[4+ EH] Slush Pool (slushpool.com); Overt AsicBoost; World First Mining Pool

[riush]

That would seem to suggest that the estimated reward on the profile page is only updated when you submit a new share.

It was falling before it found the hash, from >9 to ~5 in a few minutes.

[Beremat]

argh!

[eMansipater]

Hey slush, just wanted to say major kudos for your work on the score-based system.  It's an excellent balance between preventing cheating and maintaining low variance, with stats all beautifully available to boot.  Like many others I was getting a huge proportion of invalids so I've switched to solo to try and ease the load on your server until you can get more power online, but I'll definitely be back once everything is stabilised.  Great job!

eMansipater

[jakakxl]

[20:04:36] <mmarker> His stats
[20:04:42] <Mango-chan> how do you know
[20:04:48] <mmarker> Don't make no sense
[20:04:52] <andrew12> Mango-chan: >I think
[20:05:00] <andrew12> mmarker: neither does what you just said
[20:05:00] <Mango-chan> so where is he
[20:05:03] <mmarker> We're on block 109051
[20:05:15] <mmarker> block 109050 was just found
[20:05:37] <mmarker> and 109051 was just found as well
[20:05:44] * malfy (~malfy@unaffiliated/malfy) has joined #bitcoin-dev
[20:05:51] <mmarker> at 3:49:28 UTC

[geebus]

With the pool being split across two servers, running two bitcoind instances, what is preventing them from issuing the same (or similar) getwork to two different people? ...and are they collectively reporting shares together? How is the load being distributed?

Basically what I'm asking is...

If Miner A is connected to Server A, and Miner B is connected to Server B, is the pool looking at each round as (Shares from Server A) + (Shares from Server B) = Total Shares?

...and if it's two different bitcoind instances, aren't the two servers basically competing against each other?

[demonofelru]

I know it's probably just terrible luck but over 6 hours is very frustrating.  At average of 60 gh/s wouldn't that be like 99.9% probability or something?

[Miner-TE]

I think the block moved on and something is stuck.

[demonofelru]

It seems people are leaving the pool because of this it was about 62 ghash/s now 50, not sure if that's a good or bad sign.

[Miner-TE]

I'll stick it out.   

I'm also waiting on a payout.  My confirmed is above my threshold and has been for awhile now.  Maybe the payment script has not run yet.

[BiggieJohn]

I just started mining yesterday, so I'm not too sure how this all works yet. 
Is this the first time we ever made it to 99.99% ??

[Xerol]

I'll stick it out.   

I'm also waiting on a payout.  My confirmed is above my threshold and has been for awhile now.  Maybe the payment script has not run yet.

Yeah, at this point anyone who leaves is going to miss out on coins when it finally hits, and anyone who sticks around is going to get more than their average. Right now I'm sitting around 0.045 for this block, usually I average around 0.015. This is with 27mhash/s on a HD 4670.

[demonofelru]

Does Current shares CFD: 99.98% mean that there was a 99.98% chance we were supposed to hit by now?  Looked at explanation that links to wikipedia but don't understand?

[riush]

The score and shares just reset. Last time i checked:

Total score of current round:   442423595.2236
Shares contributed in current round:    304454

now:

Total score of current round:   6541.9620
Shares contributed in current round:   307685

Current round duration:   6:28:05

also my estimated reward went up to >4 BTC again, quickly falling now.
nothing added to unconfirmed reward.

[schnak]

slush needs to reboot the pool box i think.... this would be the problem with being in a pool.... if it stops, you stop too.

comeon slush just poke it with a stick and it'll start moving again

[FooDSt4mP]

I think there might be an overflow error with the scoring...  Look at the total score in the following vs. the contributed shares and CFD...

Well, this is not a bug, this is feature :-). I didn't want to complicate the stuff with score based system in the introduction, but as you are asking...

Using exp(round_time/C) can lead to extremely high numbers. Such big that with many workers and extremely bad pool luck, it may overflow somewhere. So in reality, the score based system is implemented with one optimization, that was not explained yesterday:

The real formula isnt score += exp(round_time/C), but:

Code:

duration = min(round_time, hour_time)
score += exp(duration/C)

where round_time is count of seconds from round beginning and hour_time is count of seconds in current hour.

And every hour, exactly at 0 minute, 0 second +- 50 miliseconds, I'm performing "renormalization". This means that all worker scores are divided by exp(min(3600, round_time)/C).

This has NO IMPACT on score system itself, but helps handling huge numbers in application. So your observation was absolutely correct, because every hour the system score drop to very low number.

The score and shares just reset. Last time i checked:

Total score of current round:   442423595.2236
Shares contributed in current round:    304454

now:

Total score of current round:   6541.9620
Shares contributed in current round:   307685

also my estimated reward went up to >4 BTC again, quickly falling now.

Scores are renormalized every hour to prevent overflow.

[demonofelru]

slush needs to reboot the pool box i think.... this would be the problem with being in a pool.... if it stops, you stop too.

comeon slush just poke it with a stick and it'll start moving again

Unfortunately depending on his time zone it might be another half a day too bad I'm too lazy to get off my but and switch to solo lol.

[schnak]

slush needs to reboot the pool box i think.... this would be the problem with being in a pool.... if it stops, you stop too.

comeon slush just poke it with a stick and it'll start moving again

Unfortunately depending on his time zone it might be another half a day too bad I'm too lazy to get off my but and switch to solo lol.

feature request - if its been more than 3x average time for a winning block, slush's computers all start blaring klaxons... thatll get his attention 

[riush]

Scores are renormalized every hour to prevent overflow.

ah, i see.
but does this explain the high estimated reward? if the pool were to find a block now would i receive that amount?
if so, it must be to the loss of someone finding shares before the renormalization? or what am i missing here?

[demonofelru]

slush needs to reboot the pool box i think.... this would be the problem with being in a pool.... if it stops, you stop too.

comeon slush just poke it with a stick and it'll start moving again

Unfortunately depending on his time zone it might be another half a day too bad I'm too lazy to get off my but and switch to solo lol.

feature request - if its been more than 3x average time for a winning block, slush's computers all start blaring klaxons... thatll get his attention 

Hehe 50 BTC bounty here.

[Beremat]

The server's officially exploded: