Pages:
Author

Topic: [XPM] [ANN] Primecoin High Performance | HP14 released! - page 41. (Read 397645 times)

full member
Activity: 122
Merit: 100
Hmm the difficulty has been rising pretty sharply for the last 20 hours or so:

http://192.241.170.170/

The site is slow to load sorry, I have not had the time to do ajax data loading. It is updated every 10 mins however.

Any speculations on what might be going on?

I could also still use some advice on how to tune my FX 8150 too Wink I've tried using -O3 and -march=native but am still only getting 0.45 chains Sad

hero member
Activity: 546
Merit: 500
My client has developed a bug on one of my machines where for over 24 hours it just sat on one block using 100% CPU meanwhile the blockchain is actually more than 1000 blocks ahead.

The client has 8 active connections and so it seems it is not a connectivity issue.

EDIT: 29 hours it seems, based on how much the restarted client must catch up. That agrees with the end of the debug log where it stops reporting mining stats at a few minutes after midnight when I was asleep:

Code:
2013-08-19 00:24:44 primemeter   4005206 prime/h  69499048 test/h  660 5-chains/h 0.671895 chain/d
2013-08-19 00:25:44 primemeter   3893025 prime/h  68190130 test/h  240 5-chains/h 0.658989 chain/d
2013-08-19 00:26:44 primemeter   3964314 prime/h  69105828 test/h  480 5-chains/h 0.669175 chain/d
dumpaddr thread stop
msghand thread interrupt
addcon thread interrupt
opencon thread interrupt
net thread interrupt
Flush(false)
DBFlush(false) ended               0ms
StopNode()
PrimecoinMiner terminated
keypool return 15
PrimecoinMiner terminated
keypool return 14
PrimecoinMiner terminated
PrimecoinMiner terminated
keypool return 17
keypool return 12
Flushed 11181 addresses to peers.dat  164ms
Committing 3 changed transactions to coin database...
Flush(true)
wallet.dat refcount=0
wallet.dat checkpoint
wallet.dat detach
wallet.dat closed
DBFlush(true) ended              94ms

full member
Activity: 122
Merit: 100
What are you guys doing to tune your AMD processors? I have an FX 8150 and only get around 0.45 chains/day on it running hp9. It also does about 1/2 of what a i7 3770 can do on jhPrimeminer.
newbie
Activity: 25
Merit: 0
With the fractional difficulty reaching a high level a useful optimization might be the "Sieve depth", represented as nBits in this line of code from prime.cpp

lpsieve = new CSieveOfEratosthenes(nSieveSize, nBits, mpzHash, mpzFixedMultiplier, pindexPrev);

Right now it it is searching for 9 chains, but increasingly these are going to be discarded because of the fraction test. It might be best to search for 10 chains now. I added a -h parameter to the pool miner as described below and it was quite effective at searching for the desired length of chain:

https://bitcointalksearch.org/topic/m.2820614

I propose -h 10 might be the best until the diff reaches 10.5 or so.

This chart also shows it being very difficult to find chains using the existing miners when the fractional part is high:
http://192.241.170.170/

I support this idea.  Although, to my knowledge all that will affect is the roundPrimorial which you can crank up by jacking with roundSievePercentage.

So should we make roundSievePercentage higher or lower, it's currently 70% at my configs. Does this also depend on what kind of CPU we're using?
hero member
Activity: 546
Merit: 500
I have this on one VPS:

    "chainsperday" : 1.13775088,

Is that about 1 block a day?

No, because the diff is ~9.75 now. The chainsperday refers only to the integer (9) difficulty not the fractional part (0.75)

So actually your chance of finding a block is closer to 0.3 per day, assuming no variance.

hero member
Activity: 532
Merit: 500
I have this on one VPS:

    "chainsperday" : 1.13775088,

Is that about 1 block a day?
I'm at .86 and found 4 blocks over 7 days.  Maybe in the long run, yes, but random still affects.

