kano (OP)
Legendary
Offline
Activity: 4536
Merit: 1847
Linux since 1997 RedHat 4
|
|
February 27, 2016, 10:52:23 AM |
|
As many would have noticed, I also did a ckdb restart in there.
There was a syncronisation issue between ckdb and ckpool so restarting ckdb resets the synchronisation and ensures no data was lost
ckdb tracks every record and every type of record it gets from ckpool since they all have continuous sequence numbers - so with the restart, as usual, I see no sequence problems so I know the synchronisation is all correct and zero data has been skipped/lost. (a ckdb restart goes back and reprocesses from the last completed/summarised shift, reading the last hour or more, in this case 40 minutes of shift data back before the ckpool restart, so it can be sure it's all ok)
|
|
|
|
StevenD
Newbie
Offline
Activity: 32
Merit: 0
|
|
February 27, 2016, 12:16:13 PM Last edit: February 27, 2016, 12:26:36 PM by StevenD |
|
Hi
Kano, Apologies if this turns out to be the most stupid post of the month so far.
I am just getting my head around how the pool pays by PPLNS, which while I have mined here I have seen the 5Nd period drop from 192hrs to 34hrs as the pool grows.
Never really considered this before today, because I haven't had to sat through a tough block that exceeded the 5Nd period, but I think this block hrs will be greater than the expected 5Nd payment period, which if I understand it, means were only going to get 1/5th of the payment for the last 36+ hrs, and we will loose out on any % payments that we didn't already receive from the previous found blocks, so that for some of the previous blocks we will loose 80%, 60% or 40% of the expected payments should we have continued at the expected 100% rather than this maddening 480% + block that were struggling with.
The only good thing that I can see is that the previous found blocks which we will loose the payments were all quite close together so the loss on previous payments is not so bad, but if we have another bad block as the next one, then we would be loosing 80% of a lot of hours work from this current block.
Is there a point when a pool hash rate gets larger, 20PH+ or whatever, when it hits a point when the 5Nd from the PPLNS actually works against the miners, because the smaller and smaller hrs of the 5Nd period and any bad luck wipes out any benefit of trying to even out the payments by using the 5Nd, is there a point when a pool gets to a larger size to actually run PPS, so to pay for the work done as its done, and then no-one ever looses out on the bad blocks that can over-run the 5Nd period and cause the loss of previous work done.
is PPLNS always best for any sized pool, or is PPS better for a certain sized pools.
If this is a stupid question just delete it
Steve
|
|
|
|
philipma1957
Legendary
Offline
Activity: 4172
Merit: 8061
'The right to privacy matters'
|
|
February 27, 2016, 01:00:02 PM |
|
Hi
Kano, Apologies if this turns out to be the most stupid post of the month so far.
I am just getting my head around how the pool pays by PPLNS, which while I have mined here I have seen the 5Nd period drop from 192hrs to 34hrs as the pool grows.
Never really considered this before today, because I haven't had to sat through a tough block that exceeded the 5Nd period, but I think this block hrs will be greater than the expected 5Nd payment period, which if I understand it, means were only going to get 1/5th of the payment for the last 36+ hrs, and we will loose out on any % payments that we didn't already receive from the previous found blocks, so that for some of the previous blocks we will loose 80%, 60% or 40% of the expected payments should we have continued at the expected 100% rather than this maddening 480% + block that were struggling with.
The only good thing that I can see is that the previous found blocks which we will loose the payments were all quite close together so the loss on previous payments is not so bad, but if we have another bad block as the next one, then we would be loosing 80% of a lot of hours work from this current block.
Is there a point when a pool hash rate gets larger, 20PH+ or whatever, when it hits a point when the 5Nd from the PPLNS actually works against the miners, because the smaller and smaller hrs of the 5Nd period and any bad luck wipes out any benefit of trying to even out the payments by using the 5Nd, is there a point when a pool gets to a larger size to actually run PPS, so to pay for the work done as its done, and then no-one ever looses out on the bad blocks that can over-run the 5Nd period and cause the loss of previous work done.
is PPLNS always best for any sized pool, or is PPS better for a certain sized pools.
