID.

lol i didnt ddos ypool and why should i?After many ddos atacks he bought a well protected firewall.Btw  jh00 tried to scam people before and he cried me about it.Now he tries do the same thing with xpm.If you dont belive come to ypool chat and check how many people complain:They cant get cashout their xpms for 5 days at least.I dont know what will be his excuse this time.Btw he cant block me :D

meh i'm just putting this guy on ignore and moving on i really don't need to see his dribble.

Either its a culture disconnection and or a language barrier. Whatever it is i think the guy is unstable and i really don't wanna entertain it anymore if you ignore them they usually go away

I am trying to get the server figured out, but would prefer a solo version with no fees.  Is anyone working on this?

need primeserver and webserver use the same database and user?

Same database, users can be different.

got it thank you

Btw, the Nvidia version of the XPM GPU miner has been open sourced recently @ Peercointalk. The dev is Jon Pry


It achieves quite good rates on GTX 750ti , the $150 bucks, max 60w Nvidia GPU based on Maxwell architecture. About 27 7ch/h. Not sure if we can make use of his code for further optimization.

0,47cpd for target 11 is really impressive.
Check your code, probably you still find BT chains for target 10, if so, you will get more modest cpd values :)

My code is Fermat test (I think it's a part of all XPM miners now :) ), other code created by MadMax. I found commented code in sieve.cl:539, but CPD calculate code also invalid (it works only with target 10).
What speed shows your miner on 11ch target now ?

My Fermat test implementation is not based on your work, it is faster a bit. But your miner is really good.
My miner shows about 0.3 cpd currently for diff 11 on 290x. However, my calculations show that miner for dif11 will be faster only at 10.99.
What's wrong with cpd calculation? It looks very nice and can be used for target 11 without changes with good approximation. But if you use target 10 for BT sieves you will get invalid cpd values like 0,47. For target 11 for BT you should use 6 layers C1 and 5 layers C2.

Ok, may be modular power can be more optimized. What do you think about 352-bit multiplication function?

I know about 6 layers from C1 & 5 from C2.. after full disabling BT shares, miner shows prime probability 0,145 and 4ch/d with target 11 :) (on 10ch version prime probability and ch/d decrease).

Post your calculations please.

primecoind -initwtdb does not work ???

Error: To use primecoind, you must set a rpcpassword in the configuration file:
/root/.primecoin/primecoin.conf
It is recommended you use the following random password:
rpcuser=primecoinrpc
rpcpassword=8nRsemgP777svRH5iWqnbkHnrYDaNALEsV4iy9VcmLte
(you do not need to remember this password)
The username and password MUST NOT be the same.
If the file does not exist, create it with owner-readable-only file permissions.
It is also recommended to set alertnotify so you are notified of problems;
for example: alertnotify=echo %s | mail -s "Primecoin Alert" admin@foo.com

primecoind: /usr/include/boost/thread/pthread/recursive_mutex.hpp:107: void boost::recursive_mutex::lock(): Assertion `!pthread_mutex_lock(&m)' failed.

Try ./primecoind -username (a_username) -password (a_password) -initwtdb

It does say what to do tho! its saying make a .conf and put a password and username in it but i would just run it via command

thank you but it doesnot create the dateabse tables

Then i'm thinking you haven't installed wt properly..try using a ppa and apt-get all the witty-dev stuff

Finally successful!

That's funny numbers :) The formula itself is good, probably something is wrong with input data.

> Post your calculations please.

Ok, check PM.

dont corrupt ex into pm secrecy

This has been covered in the past in this thread you need to CREATE ROLE as a postgresql user i had the same issue

Getting tired of waiting for claymore to release an optimized 11ch miner, 10ch just not hitting blocks anymore. I browsed through the thread here but didnt see the download links anywhere. Anyone care to recap 11ch solo mining for me? Saw the pool link, anyone know the pool/Dev fee?

Is thee any chance someone can post a simplified procedure with links may be to compile and run this on windows 64 ???

Hey eX, how do I connect to your pool? Do I need to pm my ip address to you?

Sending PM..

hey ex any chance we can grab your update 11ch pool code? Because id like to pay out more for 11ch on my pool thanks

Is it working normal i mean is it paying out normally like the old code was for the work the miners are putting in? since diff 10.955 its been so hard lol

Thanks again eX...

The pool is slightly slower for the pass 24 hours. That's why I want to try eX's pool out.

is it possible to let this miner support ypool or gpool?

