[XPM] Working on a GPU miner for Primecoin

[spirale]

1 BTC donated. Txid: 8fad5daaef41f5aac9eb01306161a4cef78e180b3c255db094dd60d126b642bc

[1715676754]

1715676754

[1715676754]

1715676754

[1715676754]

1715676754

[1715676754]

1715676754

[1715676754]

1715676754

[mineral]

Hi,

A GPU miner for PRimecoin implies to parallelize the 2 execution instances of BN_mod_exp(&r, &a, &e, &n, pctx). Paralleling of sieve would be an improvement, but it is not the central part , as is done in other OpenCL/CUDA primality tests where sieve is done in CPU.

So, it is indicative  when someone open a thread about parelellize a code and reports are about mingw and accesorials and anecdotical affairs with Windows building.

Well, for parallelizing  BN_mod_exp(&r, &a, &e, &n, pctx) this guy should to have rewrotten (at least) BigNum lib for OpenCL ....  

Good luck with the project and people who donated !!

[Pmalek]

What did you just say please?  Is it hard, impossible?

[Fallout]

It is hard but not impossible

[mikaelh]

Hi,

A GPU miner for PRimecoin implies to parallelize the 2 execution instances of BN_mod_exp(&r, &a, &e, &n, pctx). Paralleling of sieve would be an improvement, but it is not the central part , as is done in other OpenCL/CUDA primality tests where sieve is done in CPU.

So, it is indicative  when someone open a thread about parelellize a code and reports are about mingw and accesorials and anecdotical affairs with Windows building.

Well, for parallelizing  BN_mod_exp(&r, &a, &e, &n, pctx) this guy should to have rewrotten (at least) BigNum lib for OpenCL ....  

Good luck with the project and people who donated !!

Yup, if someone manages to use some level of parallelism for mod_exp on a GPU, then that will be a big thing. I'm personally still a bit doubtful whether it can be done efficiently because doing the mod operation requires division. Usually those algorithms involve unpredictable branching and that's going to slow things down on the GPU.

[mineral]

Hi,

A GPU miner for PRimecoin implies to parallelize the 2 execution instances of BN_mod_exp(&r, &a, &e, &n, pctx). Paralleling of sieve would be an improvement, but it is not the central part , as is done in other OpenCL/CUDA primality tests where sieve is done in CPU.

So, it is indicative  when someone open a thread about parelellize a code and reports are about mingw and accesorials and anecdotical affairs with Windows building.

Well, for parallelizing  BN_mod_exp(&r, &a, &e, &n, pctx) this guy should to have rewrotten (at least) BigNum lib for OpenCL ....  

Good luck with the project and people who donated !!

Yup, if someone manages to use some level of parallelism for mod_exp on a GPU, then that will be a big thing. I'm personally still a bit doubtful whether it can be done efficiently because doing the mod operation requires division. Usually those algorithms involve unpredictable branching and that's going to slow things down on the GPU.

Then, Which task are you going to send to GPU, only the sieve?

[mikaelh]

Then, Which task are you going to send to GPU, only the sieve?

Well, I do have some ideas how sieving could be implemented efficiently on a GPU. I'm not actively working on it though.

[Pmalek]

You smart guys should unite with Mtrlt for the good cause 

[gateway]

You smart guys should unite with Mtrlt for the good cause 

+1

Yea he's usually on irc ch #eligius-prime working with a few others or kicking ideas around.  Developers feel free to join and maybe provide some feedback.  Please no is it done questions or you will prob get kicked

[usahero]

@mtrlt How would you deal with refunds if you were unable to deliver a GPU miner at all?

Donation for GPU miner is a high risk, high return business.   

 No risk, if you donate on the day gpu miner will be released. Which doesn't sound too good for those of us who donated early.

[CoinBuzz]

can somebody explain to me how a primecoin mining algorithm can be partialized and distributed to core/gpu's ?

[cryptojournal]

Hey mtrlt,

I think this guy is trying to usurp your identity...
http://miner4world.blogspot.fr/search?updated-max=2013-07-20T15:00:00%2B07:00&max-results=2&start=2&by-date=false

at least the bitcoin address is different...

Sorry if this has been posted already, I didn't read the whole thread...

CJ

[usahero]

Dirty scammer!

[ReCat]

https://blockchain.info/address/1F3pakbmWsoWjsAytCBuMJCyZb1W5s3iX9

Well, at least he hasn't received much BTC. The OP should probably put a warning for this on the first post, though. THIS is the only place he's posting on. Right?

[igysa]

this is all very fishy 

[Lauda]

this is all very fishy 

No chill out for the moment 

[hasle2]

this is all very fishy 

No chill out for the moment 

I agree, I'm not worried about the extra time it is taking. Primecoin uses a new algorithm and it will take time to figure out how to get it to run efficiently on a GPU. The original deadline was unrealistic. It's not possible to predict ahead of time how much effort it will take to code something genuinely new.

[bidji29]

It's coming

[LZ]

1F3pakbmWsoWjsAytCBuMJCyZb1W5s3iX9

1F3pakbmWsoWjsAytCBuMJCyZb1W5s3iX9

Gotcha!

[itod]

1F3pakbmWsoWjsAytCBuMJCyZb1W5s3iX9

1F3pakbmWsoWjsAytCBuMJCyZb1W5s3iX9

Gotcha!

Yes, that's the scamer, good catch:
https://bitcointalk.org/index.php?action=profile;u=84753