bitHopper: Python Pool Hopper Proxy

[Sukrim]

Anyways, I think backup_latehop is ready for merging. Pull request sent.

Next on the list: Nicer graphs.
Edit:
Nah, I hate javascript - scratch that!

[MaGNeT]

PPS / Pushpool / 23Ghash

https://www.ninjacoin.com/

[jkminkov]

PPS / Pushpool / 23Ghash

https://www.ninjacoin.com/

3% for PPS is unsustainable...

[MaGNeT]

PPS / Pushpool / 23Ghash

https://www.ninjacoin.com/

3% for PPS is unsustainable...

For pool or hopper?

[Sukrim]

Added BTCPool24.com

Another new pool with prop and they even claim to also have instant pay 0% PPS.

They seem to charge a 0.01 transaction fee on payouts though... (never heard of "sendmany", eh?)

[Disposition]

* Individual Miner Stats. Based on username

The lack individual miner tracking has been putting me off from using this hopper, now I see this feature, I have a few questions.

how are the user name determined? do you connect to localhost:8337 using any username and password?

thanks.

[paraipan]

...................................................................

Absolutely perfect! I would have never been able to figure that one out. Thanks.

I've created a fork on github. Everybody check out my new stats layout / color scheme. (I also changed and capitalized the pool names in the pool.cfg and user.cfg.default, so just replace those files if you don't want to deal with that.)

https://github.com/macboy80/bitHopper

EDIT: Just updated to comment out disabled pools in pool.cfg and user.cfg.default (Removes disabled pools from the stats) index.html should copy over to latest c00w build without any changes

make a pull request pls, would love to have that

[paraipan]

* Individual Miner Stats. Based on username

The lack individual miner tracking has been putting me off from using this hopper, now I see this feature, I have a few questions.

how are the user name determined? do you connect to localhost:8337 using any username and password?

thanks.

you can use bitcoin-mining-proxy in between miner and hopper proxy for that purpose

[Disposition]

you can use bitcoin-mining-proxy in between miner and hopper proxy for that purpose

so what does this individual miner status show? do the arthor mean per pool? or is there a flexable mining proxy like feature that's built in already.


also what information should you connect to the hopper by? e.g. the username and password

[paraipan]

you can use bitcoin-mining-proxy in between miner and hopper proxy for that purpose

so what does this individual miner status show? do the arthor mean per pool? or is there a flexable mining proxy like feature that's built in already.


also what information should you connect to the hopper by? e.g. the username and password

ah ok, connect putting any username (ex: x) and any pass (ex: y)

edit: you will have to provide usernames and passwords for every pool in "user.cfg"

i don't understand the "individual miner" part though

[pbj sammich]

i don't understand the "individual miner" part though

perhaps he is thinking about separate worker stats instead of miners?

[Sukrim]

You get the total shares submitted (I think including stale/invalid) on the stats page per worker (whatever you enter as "user" in your miner when connecting to bitHopper), that's all currently.

Might be enough if you have one proxy you share with friends, or to see how many shares have been submitted by one machine, but nothing very sophisticated (yet). If you need something specific, feel free to code it in - I doubt though that there will be many developments in this direction in the near future... if you really need these stats you can anyways use a second proxy and if several people want to poolhop, they can all just use local proxys.

[c00w]

I plan on adding rejects and last time seen. Whats in the database is the shares for each worker and where they were submitted currently.

[paraipan]

@sukrim would you mind coding some "removefromlist" filter if username and pass is not found in user.cfg for a certain pool ?
I just removed accidentally some pool credentials and was getting parse error not knowing why. Thanks

[organofcorti]

I wrote a monte carlo sim for hopping and now we can hopefully put the whole mine_charity thing to bed.

If you hop on at 2 * difficulty, your efficiency never goes over about 0.83.

I'll post chartporn tomorrow and if there are any requests I'll put the sim on github.

[Sukrim]

@para:
Yeah, I'd actually like to display a more descriptive error like "user.cfg incomplete for [pool], disabling [pool]".
Not at the moment though, but within 24 hours, ok?

@organo:
Do you haveby chance monte carlo sims with prop pools to be hopped and scored (ideally) or at least PPLNS pools as backup_latehop (on backup pools, select the one with the most shares currently)?

Even though I should know better, it still "feels" to bring potentially more output if you jump in on long rounds in these payout methods, as at least scored pays for sure more if you get in late (though you never know when "late" happens).

[organofcorti]

@organo:
Do you haveby chance monte carlo sims with prop pools to be hopped and scored (ideally) or at least PPLNS pools as backup_latehop (on backup pools, select the one with the most shares currently)?

Even though I should know better, it still "feels" to bring potentially more output if you jump in on long rounds in these payout methods, as at least scored pays for sure more if you get in late (though you never know when "late" happens).

I wrote the sims from scratch to make 2 and 3d graphs of hop on v hop off v efficiency, so it's not modular. The reason i wrote them was to show that late hop on always reduces efficiency to below 1.0 . No one was following the math and everyone follows charts!

So I could write a sim for pplns to provide a similar graph of  hop on v hop off v efficiency if people want one, but for scored pools I'd need the algo for weighting shares and I don't think slush will make that happen in a hurry.

BTW, for me slush pays out well if you start at 0 and leave v early.

[Sukrim]

Slush is known to be broken, Meni's scoring claims to be hopping proof though. No worries, it's not that urgent and theoretically you should anyways just find out that it really only gives an expected value of 100%.

[organofcorti]

Slush is known to be broken

...and thats why i'd love to get my hands on his algo so i can find the max efficiency hop on/off points accurately.

I won't bother with Meni's. All those zeros would have my little macbook air churning for months! Plus I have no idea how to access opencl in R. But pplns might hold a few interesting surprises.

[joulesbeef]

Slush is known to be broken

what exactly does mine_slush do? should I be mining them? I seem to be doing well with slush, but I know this could be luck.

Yeha the more I thought about mine_charity the more I thought it was the exact opposite of hopping and could possible lead to eff below 100%, I'm glad someone else did the math.