Hello eX,

how to connect with your pool , i had sent PM !

Thx

I am torn between 10ch and 11ch Kernel, what are your experiences?

Can anyone explain why currently only ~40 Workers on eXremal's pool (about 10 times less than some days ago)?

I switched my rigs to X11 last night. SYNC was generating mad profits.
I will wait for eX to release his improved 11-chain miner and then give it a try again.

SYNC?

Yes, SYNC: https://bitcointalk.org/index.php?topic=607919
The introduction peak is already over but last night it was fantastic. I rented my rigs at 0.0144 BTC per GPU(280X) per day :)
Now prices have dropped a bit but still I get 0.007 per GPU(280X) per day, which is still great.
So today you will probably not see me on eX his pool.

Can someone point me towards the windows binary version of this miner for solo? the only miner i can find is for a pool that is now shutdown?

You have to modify the miner with the following: https://bitcointalk.org/index.php?topic=602292.200 . After that, compile and it should run fine.

Can this solo version mine other altcoin cloned from XPM ?

Which altcoin is cloned from XPM?

datacoin for example

Yes, with about 10 minutes of work to clone and add Datacoin support to the Pool Server:

2014-05-20 05:11:07 NotifyNewBlock(323189)
2014-05-20 05:11:07 ProcessBlock: ACCEPTED
2014-05-20 05:11:09 getblocks -1 to 0000000000000000000000000000000000000000000000000000000000000000 limit 500
2014-05-20 05:11:09 getblocks -1 to 0000000000000000000000000000000000000000000000000000000000000000 limit 500
2014-05-20 05:11:10 PrimeWorker 0: mInvCount = 0/700
2014-05-20 05:11:23 DatacoinMiner:
2014-05-20 05:11:23 proof-of-work found 
  target: 09.8e604c
  multiplier: 6c8ca53fb32e17a0
  CBlock(hash=d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd, hashBlockHeader=db4846af9dfd79c05aa7e445949f881a367d39a5af2c753bdc0f9e582a55c304, ver=2, hashPrevBlock=751fadb434403229378b4cb65d7d57b7df4f5f87df4fab86739489d39521c637, hashMerkleRoot=426e92de899e4bda892785b79253288469b98eb4cf18136976385ec278ac0cce, nTime=1400562666, nBits=098e604c, nNonce=34929537, vtx=1)
2014-05-20 05:11:23   CTransaction(hash=426e92de899e4bda892785b79253288469b98eb4cf18136976385ec278ac0cce, ver=1, vin.size=1, vout.size=1, nLockTime=0, data.size=0)
    CTxIn(COutPoint(0000000000000000000000000000000000000000000000000000000000000000, 4294967295), coinbase 0376ee04010100062f503253482f)
    CTxOut(nValue=10.93000000, scriptPubKey=03143e6e8cf9bc1a80193f6c2d86d0)
  vMerkleTree: 426e92de899e4bda892785b79253288469b98eb4cf18136976385ec278ac0cce
2014-05-20 05:11:23 generated 10.93
2014-05-20 05:11:23 keypool keep 492
2014-05-20 05:11:23 AddToWallet 426e92de899e4bda892785b79253288469b98eb4cf18136976385ec278ac0cce  new
2014-05-20 05:11:23 Committing 1 changed transactions to coin database...
2014-05-20 05:11:23 SetBestChain: new best=d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd  height=323190  difficulty=9.5561569 log2Work=24.116464  log2ChainWork=42.525248  tx=379969  date=2014-05-20 05:11:06 progress=0.999969
2014-05-20 05:11:23 NotifyNewBlock(323190)
2014-05-20 05:11:23 ProcessBlock: ACCEPTED
2014-05-20 05:11:23 keypool reserve 493
2014-05-20 05:11:23 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:23 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:23 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:23 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:23 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:24 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:24 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:24 received getdata for: block d77d3948688257941765a1f3de2466a7591ec145483ebf1d0f45fc3fc13614cd
2014-05-20 05:11:25 Flushing wallet.dat
2014-05-20 05:11:25 Flushed wallet.dat 189ms

can you tell me how to edit to support datacoin?

is datacoin profitable?

Same here ;-)

No, I stopped GPU mining DTC about six months ago. For me, Primecoin mining has always been more profitable.

I will release the instructions after I optimize the sieve code to target chain lengths of 9. A single R9 280X GPU should be able to find 50+ blocks/day at the current difficulty - currently achieving about 25 blocks/day. I will not have time until this coming weekend to look at the sieving code - I am new to AMD GPU (OpenCL) programing so it might take me longer than a few minutes to optimize the OpenCL code..

[GPU 0] T=-1C A=-1% E=0 primes=0.103814 fermat=1135288/sec cpd=137.39/day
[GPU 1] T=-1C A=-1% E=0 primes=0.102871 fermat=1170235/sec cpd=130.44/day
(ST/INV/DUP): 131x 7ch(1/0/0) 7x 8ch(0/0/0) 3x 9ch(0/0/0)
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 1 found share: 8-ch type 3
Share accepted.
Work received: height=323697 diff=9.5617028 latency=0ms
GPU 1 found share: 9-ch type 3
GPU 1 found BLOCK!
GPU 1 found share: 7-ch type 3
Share accepted.
Stale share.
Work received: height=323698 diff=9.5617434 latency=0ms
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 1 found share: 7-ch type 2
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 0 found share: 8-ch type 2
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 1 found share: 7-ch type 3
Share accepted.
GPU 1 found share: 7-ch type 1
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
GPU 0 found share: 7-ch type 2
Share accepted.
GPU 1 found share: 7-ch type 3
Share accepted.
GPU 0 found share: 7-ch type 3
Share accepted.
[GPU 0] T=-1C A=-1% E=0 primes=0.103545 fermat=1128918/sec cpd=133.46/day
[GPU 1] T=-1C A=-1% E=0 primes=0.102972 fermat=1176122/sec cpd=132.26/day
(ST/INV/DUP): 143x 7ch(2/0/0) 9x 8ch(0/0/0) 4x 9ch(0/0/0)

+1
though I think datacoin is a waste of time ;)

Well, I dont need miner code, only pool code for now, which, as far as  I understand it, is used only as server giving work to clients ;-)

Thanks in advance!

Guys, how many btc per scrypt mh/s would you guesstimate right now? (280x).

I'm folding and getting curecoins and it's really good but it's nice to have a fallback plan.

Thanks!

how much average you are earning per day with one R9 280x doing folding?

I'd say 2.5-4 curecoins a day.

Miner still alive, I found some optimizations for sieve, result is:
7,75CPD on Radeon R9 290X, new version will be released in next 12 hours.

amazing!

I'll give it a shot...thanks!

Is anyone have it working ok for mining datacoin?


Is that on 10-ch or 11-ch optimization?

Both.
I will release only 10ch, 11ch release required client code changing.

Right on....

would you be affecting your original code or madmax's ?

How about R9 280X ? 5~6CPD ?

He is just gonna release more .bin for windows miner folder and .cl file for linux compile! its an optimization he said and the 11ch. Says he would have to change the miner code so no its just a .bin file and .cl files for linux! for the present release he seems to be doing!

I sure hope I am not reading too much into this:

Technically, .cl/.bin are OpenCL kernel code.

I found some optimizations for MadMax's sieve code.

11ch requires cpp code changing (and recompiling for windows/linux), not only OpenCL kernel.
5.4-5,5CPD may be.. I don't have R9 280X for test.

sorry to mess arround, but CureCoin mining is way more profitable than Primecoin mining.

it consumes same power as X11  (cool n quiet fans) and im earning 0.013BTC/day per singler R9 280x

Just look out for Folding@home project, its way more profitable right now

Thanks for sharing it with us, really appreciated it. Now if you could bring your msg to the appropriate thread, that would be better; it is out of topic here.

actually I think it belong to that thread, most of the xpm miner have moved there... may-be he just feels alone  ;D

i was trying to help getting the latest info about highest profitable mining, anyway i was just trying to help others, sorry if injured someones feeling (cant think of any that will feel damaged by helping earning more.... but there world is vast i learnd i could find impossible persons)

Thanks, i speak for everyone when i say you didn't hurt anyones feelings but you are off topic and were called on it so keep your new and profitable coins to chartrooms and appropriate threads..

Thats how this works mate hence all of the miners on coins that are not the most profitable new fandangle coin, they are not interested in it hence being on topic in their specific coins threads.

Speaking of ontopic i have about 75cpd getting about 5-7 blocks per day still with this miner and the pool so thats great this increase will sit nicely with us being that we don't want to go anywhere else :)

i have tried this miner but i cannot get it to work.

with the default config.txt it runs but cannot connect to the pool.

if i add another pool in the config and do not change anything, after starting the client, it will close after 2-3 seconds

can anyone please give me some help with it?

Send eXtremal a pm he has a pool and will set you up im pretty sure if you just pm him and if you get no luck with him (if he thinks his pool is too full) then pm me and ill hook you up on mine

https://www.dropbox.com/s/fgok9qjj7ralf04/patch_v3.tar.gz


About 7,8 CPD with Radeon R9 290X.

NEWS: I've finished the solo-mining server.

The solo-mining server code is at: https://github.com/madMAx43v3r/xpmserver

The solo-mining server binaries can be downloaded at:
For Windows: https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI (https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI)
For Linux: https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k (https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k)
Note: The windows version crashes on shutdown, have not been able to fix this.

The binary zips also include a README.txt you should check out to get started.

EDIT: If you have dynamic ip: There is a bug somewhere in the bitcoin network code that can cause the server to loose all connections to the network after your external ip has changed. It will still show 8 or more connections but if you look in debug.log you will no longer see new blocks or transactions. In this case you need to restart the server.
The chance of this bug to happen is around 50%. So be careful, if your external ip changes and you don't check the server and the bug has happend you will produce 100% orphan blocks.

Good :) But it can't work with standard primecoin client from official site.

Next (experimental) patch https://www.dropbox.com/s/q9kzdvpl0rx8rf3/patch_v4.tar.gz

Why? I've tested and it works.


Nice job :)

Amazing news! :o

nice job!

Thank you eXtremal and madMax!

@eXtremal, you mentioned that you would need to patch client code? Not released yet?

@madMax, same link for Windows and Linux? Both of them are Windows.

Anyone encounter error forking madMax's github code if you already have primecoin or ppcoin forked into your repo?

Link for linux version fixed, thanks for reporting.

Linux xpmserver: https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k (https://mega.co.nz/#!NZAHWC6b!FyeF6aKZ32KPs_Ns0NwZsEGaOWkNAlwfSz76hO-Dg_k)

EDIT: Guys please, targeting 11-chains is only more profitable when diff >10.99 at least (sunny king ain't stupid and my calcs show this also)
Currently the miner does not support uneven targets (resulting in wrong CPD readings) because i didn't expect anyone to go for 11 anytime soon.
However supercomputing is testing a fix right now, so we'll see.

2 MadMax
the solo-mining server
At what address wallet coins will come in the case of finding the block.
What percentage of a developer?

I don't understand how I can mine with official wallet (http://primecoin.io/downloads.php).. it's possible only with patched wallet from github.

How much CPD on R9 280X with last version? Quote from Claymore miner's thread:
Claymore 10.0 miner shows 5,33CPD for one R9 280X.

I have a true 11ch version with 0,27CPD on R9 290X, it's unusable with current diffuculty, let's wait 10.99 (may be 10.98x).

You can't. madMax patched official wallet as he is reusing his protobuf/czmq/zmq/utf8 code from pool server, integrating into the wallet to simplify the coding.


no dev fee inside. The patched wallet probably can use the blockchain and keys on the same system where you installed the official wallet. Else, it will be a new address

Yeah, it's a new address for each block, just like the original code.
And yes, there is no dev fee.

but users are most welcome to donate to madmax for his hardwork  :D, not forgetting eXtremal too

hi everybody,

I just wanted to let you know, that I integrated madmax' source code & extremal's v3 optimizations into my primecoin pool miner, "xolominer", which can be used to mine on www.beeeeer.org
The official announcement can be found here: ppc/xpm-talk ANN thread (http://www.peercointalk.org/index.php?topic=485.msg25871#msg25871)

I integrated a developer fee (1%) pointing at madmax' donation address.
Also, it's experimental and not 100% perfect yet (missing some comfort features and cleanup).

Well, uhm... thank you so far.

- xolokram

ps. I'll try this monitor this thread here, but I'm more active on the ppc/xpm-talk board

nice and works on 69xx series wish there was a solo ver too

Hello,

Anyone tested on old GPU like HD-5970?

Errrm have you read this lately
Maybe i'm misunderstanding what it is you want!?

said it clearly a solo ver that works with 69xx series

try it out works on 6970/50

+1

While it may work, pre GCN cards are bad at primecoin mining, there's no way around it.

R7 240 is just as quick as a HD5870/6970. VLIW isn't optimize to do primecoin mining.

Im still really confused are you wishing beerminer was solo even though its mads code anyhow cause mad has a solo version you know that right? Both open source so that means optimization on one will be pulled and used on another being that they both use the same fermat, sieve

@eXtremal:

Tyvm good sir, again!

Finally I had time and was able to produce full, remote pools with databases, webstats and workers able to get for XPM & DTC.
Workers even found some blocks lol.

its a mix of his code plus u mean fermat. like saying cgminer and its forks are the same

Cool do what you want but have you tried out mads solo client yet?

Is the mega link in op's post a binary of the latest code changes (7 hours ago on github)

This link is the one i'm talking about: https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI

Does anyone have this file, the link is expired...

This post (https://bitcointalk.org/index.php?topic=602292.msg6714779#msg6714779) is from the guy who made the optimizations to madmax's code so go through his posts to find the ones that work for you we are up to v3 but im still using that one v2


HINT: use the 10ch

I have posted in this thread before on how to compile your own OpenCL bin, which is quite simple: You just need to put v3 or v4's .cl files in gpu dir and run the miner program.

This is the solo miner made by madmax that patches the primecoin wallet to talk with xpmclient.

Well, I have 3 of those and an r9 290...might as well try.

 

r7 240 is not as quick as a 5870/6970 in anything and none of the architectures are optimized for mining

I've done a 24hours test on a 5870, 6970 and R7 240, they all do 0.50 CPD each. I didn't say it was optimized, I said "isn't optimize"...

well thats still wrong on the r7 240 being the same as a 5870/6970 they are 4 to 5 times the card and saying it isnt optimized is like saying some are when none are

also i get more than .5 cpd on 6970

Anyone was able to make ADL work under linux? ;-)

*bump*

can someone PM me the IP and port number of the XPM pool? plz

Look on page 11

port 8081 is definatelly not the port to connect thru.... anybody know the port number?

Try 162.244.25.26 no port, no fee, use at your own risk

Pool donation addy is AWHqZqmAwPgmjiDnZVHWdD8t4ug2QF6tPi

i want to use extremal pool, i dont trust anonymous or unknown pools

8081 is the web interface.

The pool is safe, been using it for a week or so now with no issues...

eXtremal, any updates to the miner?

Hi,

I've made new pool for datacoin.

In miner config.txt use:

server ip: 92.222.24.57, port: 6666, as address use your datacoin address.

Web stats are here: http://92.222.24.57:6680/
Pool will send coins to your datacoin address when you mine more than 11 DTC, when they have mature to 100 confirmations. With GPUs that will be faster than you think :-D

Sample miner working:
Work received: height=339420 diff=9.7432844 latency=32ms
GPU 0 found share: 7-ch type 2
GPU 0 found share: 8-ch type 2
Share accepted.
Share accepted.



Use on your own risk! Good luck and have fun!

Marcus.

Does the madMAx xpm work with it? Or does it need a modded version?

Yes, miner from 1st page of this thread  works in 100%, I use linux version mostly, but just checked it on my 270X on Windows, and was able to get 50 cpd with little tweaking.

Is your pool ddos protected? or you will suffer same problem with madMAx510510

I am giving it a try for a day and see how it actually got paid compare with XPM... Is there a fee in your pool?

re: CPD, xpmclient still shows the original primecoin 10-ch data... I assume it is roughly 10x for 9-ch of DTC but then 270X with 50cpd is doing much more than 10x...

EDIT: "Your workers" on the web interface doesn't work neither... so it is no way to know the true cpd...

Datacoin diff is 9 so i think he means 50x 9ch! Your miner says the true cpd targeting 10ch!

Yes I know the miner shows 10ch, but if 9ch cpd is approximately 10 times of 10ch than his little tweak is way more than little as it means a 270x can do 5cpd for 10ch... of course, it is all rough estimation?

idk mate i wouldn't take his calculations as gospel id just mine and be happy with it. Go and talk on the datacoin threads for someone to tweak it i guess

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

edit:

270X settings for DTC:

What I noticed is when I use single card its CDP is reported to pool, but when I use f.e. 4 cards miner shows CDP but pool is not.
I will try to investigate it in next few days and broken ADL under linux too.

how to make the kernel.bin?

You have to have gpu/ folder inside miner.exe location, with *.cl files.
Rename old kernel to kernel.bin.old for safety. Run miner, new kernel should compile. Done.