If this is a stupid question just delete it
Steve
If you do PPLNS anytime you rip back to back cdfs of 99% for your blocks you are fucked.. But just for the two blocks. The key is think in 100 block segments not 2 block segments. What is the record for the last 100 blocks? 116.05% now if you did PPS at a 99% payrate for those 100 blocks you would have lost out on the 17% over the 99%.
|
|
|
|
wolfen
|
|
February 27, 2016, 01:04:37 PM |
|
I need a drink Maybe some Valium.
|
For those about to block we salute you! AC->BTC
|
|
|
sorry2xs
Legendary
Offline
Activity: 924
Merit: 1000
Dark Passenger Bitcoin miner 2013,Bitcoin node
|
|
February 27, 2016, 01:06:13 PM |
|
i nice Chaos Calmer will do
|
Please tip the Node 1MPWKB23NsZsXHANnFwVAWT86mL24fqAjF; KO4UX THAT NO GOOD DO GOODER BAT!!!
|
|
|
philipma1957
Legendary
Offline
Activity: 4172
Merit: 8061
'The right to privacy matters'
|
|
February 27, 2016, 01:06:40 PM |
|
I need a drink Maybe some Valium. Well Yeah we are at 496% which is over a cdf of 99%. This is the 2nd worst block in history of the pool. I have some high quality vodka that I dropped 6 vanilla sticks into. Best vanilla vodka you could ever have.
|
|
|
|
DarKSm0ke
|
|
February 27, 2016, 01:28:20 PM |
|
BLOCK
|
|
|
|
Jake36
|
|
February 27, 2016, 01:30:35 PM |
|
BLOCK Finally! Now "WE" can get on with Block Saturday, to make up for a non-block Friday.. And with 0.55687924 BTC in fee's!
|
|
|
|
ZACHM
|
|
February 27, 2016, 01:31:15 PM |
|
There it is.
|
|
|
|
hoosier_13
Member
Offline
Activity: 62
Merit: 10
|
|
February 27, 2016, 01:31:53 PM |
|
Thank God, was getting worried
|
Bitrated user: TICH13.
|
|
|
philipma1957
Legendary
Offline
Activity: 4172
Merit: 8061
'The right to privacy matters'
|
|
February 27, 2016, 01:32:01 PM |
|
Hi
Kano, Apologies if this turns out to be the most stupid post of the month so far.
I am just getting my head around how the pool pays by PPLNS, which while I have mined here I have seen the 5Nd period drop from 192hrs to 34hrs as the pool grows.
Never really considered this before today, because I haven't had to sat through a tough block that exceeded the 5Nd period, but I think this block hrs will be greater than the expected 5Nd payment period, which if I understand it, means were only going to get 1/5th of the payment for the last 36+ hrs, and we will loose out on any % payments that we didn't already receive from the previous found blocks, so that for some of the previous blocks we will loose 80%, 60% or 40% of the expected payments should we have continued at the expected 100% rather than this maddening 480% + block that were struggling with.
The only good thing that I can see is that the previous found blocks which we will loose the payments were all quite close together so the loss on previous payments is not so bad, but if we have another bad block as the next one, then we would be loosing 80% of a lot of hours work from this current block.
Is there a point when a pool hash rate gets larger, 20PH+ or whatever, when it hits a point when the 5Nd from the PPLNS actually works against the miners, because the smaller and smaller hrs of the 5Nd period and any bad luck wipes out any benefit of trying to even out the payments by using the 5Nd, is there a point when a pool gets to a larger size to actually run PPS, so to pay for the work done as its done, and then no-one ever looses out on the bad blocks that can over-run the 5Nd period and cause the loss of previous work done.
is PPLNS always best for any sized pool, or is PPS better for a certain sized pools.
If this is a stupid question just delete it
Steve
If you do PPLNS anytime you rip back to back cdfs of 99% for your blocks you are fucked.. But just for the two blocks. The key is think in 100 block segments not 2 block segments. What is the record for the last 100 blocks? 116.05% now if you did PPS at a 99% payrate for those 100 blocks you would have lost out on the 17% over the 99%. To follow up f2pool pays 96% not 99% and f2pool takes the tx id fees. while give some NMC nets to 96% so over the last 100 blocks so if 100% was 1 btc you would get .96 btc from f2pool maybe .97 btc kano you would get 1.16 btc maybe 1.164 btc since he gives the tx id fees. so .97 to 1.16 is a good estimate for a 100 block period. But for this block right now PPS is better. Small samples of 1, 2, 3, 4 blocks do not mean much. a cdf of .993 second best ever bitminerpro_S3R.2059A0C51E43 thanks hey was this an s3?
|
|
|
|
sorry2xs
Legendary
Offline
Activity: 924
Merit: 1000
Dark Passenger Bitcoin miner 2013,Bitcoin node
|
|
February 27, 2016, 01:34:25 PM |
|
church lady to pool Well aint that special
|
Please tip the Node 1MPWKB23NsZsXHANnFwVAWT86mL24fqAjF; KO4UX THAT NO GOOD DO GOODER BAT!!!
