[XPM] BFGMiner Branch.. on github

[A L I E N]

I did not publish this for people to use, but rather for developers to play with.
It is not production-quality code, and without understanding the code you probably will not be able to get it working.

Since it is also only used by a probable-scamcoin at this point, I'd also like to express my preference that others not encourage/help non-developers use it for now, outside of improving the code/documentation in git.
(However, to be clear, I will also not hold any grudge or hard feelings against those who do.)

If any developers wish to work with or improve the code, perhaps even clean it up sufficient for merging into mainline BFGMiner, that is another matter entirely: I am more than willing to help/collaborate on this (but do expect me to ignore you if your questions suggest to me that you don't know the first thing about development!).

Why do you think Primecoin is a scamcoin?

[1714735531]

1714735531

[1714735531]

1714735531

[1714735531]

1714735531

[Zalfrin]

With my first compile, --algo wasn't a supported switch at runtime.

I recompiled with:
CFLAGS="-O2 -msse2" ./configure --enable-cpumining --enable-prime
make

I don't know if --enable-prime did anything or not, but it didn't hurt.

Now I can launch with --algo prime and it asks for pool address. I will see if I can point to my local primecoind RPC port.

./bfgminer --algo=prime -o localhost:8000 -u rpcuser -p rpcpass

8000? default port is 9912, did you change it?

Whatever the daemons is running

This got me past the crash on launch anyways. Not sure if I was missing the "=" after --algo or what.

Now I can see "new blocks" arriving, although they don't line up with when I see new blocks in QT debug terminal. Difficulty is being reported as 0, could be an unimplemented piece of the code or could be an actual problem.



Doubtful that it's actually functional.

[Luke-Jr]

What's the difference between primecoind mining and this miner in terms of perfomance?.

About 100x.

[blastbob]

What's the difference between primecoind mining and this miner in terms of perfomance?.

About 100x.

Did you manage to block with this version? or is the mostly primemeter stats?

Is each primemeter stats line per cpu?

Anyway cred for starting something

[Mike270]

This got me past the crash on launch anyways. Not sure if I was missing the "=" after --algo or what.

Now I can see "new blocks" arriving, although they don't line up with when I see new blocks in QT debug terminal. Difficulty is being reported as 0, could be an unimplemented piece of the code or could be an actual problem.



Doubtful that it's actually functional.

You need to pass it a --scan-time 1, that way it will register new blocks faster (necessary for those short intervals we have)
It seems to pass whatever it finds, though, no proof-of-work check implemented yet as it seems.
So far all my block candidates it found on testnet were rejected(by primecoind). No block candidates on production yet.

[oroqen]

This got me past the crash on launch anyways. Not sure if I was missing the "=" after --algo or what.

Now I can see "new blocks" arriving, although they don't line up with when I see new blocks in QT debug terminal. Difficulty is being reported as 0, could be an unimplemented piece of the code or could be an actual problem.



Doubtful that it's actually functional.

You need to pass it a --scan-time 1, that way it will register new blocks faster (necessary for those short intervals we have)
It seems to pass whatever it finds, though, no proof-of-work check implemented yet as it seems.
So far all my block candidates it found on testnet were rejected(by primecoind). No block candidates on production yet.

I've gotten the same results as you, I even went as far as too modify coindbase-addr too accept the addresses

[shinkicker]

This is what I am seeing, but no idea if this is working correctly:



It states 680pps

I need to have a really good look at the code as its my first time playing around with this miner. Looks pretty awesome so far.

[Zalfrin]

You need to pass it a --scan-time 1, that way it will register new blocks faster (necessary for those short intervals we have)
It seems to pass whatever it finds, though, no proof-of-work check implemented yet as it seems.
So far all my block candidates it found on testnet were rejected(by primecoind). No block candidates on production yet.

Thanks for the info. Failing on testnet doesn't seem like a good sign, but let us know if you see any successes.

[Zalfrin]

Heh... mine seems to be trying to use the gpu for this, once I turn off generation in QT my CPU utilization drops to 0. Also my GPU mining performance for another altcoin is degraded.

[cosurgi]

are there some problems reporting correct difficulty? For me bfgminer reports:

Connected to localhost diff 0 without LP as user primecoinrpc
Block: ...d85d7682505eee91  Diff:0 (  inf?h/s)  Started: [00:24:02]

So it's difficulty zero.

[Mike270]

Heh... mine seems to be trying to use the gpu for this, once I turn off generation in QT my CPU utilization drops to 0. Also my GPU mining performance for another altcoin is degraded.

I guess you should do --disable-opencl and do a build only for cpu mining...
Testing my own optimizations right now, so there won't be further findings from me for a while :-)

[Mike270]

are there some problems reporting correct difficulty? For me bfgminer reports:

Connected to localhost diff 0 without LP as user primecoinrpc
Block: ...d85d7682505eee91  Diff:0 (  inf?h/s)  Started: [00:24:02]

So it's difficulty zero.

I like the infinitely unknown network hash rate :-)

[cosurgi]

hmm. Everything gets rejected. I tried on testnet, with a slightly lower difficulty, and I already have like 10 rejects:

[2013-07-11 23:40:06] Pool 0 stale share detected, submitting as user requested
[2013-07-11 23:40:06] Rejected f521f1b3 CPU55  Diff 0/18.4E
[2013-07-11 23:40:09] primemeter    626421 prime/h   4311771 test/h   174pps

[mkmen]

yeah, same here, only rejects on testnet, 2 on CPU0, 3 on CPU1 and 180 and rising on CPU2 (got just 3 enabled)

[96redformula]

On this step I get errors any ideas?


./autogen.sh


which throws the error

 Makefile.am:67 variable 'EXTRA_bfgminer_dependencies' is defined but no program or library has 'EXTRA_bfgminer' as canonical name

[96redformula]

On this step I get errors any ideas?

./autogen.sh

which throws the error

Makefile.am:67 variable 'EXTRA_bfgminer_dependencies' is defined but no program or library has 'EXTRA_bfgminer' as canonical name

[human]

This is what I am seeing, but no idea if this is working correctly:



It states 680pps

I need to have a really good look at the code as its my first time playing around with this miner. Looks pretty awesome so far.

Same picture here. Primemeter says something, but does it actually mine? (Almost) everything else is zero.
primemeter   6464834 prime/h  43856034 test/h  1795pps

Do I understand it right, that primecoind has to run in the background with generate=1 or maybe generate=0?

[Zalfrin]

Heh... seems configuring the make for disabled FPGAs and Avalon just doesn't work. Breaks the compiler from missing references. Oh well, disabling opencl works at least. Until runtime when it crashes. LOL.  Adding -S cpu:auto to enable CPU mining to my build with opencl still enabled gives a crash soon after start.

[AgentME]

Since it is also only used by a probable-scamcoin at this point, I'd also like to express my preference that others not encourage/help non-developers use it for now, outside of improving the code/documentation in git.
(However, to be clear, I will also not hold any grudge or hard feelings against those who do.)

What makes you say that? As far as I see, primecoin and the same developer's ppcoin look like the only bitcoin-derivatives to bring something new and worthwhile to the table since namecoin.

[shinkicker]

This is what I am seeing, but no idea if this is working correctly:



It states 680pps

I need to have a really good look at the code as its my first time playing around with this miner. Looks pretty awesome so far.

Same picture here. Primemeter says something, but does it actually mine? (Almost) everything else is zero.
primemeter   6464834 prime/h  43856034 test/h  1795pps

Do I understand it right, that primecoind has to run in the background with generate=1 or maybe generate=0?

No. you don't want primecoind mining. Remove gen= from your primecoin.conf and just run is as daemon

primecoind --daemon

From there don't do any setgenerate commands.

I am closed to giving up on this coin now, been running for another 24 hours without one block found (on the primecoind miner)! I don't do this for money, its a hobby, but this is not so much fun when there is little reward to see.