I didn't mine much XPM so even I plan to donate 10% to MadMax it is just a few coins... and I may not mine XPM any more as the difficulties raised to a point that makes it not much profitable...

But I think I can make some contribution instead of donation  :)

As MadMax said he used cgminer adl.cpp, which I think is a copy of some sample code too.
He changed it from c to cpp but nothing big is changed.

I figured out there are 2 things to make it works.

Like using cgminer, you need to set
export DISPLAY=:0

otherwise the ADL_Main_Control_Create seems thinking no active adaptor available giving "ADL Initialisation Error! Error"

After fixing this there is a second error from ADL_Adapter_AdapterInfo_Get. I checked the error code and it said one parameter is wrong/mismatch.
This is subtle, and not sure if only works in my case. I found that it is sizeof argument ... maybe in Windows it is default to 32bit but in Linux 64bit for etc.
I cast it to unsigned long and it works. i.e.

                result = ADL_Adapter_AdapterInfo_Get (lpInfo, (unsigned long) sizeof (AdapterInfo) * iNumberAdapters);

See if MadMax or eXtremal can compile xpmclient for Linux distribution.

I found out how to compile the cl when I checked the code too :D
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 will compile it today evening, now I'm in work, and then 400km to travel :-)
Thanks for sharing solution!

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 :D

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

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

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.

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...

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.

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?

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

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.

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

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

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.

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.

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.

+1  :D

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

You are free to port his source to this miner :-D

I think it's safe to assume that this open source project has died out?

eXtremal did say he was working on a Fermat test optimization.

Currently the XPM price has fallen quite a bit .. ppl likely to shift their computes elsewhere.

https://www.dropbox.com/s/81rzx2cwlj6zzfn/patch_v5.tar.gz
Radeon R9 290X @ 1040/1250MHz:

Fermat test was optimized by 15%, but this opensource miner slower than proprietary windows-only miners about 20%, not good :(

well, proprietary one has a 10% dev fee... so don't give up  ;)

BTW, last time I said I found the Linux adl bug but it didn't work...
I re-examined it and I found that I pointed out the wrong coding sections.
It should be the missing of casting of the sizeof to the malloc function @ line #279

            lpInfo = (LPAdapterInfo) malloc ( (unsigned) sizeof (AdapterInfo) * iNumberAdapters );

Having this open sourced and running on linux makes it worth it.
Even with the speedups, it's still competitive.

Got some instructions to set this up on PiMP 64 bit?

What pools can we use?

Great that this thread didn't die and miner is still maintained :-)

You can use it for datacoin mining: http://datacoin.graymines.net

I use fixed client version with working ADL and fan control on Linux... this version will be releases soon, in forked git repository.

Try copy binary and "gpu" directory with kernel.bin to usb stick with PiMP. If it will not work, you will need build miner from sources.

XPM pool at http://89.179.126.88:8081 is alive.
Here is solo miner https://bitcointalk.org/index.php?topic=602292.msg6881582#msg6881582
For DTC mining use MarcusDe's pool (DTC difficulty still 9.xxx, you need patched kernel.bin for it).

Thanks a lot! Will need testing to get some profitability numbers now.

In free time I will make www with links and downloads for all kinds of cards and kernels to have stuff in one place with easy access.
Anyway, for now on DTC pool you will get good reward 9 DTC for every 10-chain you find, so you shoud think 2 times before changing kernel :-P

Hey eXtremal, could you please check your pool as it seems to be refusing connections? It was downloading the blockchain earlier today, and now my client can't go past "Connecting to frontend: 89.179.126.88:8081 ...". Thank you in advance!

Edit: It appears that frontend is now at  89.179.126.88:6666, works just fine.

https://www.dropbox.com/s/nqsdj0u9sweiq2w/patch_v6.tar.gz
A little sieve optimization, +8-9% (up to 8,9-9,0CPD on R9 290X).

Blockchain was corrupted, already fixed.

Good work! With patch v6 cpd for datacoin is doubled compared to month ago version.
Go miners, download it! :-)

This thread is a hot mess.  There are pages worth of hijack topic that don't belong here but what's worse is the lack of coherency towards any actual mining solution. 

I downloaded the solo server from here:

https://mega.co.nz/#!hRAmGRAQ!cbYlm-meXprzImw-vQ5COnBQiWtgPFXJeJtTs-041AI

Can someone tell me how I actually "compile" the miner code "at: https://github.com/madMAx43v3r/xpmclient" ? 

I've downloaded eclipse, Java SDK/JRE, GIT, etc and read numerous blogs/forums trying to find the answer.

There's a lot of good info in this thread but I've tried all day and can't crack this nut so far.

edit: feel free to PM me.  thx

It doesn't get much clearer then that!

Not coherent LMAO! the first bloody post page has all you need!
Doesn't get much more coherent then that mate sorry if i offend but cmon.

You come here offend everyone who has contributed to the thread then you ask for help..

Sorry, didn't mean to offend anyone. 

So what is it you need help with specifically?

I think I have all the dependencies now.  But I'm not versed in compiling.  I'm not sure about the config4Cpp.. couldn't find any place that had that... just the source.

I read the read me about config4cpp but can't get any of that to work:

Documentation
-------------

Comprehensive documentation for Config4Cpp is available. However, it is
distributed separately from the source code. You can find the
documentation (in PDF and HTML formats) on www.config4star.org.


Compilation instructions
------------------------

The build system has been tested on: (1) Linux with G++, (2) Cygwin with
G++, and (3) Windows with Visual C++ 6.0.

To build on Linux or Cygwin, run the following commands:

   make

To build on Windows with Visual C++, run the following commands:

   vcvars32.bat
   nmake -f Makefile.win

If you are building on another operating system, or with another
compiler, the you might need to edit "Makefile.inc" (if you are on a
UNIX-like operating system) or "Makefile.win.inc" (if you are on
Windows). You might also need to edit "src/platform.h" and
"src/platform.cpp".

Executables will be put into the "bin" directory, and library files will
be put into the "lib" directory.



I also just tried to run eclipse and open the makefile but I really don't know what I'm doing:

http://help.eclipse.org/juno/index.jsp?topic=%2Forg.eclipse.cdt.doc.user%2Fconcepts%2Fcdt_c_build_over.htm

Feel like I'm getting closer

Edit: I just opened the makefile.win in eclipse and ran it but got an error.

How do I "Open the Eclipse C++ Project"?  And then what do I do with it when I have it open?  Configure include paths, etc?  what is that?  Am I working with the .project file?

Under ubuntu most of stuff is avaiable with 'apt-get install', afair only zeromq had to be compiled from source.
If you have all libraries and tools needed for whole system, you just run 'make' in "Release blabla linux" folder and miner will compile under linux (and you don't have to use eclipse at all)

I'm on windows and it's not so much that I can't open programs it's more that I don't know which specific files to open or other specifics of how to make it work.

If you use Windows, why you want to compile miner yourself to use it, when its ready to use already (if not for sport)?
You just set config file, click on .exe file and its all :-)

Latest update: https://www.dropbox.com/s/ujfsc7vnwjtbv4a/patch_v6.1.tar.gz

3xRadeon R9 290X 1040/1250MHz

Can you please provide a link to this windows miner that is ready to use?  All I see is the link on the front that is to the github source.  Sorry, I must be blind.

Edit: also, I think the issue may also be with the sense that I want to modify the source to try out solo server: https://bitcointalk.org/index.php?topic=602292.msg6733542#msg6733542 

grendel25
https://bitcointalk.org/index.php?topic=598542.0 - see first post. Download miner for your OS and latest patch, replace kernel.bin file and pool address in config.txt... Also you can use solo miner.

from the first post - these are for the solo server:

Okay, cool. Think I'm getting the hang of mingw now too.  Thanks for the help.

What do I do with the "gpu" folder?  So far I just placed it in the same folder with the xpmclient.exe.  

Also, when I try to run it I get an error:


edit:  ahh... latest drivers... Yeah, I'm running 13.12 for another miner right now.  Oh well.  I'll come back to this in a week or so.  I think claymore is still faster for now.  Good luck

Opensource miner comeback: https://www.dropbox.com/s/zdq674972kpllup/xpmclient_v7.tar.gz
http://s019.radikal.ru/i637/1407/8b/32c25d7f5edb.png
It's single Sapphire Tri-X Radeon R9 290X stock on DTC pool, 12,7CPD.

Archive contains binaries for Linux x86-64 and Windows (sources not ready for release now, after 2-3 days I share updated source code in separate git repository) and config.txt templates for currently working XPM and DTC pools.
Linux version has problems with overclocking, 12CPD is limit for R9 290X now in Linux.