|
|
|
generalt
Legendary
Offline
Activity: 1096
Merit: 1021
|
|
February 27, 2016, 01:35:23 PM |
|
BLOCK That was a tough one at 502.797%
|
BTC: 1GENERALrtBAjEv2Ps5cmEW1FADnXh1bCZ
|
|
|
sorry2xs
Legendary
Offline
Activity: 924
Merit: 1000
Dark Passenger Bitcoin miner 2013,Bitcoin node
|
|
February 27, 2016, 01:38:55 PM |
|
it was a bit of a drain i say
|
Please tip the Node 1MPWKB23NsZsXHANnFwVAWT86mL24fqAjF; KO4UX THAT NO GOOD DO GOODER BAT!!!
|
|
|
philipma1957
Legendary
Offline
Activity: 4172
Merit: 8061
'The right to privacy matters'
|
|
February 27, 2016, 01:41:36 PM |
|
we need a bang, bang, bam in the next hour or two.
|
|
|
|
wolfen
|
|
February 27, 2016, 01:42:27 PM |
|
AAAAAAAAAAAAAAAAAAAAAAAAAAHHHHHH i'll take it.
|
For those about to block we salute you! AC->BTC
|
|
|
xgeniy
Newbie
Offline
Activity: 18
Merit: 0
|
|
February 27, 2016, 01:43:32 PM |
|
go go kano! let's more green block)))
|
|
|
|
sorry2xs
Legendary
Offline
Activity: 924
Merit: 1000
Dark Passenger Bitcoin miner 2013,Bitcoin node
|
|
February 27, 2016, 01:48:22 PM |
|
|
Please tip the Node 1MPWKB23NsZsXHANnFwVAWT86mL24fqAjF; KO4UX THAT NO GOOD DO GOODER BAT!!!
|
|
|
WBF1
|
|
February 27, 2016, 01:50:43 PM |
|
Hi
Kano, Apologies if this turns out to be the most stupid post of the month so far.
I am just getting my head around how the pool pays by PPLNS, which while I have mined here I have seen the 5Nd period drop from 192hrs to 34hrs as the pool grows.
Never really considered this before today, because I haven't had to sat through a tough block that exceeded the 5Nd period, but I think this block hrs will be greater than the expected 5Nd payment period, which if I understand it, means were only going to get 1/5th of the payment for the last 36+ hrs, and we will loose out on any % payments that we didn't already receive from the previous found blocks, so that for some of the previous blocks we will loose 80%, 60% or 40% of the expected payments should we have continued at the expected 100% rather than this maddening 480% + block that were struggling with.
The only good thing that I can see is that the previous found blocks which we will loose the payments were all quite close together so the loss on previous payments is not so bad, but if we have another bad block as the next one, then we would be loosing 80% of a lot of hours work from this current block.
Is there a point when a pool hash rate gets larger, 20PH+ or whatever, when it hits a point when the 5Nd from the PPLNS actually works against the miners, because the smaller and smaller hrs of the 5Nd period and any bad luck wipes out any benefit of trying to even out the payments by using the 5Nd, is there a point when a pool gets to a larger size to actually run PPS, so to pay for the work done as its done, and then no-one ever looses out on the bad blocks that can over-run the 5Nd period and cause the loss of previous work done.
is PPLNS always best for any sized pool, or is PPS better for a certain sized pools.
If this is a stupid question just delete it
Steve
Short answer is PPLNS is better than PPS because PPS is capped. For example, if you mine at a pool with 5% fee, paying PPS, the most you will ever make compared to expected payout is 95%... even if thepool hits a string of blocks at 50% dificulty (which means the pool income is roughly 200% expected income but you only get 95%). Works the other way too where a string of 200% difficulty blocks has the pool at roughly 50% expected income but you get 95%. PPS will always have higher fee to make up for variance. In the above scenario, on PPLNS, you get 200% of expected income when the pool hits a string of fast blocks at 50% difficulty and 50% expected income when with a string of 200% blocks. BUT PPLNS smooths out this variance using a window of size N to look backwards and see what portion of the payout you get. The only time work isever unpaid is when work on one block goes over 500% (because N here is 5d where d is difficulty). This hasn't happened very often. And like phillipma1957 said, you have to look at it over a longer time span. At time spans of 7 days or more, this pool is over 100%.
|
|
|
|
dontetris
Newbie
Offline
Activity: 47
Merit: 0
|
|
February 27, 2016, 01:51:11 PM |
|
And there it is, 502% diff
|
|
|
|
|