Introducing CherryPicking - new Windows & Linux Pool Hopper

[ampirebus]

Bloodred you are the man! I can confirm with your help I am up and running at 100% 

[1714917586]

1714917586

[1714917586]

1714917586

[1714917586]

1714917586

[muyoso]

OK, weird.  Polmine is at 41% and CP just switched to Slush, which is at 98.7% of difficulty.  This is almost opposite behavior of bithopper.  Can't wait to see the results.

Edit:  I don't see how jumping late into a round at Slush is profitable.  Jumping in at the very beginning and staying for 10% of difficulty number of shares is profitable because you are only risking a small amount of shares and when they have a fast round you get a relatively massive payout (2-300% efficiency) and on long rounds you get nothing.  Works out to somewhere in the range of 100-150% efficiency from the days I have tested it on BH.  From the looks of it, you are trying the opposite.  Jumping in on long rounds for the last 10% of difficulty instead of the first.  The reason I don't think that would work as well is that you are betting on such a small range.  The round literally has to end at difficulty + 5-10% for the maximum payout, if what I see in the console is correct, which rarely happens.

Edit2:  Loving CherryPicking so far.  Would be absolutely amazing if you could do some sort of website output like bithopper and even better if you figured out a way to feed multiple instances of CP into one website output.  


Edit3: Getting this error on rigs not running a bitcoin server:

Code:

[Miner0] mtred.com:8337 Problems communicating with bitcoin RPC 0 2

Doesn't seem to be affecting anything.  Just want to make sure that I don't need to be running the bitcoin server for CP to work.

[Stupidpal]

Does SMPPS work with 0.6.3?

[MrWizard]

Does SMPPS work with 0.6.3?

You can set one as your backup pool.  That's what I have done.

[Bloodred]

Edit2:  Loving CherryPicking so far.  Would be absolutely amazing if you could do some sort of website output like bithopper and even better if you figured out a way to feed multiple instances of CP into one website output. 

Glad to hear that! About the website, I'll be honest about it: I suck at developing web-based stuff, so I can't make any promise, at least for the time being. However, if CherryPicking becomes more and more popular maybe somebody will make a sort of web front-end that scrapes info from the log or from the future .csv files and displays it online.
As for multiple instances, I assume you mean on different machines and I was already thinking of making it work in "master" and "slave" mode or something like that. Basically, you'd have slave instances of CP connecting to the master instance, the master instance would be the only one gathering stats and then sending them to the slaves, to avoid sending the same requests from all machines if they're on a single network. If I ever get around to doing this there will also be a few commands available to the master instance and I guess having it gather mining stats from the slaves wouldn't be too hard, so it could dump those to a file and the web front-end could get the info there.

This is still a sort of concept idea, I wouldn't say it's the lowest priority but there are other things that are more important right now, like tweaking the hopping algorithm for Score and seeing if PPLNS is worth it or not.
I'll check out that hop you mentioned, it shouldn't hop on Slush at 98.7%. Was that the actual diff percentage or the number it was displaying? The number is a priority calculated differently for different pool profiles.

Edit3: Getting this error on rigs not running a bitcoin server:

Code:

[Miner0] mtred.com:8337 Problems communicating with bitcoin RPC 0 2

Doesn't seem to be affecting anything.  Just want to make sure that I don't need to be running the bitcoin server for CP to work.

That's a message coming from poclbm itself (anything that has [Minerx] in front is coming directly from a poclbm instance), CP is just relaying it to you and it has nothing to do with it. There probably was lag spike or something and I think the bitcoin server stuff is purely coincidental, CP has absolutely no need of it. I'd ignore it if it isn't affecting mining. I've seen messages like those too while mining but they didn't seem to affect anything.
One thing is worth mentioning I guess, if CP gets 8 of those messages (from Miner0) in the span of 30s it'll mark the pool as lagging/down for 10 minutes and disable any mining on it. This is so it automatically switches from dead pools even if the stats API/web page is working. It doesn't connect to the miner hosts itself so it has to monitor poclbm's error output instead.

Does SMPPS work with 0.6.3?

No, there's no profile for SMPPS in 0.6.3, but you can indeed use them as BACKUP.

[flower1024]

if you guys want to use multiple miner pc's you could try my mod: https://bitcointalk.org/index.php?topic=34514.0

it make poclbm connect to a running bithopper and switches without stales.

as its a normal poclbm-pool connection there are no proxy-related problems. you can even restart bithopper without loosing shares.

[muyoso]

cherrypicker refuses to go to a backup for me.  It just jumps to the lowest prop pool even if its over the 43% threshold. What am i doing wrong?

[ampirebus]

cherrypicker refuses to go to a backup for me.  It just jumps to the lowest prop pool even if its over the 43% threshold. What am i doing wrong?

thats weird cause mine is pretty much just going to backup pools as of last nights log. spent half the night at eligius and went to slush's for 1 or 2 rounds and the other half was pretty much all mtred.

[Psychonautchn]

Could you make up a config. for ABCPool.co (2% pps)? Also, I'm still having problems with the security certificate for ARS, Mt.Red, and RFCPool, using both your updated files and manually updating jssecacerts. Otherwise CP works great, if I can keep 24hrs of up time without my internet cutting out I will give a definitive earnings comparison.

[Bloodred]

cherrypicker refuses to go to a backup for me.  It just jumps to the lowest prop pool even if its over the 43% threshold. What am i doing wrong?

Are the backup pools set properly? Can you please copy-paste the info it displays when updating the pools? I'll be able to tell what's happening from that. You need a working pool set as BACKUP or it will do just what you've described.

Could you make up a config. for ABCPool.co (2% pps)? Also, I'm still having problems with the security certificate for ARS, Mt.Red, and RFCPool, using both your updated files and manually updating jssecacerts. Otherwise CP works great, if I can keep 24hrs of up time without my internet cutting out I will give a definitive earnings comparison.

I've added ABCPool to the archive. As for the certs, I'm trying to find out what may be causing all these issues but it's difficult to diagnose since I can't reproduce the errors. Since you're getting errors with so many pools, are you sure you've copied the file in the correct location?
Somebody also messaged me and told me he was getting these errors because something was wrong with his JRE install and he had fixed it, but didn't mention what or how. Maybe reinstall the JRE?

[muyoso]

cherrypicker refuses to go to a backup for me.  It just jumps to the lowest prop pool even if its over the 43% threshold. What am i doing wrong?

Are the backup pools set properly? Can you please copy-paste the info it displays when updating the pools? I'll be able to tell what's happening from that. You need a working pool set as BACKUP or it will do just what you've described.

I have ars set to backup in the latest stable release.

Edit:  OK, found the problem.  In the config files you have the port set to 8334.  The port that works for me is 8344.  I think the reason that it wasn't switching to Ars for me last night is that Ars is set to 1.00264 and BTCPool24 was below that, although I will have to keep an eye on this to confirm it.  Seems to be working perfectly now that I changed the port.  Going to setup another backup JIC. 

[ampirebus]

sorry if this is a n00b question but why isnt deepbit hopped

[muyoso]

sorry if this is a n00b question but why isnt deepbit hopped

They delay their stats for an hour.  Basically its impossible right now to know when they found a block and hop to them. 

[ampirebus]

Uploaded with ImageShack.us

not so happy with this. i would have thought the hopper would have seen those short rounds and gone in... it completely avoided all the best rounds! and the ones that it did join are no payout at all . Spent time between eligius and mtred between then, which also ended up having very long rounds.  

below is a comparison of what i get on slush w/o cherrypicker on...

7140    2011-08-04 19:40:26    0:12:41    355009    0.01410793    139620    confirmed
7139    2011-08-04 19:27:46    0:06:22    178960    0.01227752    139618    confirmed

[hawks5999]

what's your estimated payout at mtred? what's your balance at eligius?

slush is score based and only experimentally hoppable, btw. You shouldn't expect big payouts from slush.

[muyoso]

Uploaded with ImageShack.us

not so happy with this. i would have thought the hopper would have seen those short rounds and gone in... it completely avoided all the best rounds! and the ones that it did join are no payout at all . Spent time between eligius and mtred between then, which also ended up having very long rounds.  

below is a comparison of what i get on slush w/o cherrypicker on...

7140    2011-08-04 19:40:26    0:12:41    355009    0.01410793    139620    confirmed
7139    2011-08-04 19:27:46    0:06:22    178960    0.01227752    139618    confirmed

Yea, the way cherrypicker does slush is completely backwards.  It should be doing it exactly like bithopper does, which is jumping in at the start of every round for 10% of difficulty shares and then hopping out.  I can confirm that this method absolutely works and gives you 100-150% efficiency over time.

Cherrypicker:


vs

Bithopper:


I've just disabled slush for now.  Been a boring day for hopping so far.  Been on Arsbitcoin pretty much all day.

[ampirebus]

what's your estimated payout at mtred? what's your balance at eligius?

slush is score based and only experimentally hoppable, btw. You shouldn't expect big payouts from slush.

well compared to not hopping my payouts are pretty much 0. my balance at eligus/mtred combined is .18

im not expecting big payouts from slush, just payouts. every single round on slush if i were not hopping i would get .01~.02. going from .02 to .00002

[hawks5999]

I have to say I've been a little nervous about all the additional small pools (even though I requested them  ). It seemed like too much time was going to bithasher.

But I'll be danged. We just solved a block on bithasher while picking (534951 total shares) and now I have 1.73 BTC waiting for confirm. That should keep my daily moving average happy 

[muyoso]

it completely avoided all the best rounds

there is no way to know what rounds will be the best because CP cant predict the future so mining in short rounds is always a tricky thing
a task was open for this and maybe a child score profile for short rounds can fix your needs

Thats why you mine all rounds for no more than 10%.

Take these rounds for example:



Each round had ~55 shares submitted.

7140 - 201.79% efficiency
7139 - 1201.46% efficiency
7138 - 422.01% efficiency

So on the longer rounds when you get nothing, these go a long way to keeping overall efficiency high.

[muyoso]

I have to say I've been a little nervous about all the additional small pools (even though I requested them  ). It seemed like too much time was going to bithasher.

But I'll be danged. We just solved a block on bithasher while picking (534951 total shares) and now I have 1.73 BTC waiting for confirm. That should keep my daily moving average happy 

Damn you, just made me add ANOTHER pool.  Ive had like 15 tabs open just for pools for like a month now.