[500 GH/s]HHTT -Selected Diff/Stratum/PPLNS/Paid Stales/High Availability/Tor - page 32.

fireduck

sr. member

Activity: 392

Merit: 251

Quote from: crazyates on August 29, 2012, 01:00:40 PM

Is it even possible to mine at 1GH/s, and all diff>=32 get submitted to your pool, and any diff<32 get submitted to another pool?

Negative. For the same reason that it is impossible to mine on a pool and take any found blocks yourself.

When mining you are doing hash(DATA_FROM_POOL + NONCE + TIME). Your client changes the nonce and to a certain extent the time. The DATA_FROM_POOL includes a hash of a transaction group that includes a payment to the pool. You can't change that without invalidating the work. Pools also check that the work unit you are submitting is one they gave out.

crazyates

legendary

Activity: 952

Merit: 1000

Quote from: fireduck on August 29, 2012, 10:31:00 AM

Quote from: cmg5461 on August 29, 2012, 10:23:59 AM

Just a thought, but the shares people produce are random. Some shares can be seen as > 32 difficulty while being mined on 1 difficulty. What's to say the user can't submit false D data based on the share they produce?

Say you produce one share every second at 1 D with 1 GH

while someone with xx Gh produces 1 share every second at 32 D

a few of those shares for the 1 gh will be > 1 difficulty, so they can be seen as mined with a higher difficulty, falsely representing their true hash rate.

In the case of my pool, the difficulty is configured and used for everyone. So iff a share is good enough for difficulty 32 it gets accepted. If someone were to setup dynamic difficulty or user selectable difficulty they would have to track which work they gave at what difficulty to avoid the problem you are describing. That shouldn't be hard, since the pool needs to track which work it has issued anyways.

Is it even possible to mine at 1GH/s, and all diff>=32 get submitted to your pool, and any diff<32 get submitted to another pool?

Lethos

sr. member

Activity: 476

Merit: 250

Keep it Simple. Every Bit Matters.

Your going to have to explain that magic pony...

fireduck

sr. member

Activity: 392

Merit: 251

Quote from: Lethos on August 29, 2012, 11:59:41 AM

Well since the password can be anything at the moment, if a user puts it down as d20, that could be the difficulty they get given, if it's something that doesn't look like a difficulty request you'd use a default.

That is a good idea and for some parts might be easier since then the username doesn't need to be manipulated. I might make the password be a HTTP style parameter string in case more things need to be added.

Something like:
difficulty=20&magic_pony=true&whatever=something

Lethos

sr. member

Activity: 476

Merit: 250

Keep it Simple. Every Bit Matters.

Quote from: fireduck on August 29, 2012, 10:34:56 AM

Quote from: Meni Rosenfeld on August 29, 2012, 10:29:39 AM

The share difficulty needs to be decided in advance. If a miner configured as a 32 miner submits a 1 share it will be rejected. If a miner configures as 1 submits a 32 share it will count just like a normal 1 share.

Yep. I was thinking about doing something where the user specifies the difficulty in the username, so it would be address_difficulty. That way miners can decide what difficulty made sense for their rigs. Then I would also define a sliding scale where higher difficulty means less fees. This would allow me to encourage selecting a higher difficulty which reduces DB strain on my end.

Well since the password can be anything at the moment, if a user puts it down as d20, that could be the difficulty they get given, if it's something that doesn't look like a difficulty request you'd use a default.

fireduck

sr. member

Activity: 392

Merit: 251

Quote from: Meni Rosenfeld on August 29, 2012, 10:29:39 AM

The share difficulty needs to be decided in advance. If a miner configured as a 32 miner submits a 1 share it will be rejected. If a miner configures as 1 submits a 32 share it will count just like a normal 1 share.

Yep. I was thinking about doing something where the user specifies the difficulty in the username, so it would be address_difficulty. That way miners can decide what difficulty made sense for their rigs. Then I would also define a sliding scale where higher difficulty means less fees. This would allow me to encourage selecting a higher difficulty which reduces DB strain on my end.

fireduck

sr. member

Activity: 392

Merit: 251

Quote from: cmg5461 on August 29, 2012, 10:23:59 AM

Just a thought, but the shares people produce are random. Some shares can be seen as > 32 difficulty while being mined on 1 difficulty. What's to say the user can't submit false D data based on the share they produce?

Say you produce one share every second at 1 D with 1 GH

while someone with xx Gh produces 1 share every second at 32 D

a few of those shares for the 1 gh will be > 1 difficulty, so they can be seen as mined with a higher difficulty, falsely representing their true hash rate.

In the case of my pool, the difficulty is configured and used for everyone. So iff a share is good enough for difficulty 32 it gets accepted. If someone were to setup dynamic difficulty or user selectable difficulty they would have to track which work they gave at what difficulty to avoid the problem you are describing. That shouldn't be hard, since the pool needs to track which work it has issued anyways.

Meni Rosenfeld

donator

Activity: 2058

Merit: 1054

Quote from: cmg5461 on August 29, 2012, 10:23:59 AM

Just a thought, but the shares people produce are random. Some shares can be seen as > 32 difficulty while being mined on 1 difficulty. What's to say the user can't submit false D data based on the share they produce?

Say you produce one share every second at 1 D with 1 GH

while someone with xx Gh produces 1 share every second at 32 D

a few of those shares for the 1 gh will be > 1 difficulty, so they can be seen as mined with a higher difficulty, falsely representing their true hash rate.

The share difficulty needs to be decided in advance. If a miner configured as a 32 miner submits a 1 share it will be rejected. If a miner configured as 1 submits a 32 share it will count just like a normal 1 share.

cmg5461

sr. member

Activity: 369

Merit: 250

Just a thought, but the shares people produce are random. Some shares can be seen as > 32 difficulty while being mined on 1 difficulty. What's to say the user can't submit false D data based on the share they produce?

Say you produce one share every second at 1 D with 1 GH

while someone with xx Gh produces 1 share every second at 32 D

a few of those shares for the 1 gh will be > 1 difficulty, so they can be seen as mined with a higher difficulty, falsely representing their true hash rate.

Lethos

sr. member

Activity: 476

Merit: 250

Keep it Simple. Every Bit Matters.

Quote from: fireduck on August 29, 2012, 09:49:06 AM

Quote from: Lethos on August 29, 2012, 09:21:44 AM

Just testing out your Pool, Fireduck.
Surprised it's still this small, already got my first payments came through.

Is 32 difficulty really that scary to everyone?

It is probably my terrible web page, that I call all the users suckers on the web page, that the hash rate is low and I don't offer merged mining. All these things add up. I imagine I'll get more traffic once the ASICs start hitting but I'd be surprised if other pools didn't have a way to increase their difficulty by then as well.

I've been mining long enough to realise merged mining means very very little difference to overall bitcoins I end up getting. Merged mining just makes it so that all these other bitcoin copies end up staying around longer than they should.

What I saw as nice was a PPS rate that is done exactly like everyone else, but with only a 2% fee, which is quiet tempting and to be honest, not having to register an account to mine their made the process really a lot easier.

Sure you could give it a face lift, but if you don't it's not the end of the world. I just want to see how my FPGA's handle doing a higher difficulty.

fireduck

sr. member

Activity: 392

Merit: 251

Quote from: Lethos on August 29, 2012, 09:21:44 AM

Just testing out your Pool, Fireduck.
Surprised it's still this small, already got my first payments came through.

Is 32 difficulty really that scary to everyone?

It is probably my terrible web page, that I call all the users suckers on the web page, that the hash rate is low and I don't offer merged mining. All these things add up. I imagine I'll get more traffic once the ASICs start hitting but I'd be surprised if other pools didn't have a way to increase their difficulty by then as well.

Lethos

sr. member

Activity: 476

Merit: 250

Keep it Simple. Every Bit Matters.

Just testing out your Pool, Fireduck.
Surprised it's still this small, already got my first payments came through.

Is 32 difficulty really that scary to everyone?

fireduck

sr. member

Activity: 392

Merit: 251

I've thought about user selected difficulties, maybe with a lower pool fee the higher the difficulty which gives people incentive to reduce database work for the pool. I think that makes a lot of sense but as I don't know pushpoold code base very well it will probably take me some time to work out.

As an aside, apparently long polling works just fine. Just not via the proxy I was using with my miners so I thought it didn't work.

coinotron

legendary

Activity: 1182

Merit: 1000

Quote from: organofcorti on July 27, 2012, 01:49:02 AM

This is a truly good idea - higher pool D is going to be necessary at some point soon. However, if you can make D user defined, then low hashrate will be less affected by variance - at 200Mhps, the standard deviation in share submission would be about 9% of the mean when going from D1 to D 32, but only 2.8% of the mean at 2Ghps. Sections 7.5 and 7.6 of Meni Rosenfeld's AoBPMRS covers this idea in more detail. While 200 Mhps might not be profitable soon, you might also have to increase the pool D, so this could be a good idea regardless.

Well done on being the first! (afaik)

Good idea.
Just FYI, most LTC pools have implemented this idea for months. But very first implementation was (AFAIK) in Solidcoin 2.0 in November 2011.

organofcorti

donator

Activity: 2058

Merit: 1007

Poor impulse control.

This is a truly good idea - higher pool D is going to be necessary at some point soon. However, if you can make D user defined, then low hashrate will be less affected by variance - at 200Mhps, the standard deviation in share submission would be about 9% of the mean when going from D1 to D 32, but only 2.8% of the mean at 2Ghps. Sections 7.5 and 7.6 of Meni Rosenfeld's AoBPMRS covers this idea in more detail. While 200 Mhps might not be profitable soon, you might also have to increase the pool D, so this could be a good idea regardless.

Well done on being the first! (afaik)

fireduck

sr. member

Activity: 392

Merit: 251

Payouts are working and detail pages are up for users.

fireduck

sr. member

Activity: 392

Merit: 251

I'd like to announce the Horrible Horrendous Terrible Tremendous Mining Pool

http://hhtt.1209k.com/

The special feature here is higher difficulty work units to reduce network chatter. It uses difficulty 32 work units (for now). Getting ready for higher hash rate butterflies in October.

If you have a 1GH/s setup now and see the shares scrolling by at one every few seconds, imagine what it will be like with a 40GH/s BFL SC Single.

There is simply no need for all that traffic. The time when difficulty one pool proof of works made sense is over.

Topic: [500 GH/s]HHTT -Selected Diff/Stratum/PPLNS/Paid Stales/High Availability/Tor - page 32. (Read 56478 times)