Alright, thanks. I am still deciding to mine this or other alt coins w/ CPU power.  Roll Eyes
Personally I don't think the other CPU only coins are as profitable.  I could be wrong though.
hero member
Activity: 532
Merit: 500
I have this on one VPS:

    "chainsperday" : 1.13775088,

Is that about 1 block a day?
I'm at .86 and found 4 blocks over 7 days.  Maybe in the long run, yes, but random still affects.
hero member
Activity: 532
Merit: 500
I have to admit I'm having mixed feelings about -hp9.  7 days ago I took my 2400 and 2500 out of ypool and started running -hp9.  The 2500 found a block that very day, I was happy.  Day 3 the 2500 found another one.  Day 6 and 7 it  also finds blocks.  The 2400 completely struck out and been switched back to ypool.  I estimate I was getting 5-7 XPM/day at ypool before I switched.  The 4 blocks plus 9xpm at ypool seems to indicate I did well, but that goosegg on the 2400 kind of rankles :/
sr. member
Activity: 520
Merit: 253
555
Hi, I am just starting to look at Primecoin and wanted to know if anyone has advise on which miner is best. After looking through the threads it seems there are two main options:

Holy Satoshi! I thought this thread was all about VPS providers, rather than the actual technology and the coin Wink Thanks for getting it back on track Smiley
full member
Activity: 162
Merit: 100
So what are some good VPS hosts that everyone is using?

We're at a point where all "good" vps providers are too expensive !

At least all I know of, dif is still going up even if really slowly so some must have found better farming place than I did...
The ones I have researched and used are now break even at best. Some others may have better luck or better deals. Buying old kit off EBay sounds a good idea from one of the other posters.
sr. member
Activity: 476
Merit: 250
So what are some good VPS hosts that everyone is using?

We're at a point where all "good" vps providers are too expensive !

At least all I know of, dif is still going up even if really slowly so some must have found better farming place than I did...
sr. member
Activity: 476
Merit: 250
With the fractional difficulty reaching a high level a useful optimization might be the "Sieve depth", represented as nBits in this line of code from prime.cpp

lpsieve = new CSieveOfEratosthenes(nSieveSize, nBits, mpzHash, mpzFixedMultiplier, pindexPrev);

Right now it it is searching for 9 chains, but increasingly these are going to be discarded because of the fraction test. It might be best to search for 10 chains now. I added a -h parameter to the pool miner as described below and it was quite effective at searching for the desired length of chain:

https://bitcointalksearch.org/topic/m.2820614

I propose -h 10 might be the best until the diff reaches 10.5 or so.

This chart also shows it being very difficult to find chains using the existing miners when the fractional part is high:
http://192.241.170.170/

I support this idea.  Although, to my knowledge all that will affect is the roundPrimorial which you can crank up by jacking with roundSievePercentage.

Interesting !

I'm at 80% already ? you guys ?
sr. member
Activity: 476
Merit: 250
I just rented two vps here : Inception Hosting

They got quickly shut down as abuse of their terms of service, I did not read anything regarding mining on their terms but whatever...

I told them I would warn you guys not to loose your time with them !

I'm supposed to be refunded so they aren't that bad either...
full member
Activity: 168
Merit: 100
With the fractional difficulty reaching a high level a useful optimization might be the "Sieve depth", represented as nBits in this line of code from prime.cpp

lpsieve = new CSieveOfEratosthenes(nSieveSize, nBits, mpzHash, mpzFixedMultiplier, pindexPrev);

Right now it it is searching for 9 chains, but increasingly these are going to be discarded because of the fraction test. It might be best to search for 10 chains now. I added a -h parameter to the pool miner as described below and it was quite effective at searching for the desired length of chain:

https://bitcointalksearch.org/topic/m.2820614

I propose -h 10 might be the best until the diff reaches 10.5 or so.

This chart also shows it being very difficult to find chains using the existing miners when the fractional part is high:
http://192.241.170.170/

I support this idea.  Although, to my knowledge all that will affect is the roundPrimorial which you can crank up by jacking with roundSievePercentage.
sr. member
Activity: 604
Merit: 250
With the fractional difficulty reaching a high level a useful optimization might be the "Sieve depth", represented as nBits in this line of code from prime.cpp

lpsieve = new CSieveOfEratosthenes(nSieveSize, nBits, mpzHash, mpzFixedMultiplier, pindexPrev);

Right now it it is searching for 9 chains, but increasingly these are going to be discarded because of the fraction test. It might be best to search for 10 chains now. I added a -h parameter to the pool miner as described below and it was quite effective at searching for the desired length of chain:

https://bitcointalksearch.org/topic/m.2820614

I propose -h 10 might be the best until the diff reaches 10.5 or so.

This chart also shows it being very difficult to find chains using the existing miners when the fractional part is high:
http://192.241.170.170/
hero member
Activity: 546
Merit: 500
Hi, I am just starting to look at Primecoin and wanted to know if anyone has advise on which miner is best. After looking through the threads it seems there are two main options:

1) Use mikaelh's miner from this thread and solo mine only.
2) Use the jhPrimeminer and pool mine at yPool.

From reading, it seems that mikaelh's miner is more efficient and will mine more blocks over time. But if you want the stability of a pool the only option is jhPrimeminer which sounds like it is a little bit less efficient.

Is this correct? If not what's the best option to get started mining? Any opinions appreciated. Thanks

You can install both and then compare them. With mikaelh's HP9, make a shortcut to the .exe and append '-gen' to the end of the target field. Then let it download the blockchain and let it start mining for a few minutes. Use 'getmininginfo' in the console and see what your CPU gives you. Currently we are taking 'primesperday' statistic as the most accurate estimate of performance.

You can then also pool mine for a while. Maybe a few days running each to reduce variance (your 'luck'). And compare your results.

If you are only going to mine on 1 or 2 CPU's and/or they are slow, pool mining might be better because it may otherwise take days to find a block mining solo. On the other hand some miners have many CPU's and usually find at least one block per day.

If you tell me what CPU's you have (model number: right click properties on Computer in most versions of windows, or, use CPU-Z) I could estimate your machines performance roughly.
full member
Activity: 322
Merit: 113
Sinbad Mixer: Mix Your BTC Quickly
Hi, I am just starting to look at Primecoin and wanted to know if anyone has advise on which miner is best. After looking through the threads it seems there are two main options:

1) Use mikaelh's miner from this thread and solo mine only.
2) Use the jhPrimeminer and pool mine at yPool.

From reading, it seems that mikaelh's miner is more efficient and will mine more blocks over time. But if you want the stability of a pool the only option is jhPrimeminer which sounds like it is a little bit less efficient.

Is this correct? If not what's the best option to get started mining? Any opinions appreciated. Thanks
Yes, basically correct. There are several versions of jhPrimeminer to tinker with, though. And you won't necessarily get more blocks solo'ing, but there's a chance you will get some and a chance you won't, whereas pool mining will give you a steady trickle of XPM.
legendary
Activity: 1153
Merit: 1000
Hi, I am just starting to look at Primecoin and wanted to know if anyone has advise on which miner is best. After looking through the threads it seems there are two main options:

1) Use mikaelh's miner from this thread and solo mine only.
2) Use the jhPrimeminer and pool mine at yPool.

From reading, it seems that mikaelh's miner is more efficient and will mine more blocks over time. But if you want the stability of a pool the only option is jhPrimeminer which sounds like it is a little bit less efficient.

Is this correct? If not what's the best option to get started mining? Any opinions appreciated. Thanks
hero member
Activity: 546
Merit: 500
While sitting in your chair running VPS remotely is nice, I preferred to take an alternate route.

I originally got a bunch (read: 12) old HP servers with 64-bit CPU's at an IT auction. Turned out they were far too inefficient (about 350 watts for <0.2 chainsperday). Ended up selling most of these and making a nice profit anyway.

A much better option is 2nd hand desktop PC's, core2duo CPU are the minimum I settle for now. These kind of machines can be had for more or less free if you find companies trying to get rid of their end of life systems and avoid disposal costs. Even at auctions I see systems like these go for <$20 every week.

I am running four on a wattmeter right now, about 300 watts total which is about $60/month in electricity for 1.2 chainsperday. It's not as great as some VPS by numbers alone, but at the end of the day I have accumulated a bunch of computer systems and parts that are still can still be sold as tested parts on eBay. I already operate a store on eBay so this is in my interests anyway.
hero member
Activity: 675
Merit: 514
I thought the Phenom was better at running this, my 2400 and 2500 are at .86/.88.
0.80 sounds about right for 4 cores/3.2 GHz. My Phenom II X4 910e makes about 0.67 cpd with 2.6 GHz.
Pages:
Jump to: