[XPM] [ANN] Primecoin Release - First Scientific Computing Cryptocurrency

[Sunny King]

Does anyone have a sense of what's involved (or could point me in the right direction) in adding an option to mine these primecoins with GPUs?  I've got some experience working with nVidia's CUDA and might be able to contribute something of value here.  That being said, I could very well be far out of my league...

The built in miner uses a basic sieve. The mining process involves two steps, first generating a sieve, then do power test (Fermat) to check primality. For simplicity maybe a GPU core for the sieve can be first developed, and throw an API call to the client to check primality. You can investigate prime.h and prime.cpp for the basic sieve code.

[1714046968]

1714046968

[1714046968]

1714046968

[mercSuey]

Does anyone have a sense of what's involved (or could point me in the right direction) in adding an option to mine these primecoins with GPUs?  I've got some experience working with nVidia's CUDA and might be able to contribute something of value here.  That being said, I could very well be far out of my league...

The built in miner uses a basic sieve. The mining process involves two steps, first generating a sieve, then do power test (Fermat) to check primality. For simplicity maybe a GPU core for the sieve can be first developed, and throw an API call to the client to check primality. You can investigate prime.h and prime.cpp for the basic sieve code.

Sunny, theoretically, won't the primes get bigger as we find more of them, so that it's impossible to maintain a uniform time in between blocks?  So that in fact, if the network hash rate remains constant over some period of time, difficulty will adjust lower?

[Chemisist]

Does anyone have a sense of what's involved (or could point me in the right direction) in adding an option to mine these primecoins with GPUs?  I've got some experience working with nVidia's CUDA and might be able to contribute something of value here.  That being said, I could very well be far out of my league...

The built in miner uses a basic sieve. The mining process involves two steps, first generating a sieve, then do power test (Fermat) to check primality. For simplicity maybe a GPU core for the sieve can be first developed, and throw an API call to the client to check primality. You can investigate prime.h and prime.cpp for the basic sieve code.

I'll have a look tomorrow, thanks Sunny.

[Sunny King]

Sunny, theoretically, won't the primes get bigger as we find more of them, so that it's impossible to maintain a uniform time in between blocks?  So that in fact, if the network hash rate remains constant over some period of time, difficulty will adjust lower?

Difficulty is measured by chain length, not by prime sizes. Block spacing is maintained via continuous difficulty adjustment.

[KrLos]

i I would love some please... i had a sh**** pc. And been trying and nothing

Aeg98BVBRTq7jeH9u76FFRXaf769o8uxS5

[hak8or]

I have been mining away for the past hour using setgenerate true 3 on an I5-3570k (4 core stock) with this in the getmining info,

Code:

{
"blocks" : 709,
"currentblocksize" : 1000,
"currentblocktx" : 0,
"errors" : "",
"generate" : true,
"genproclimit" : 3,
"primespersec" : 89,
"pooledtx" : 0,
"testnet" : false
}

the primespersec varies from a low of 59 to an upper 115. Is this normal or what? I do not believe I have found a block yet sadly, based on transactions not showing anything nor getting any blockfound in the debug console.

[TheSpiral]

I have been mining away for the past hour using setgenerate true 3 on an I5-3570k (4 core stock) with this in the getmining info,

Code:

{
"blocks" : 709,
"currentblocksize" : 1000,
"currentblocktx" : 0,
"errors" : "",
"generate" : true,
"genproclimit" : 3,
"primespersec" : 89,
"pooledtx" : 0,
"testnet" : false
}

the primespersec varies from a low of 59 to an upper 115. Is this normal or what? I do not believe I have found a block yet sadly, based on transactions not showing anything nor getting any blockfound in the debug console.

That seems about right. I get a little bit higher on the same CPU on 3 cores as well. Found 1 block, it takes a while regardless. There's only been ~750 total blocks found so far, so we're all (relatively) in the same boat.

[BigJon901]

It's more of luck. I mined for 20 minutes and got a block on a Core-i5-2500k (no overclock). Not sure if the code is AVX instruction set optimized. If so newer CPUs have an advantage over older ones.

[TheMightyX]

keep in mind theres some degree of luck when mining. Even a crap pc has a slight chance of stumbling on the correct number and completing the correct calculation. A faster pc will go through more of those calculations, thus increasing its odds of finding the correct one.
my faster PC has been mining for several hours with no luck, my slower pc has already clinched two rewards. go figure.

[Vorksholk]

keep in mind theres some degree of luck when mining. Even a crap pc has a slight chance of stumbling on the correct number and completing the correct calculation. A faster pc will go through more of those calculations, thus increasing its odds of finding the correct one.
my faster PC has been mining for several hours with no luck, my slower pc has already clinched two rewards. go figure.

Same!! My crappy $300-$400 Toshiba laptop has gotten two blocks with a dual-core AMD processor, my i5-2500k has gotten one block, and my 8350 has gotten zero.

[hak8or]

Well, thanks for the information guys! I shall happily be awaiting my first block and seeing how this crypto goes.

[ReCat]

Probability mathematics/physics is a really damn confusing thing.

[Vorksholk]

keep in mind theres some degree of luck when mining. Even a crap pc has a slight chance of stumbling on the correct number and completing the correct calculation. A faster pc will go through more of those calculations, thus increasing its odds of finding the correct one.
my faster PC has been mining for several hours with no luck, my slower pc has already clinched two rewards. go figure.

Same!! My crappy $300-$400 Toshiba laptop has gotten two blocks with a dual-core AMD processor, my i5-2500k has gotten one block, and my 8350 has gotten zero.

As well, I have two Large elastic instances from digitalocean. Each scores >10000 on Geekbench, yet and each one reports somewhere between 0 and 2 primes per second. :sig:

[ReCat]

I'm pretty sure that hosting instances are nerfed in the release binaries so that nobody can get an unfair advantage with it.

[rethaw]

I'm pretty sure that hosting instances are nerfed in the release binaries so that nobody can get an unfair advantage with it.

Uh, how?

[ReCat]

CPUID checks. Hostname checks. Server specific information checks. It's not hard.

[jjiimm_64]

CPUID checks. Hostname checks. Server specific information checks. It's not hard.

this is why i scuffed at the guy that said this is a dead coin because someone has 100 EC2's on it.... 

[drummerjdb666]

Ok lol!  Apparently this was much easier to figure out than I made it.. cause i was also ina hurry to leave for work earlier.. but yea!  

You don't need to create any config file for this at all..  It just works without a config.. maybe you will need to reboot after the first time you try to open it.. but.. it only worked for me after i deleted the .conf i tried to create completely.. haha!  oops!  

Thnx a lot to the people that responded though!    

[hak8or]

Well whatdayaknow, I got my first mined block just now, hah.

[rethaw]

CPUID checks. Hostname checks. Server specific information checks. It's not hard.

With the code on github it would have to be obfuscated somehow. Care to point out what you think it is? I would be astonished if he intentionally broke virtual CPUs. That is just rewarding botnets and punishing everyone else.