Wow you did it, wonderful! I have tested it briefly, on Linux

• 280X's cpd is 9.xx !! I think this miner becomes fastest again!!
• for 270, however, cpd is much lower than v6 (3.5 for v6, 2.7 for v7). (I have tried both 5 and 6 for hashprimorial.)
• I've got 2 warning messages per card after each new work received

For the fan speed, does it mean a fixed fan speed or max fan speed for auto?

ADL only recognition the first card in win7, please fix it, thanks for your work.
linux work well


[GPU 0] T=77C A=98% E=0 primes=0.105508 fermat=576426/sec cpd=8.51/day
[GPU 1] T=-1C A=-1% E=0 primes=0.105716 fermat=573558/sec cpd=8.64/day
[GPU 2] T=-1C A=-1% E=0 primes=0.106862 fermat=566284/sec cpd=9.50/day
[GPU 3] T=-1C A=-1% E=0 primes=0.106452 fermat=607183/sec cpd=9.80/day

try 2 or 3 for hashprimorial cpd is 5.xx for 270x

hashprimorial now not used and will be removed from config.txt, you can change sieveprimorial with range 13-15 for R9 270/280.

It's debug message, 2-3 messages after work is normal.

It's fixed fan speed.

Ok, I try fix it, but I have no access to multi-gpu windows machine for testing..

In latest version hash primorial not used.

After next update R9 290X will show over 13CPD :)

pm me i can give you a multirig for testing if you have teamviewer its windows 7 also ill put 2x 290x on it if you like

Well done eXtremal! Can't wait for the source

i think he can release source code when next  binary released, therefore to avoid someone use  his  code  for himself's paid miner, which is not good.  quote author=primer10 link=topic=602292.msg7790612#msg7790612 date=1405095507]
Well done eXtremal! Can't wait for the source
[/quote]

I don't finished work on source, it will be released tomorrow (14Jul) 99% :)

A little update now, version 7.1: https://www.dropbox.com/s/029gt5jk4tt8a2y/xpmclient_v7_1.tar.gz
http://i058.radikal.ru/1407/b9/b532dd7ac918.png
~13,45CPD on stock Sapphire Tri-X Radeon R9 290X.

My R9 280 eror!


--------------------------------------------------------------------------------------------------------
Found 2 devices
Using device 0 as GPU 0
Using device 1 as GPU 1
N=12 SIZE=4096 STRIPES=840 WIDTH=20 PCOUNT=40960 TARGET=10
GPU 0: has 32 CUs
GPU 0: primorial = 13082761331670030 (54 bits)
GPU 0: sieve size = 110100480 (27 bits)
GPU 0: max bits used 357/384
N=12 SIZE=4096 STRIPES=840 WIDTH=20 PCOUNT=40960 TARGET=10
GPU 1: has 32 CUs
GPU 1: primorial = 13082761331670030 (54 bits)
GPU 1: sieve size = 110100480 (27 bits)
GPU 1: max bits used 357/384
prepare_adl success
GPU 0 iAdapterIndex 0 strUDID PCI_VEN_1002&DEV_6798&SUBSYS_3001174B&REV_00_4&1B8
9D1F&0&00E0A iBusNumber 2 iDeviceNumber 0 iFunctionNumber 0 iVendorID 1002 strAd
apterName  AMD Radeon R9 200 Series
Failed to ADL_Adapter_ID_Get. Error -5Failed to ADL_Adapter_ID_Get. Error -5Fail
ed to ADL_Adapter_ID_Get. Error -5Failed to ADL_Adapter_ID_Get. Error -5Failed t
o ADL_Adapter_ID_Get. Error -5Failed to ADL_Adapter_ID_Get. Error -5Failed to AD
L_Adapter_ID_Get. Error -5ADL found less devices than opencl!There is possibly m
ore than one display attached to a GPUUse the gpu map feature to reliably map Op
enCL to ADLWARNING: Number of OpenCL and ADL devices did not match!Hardware moni
toring may NOT match up with devices!GPU 0 AMD Radeon R9 200 Series hardware mon
itoring enabled
Set fanspeed not supported
set_fanspeed(1, 70) failed.
Connecting to frontend: 89.179.126.88:6666 ...
Connecting to bitcoin: 89.179.126.88:60002 ...
Connecting to signals: 89.179.126.88:60003 ...
Work received: height=631593 diff=10.962315 latency=204ms
 * warning: no enough candidates available
 * warning: no enough candidates available
 * warning: no enough candidates available
 * warning: no enough candidates available
error: mpz_divisible_ui_p failed.