Topic: [9 TH] Bitparking Pool, DGM 0%,vardiff,stratum,Merge Mining - page 81. (Read 163780 times)

Well that was fun - unexpected pool downtime for 20 minutes or so. I've temporarily disabled DGM stats collection which was causing score updates to fall behind.

I installed a queuing system for updating share data recently to help with performance. Unfortunately this resulted in hash rate estimates to be under-reported in times when the pool comes under load and the queue builds up. I'll fix this in the next update, sorry.

The pool hash rate has picked up again, ranging 200-300+ at the moment with some miners that mine for a time period then leave. I'm wondering if it's the result of being a backup pool for people that kicks in when their main pool has issues.

Is it just me, or do personal hash rate estimates on the pool seem kind of low/fluctuating right now?

BTW, awesome pool DoubleC, thanks!

Luckily we had a few good short blocks earlier so it balances out. The pool ran really bad in the early days, and someone ripped the pool off for about 300 btc back then after finding a flaw, so it's still catch up. In total it's 800 btc behind or so. Variance is crazy. It an easily be +-600 btc at any time.

I haven't decided on whether to go DGM yet - I'm just running stats for curiosities sake for now. If I did I'd do them side by side.

Ouch, I bet that hurt you?


Question, when your implement DGM will PPS stay and run side by side?

That was a horrible block at 6,907,591 shares. Hopefully we don't get too many of those!

There was a few seconds of downtime a short time ago as I installed some pool updates to help with performance.

Not that I'm aware of. You'd have to ask Vircurex.

Can i get the vircurex discount if i've already made an account with them? without making a new one using the register link...

And that's exactly why I don't like DGM: it has variance. During periods of bad luck it can be very frustrating! I prefer the nice and steady payouts of PPS, even if it returns a little bit less over the long term. Another reason is that DGM was designed to discourage pool hoppers. That means it's not a good idea to set a DGM pool as backup because that makes you a pool hopper and you get paid less. Or if you have to turn off your rig for a little while to do maintenance, when you switch it back on you get paid less for a while than what you would get with PPS.

PPS is so much more user friendly.

With 11 blocks solved since collecting DGM stats the numbers are interesting. So far DGM would have paid out 140+ bitcoins more than PPS. This is a bit of a skewed view though since the pool has been running well with no blocks with scores greater than difficulty since I started running the stats. I'm sure it'll balance out when that happens. Pity I didn't run the stats two weeks earlier when the pool was running bad!

Interesting study. My numbers end up very similar, but then I am recently returned to mining here, so hard to consider this a long term picture. Nice to see some creative thinking with the pool.

I've added a user id 'estimated payout' portion of http://mmpool.bitparking.com/blockstats page. This Id is the first 8 characters of the sha256 hash of the username of the user. You can use a page like http://www.movable-type.co.uk/scripts/sha256.html to convert your username to a hash to see where you rank.

Edit: And now there's a rough estimate of hash rate there as well.

The best stats page that I've seen is this one: http://www.litecoinpool.org/stats (Look at the variance of blocks found per user!). If you can create something similar then that would be great!

I'm not a fan of DGM and I will mine at a PPS pool as long as there is one.

I've been looking at comparing other payment methods to PPS and have added a quick 'stats' page to which shows some data. The first payment method I'm comparing to is DGM as this seems to be a reasonable alternative to PPS which helps lower variance for the pool operator. For the curious the stats are here: http://mmpool.bitparking.com/blockstats

This page shows under the 'Current' heading the current number of difficulty 1 solutions (as Score) and the amount of PPS that has been paid so far for the current block being worked on. On average a block will be solved by the pool every 'Difficulty' score submissions. This number gives you an idea of how far into a block we are and how unlucky we are if we're way above difficulty.

Under 'Blocks' is a list of the blocks that have been solved since I added the payment comparison stats. So far there have only been two full blocks. They show the total amount that would be paid to all miners using the DGM payment method, the total amount actually paid using the existing PPS system, the difference between the two, and the score total that solved the block (number of difficulty 1 solutions submitted by miners).

As you can see with the two block solved there is a bit of variance there with DGM being higher in the 2nd block and PPS in the first. In both cases the PPS is less than the 50 for the block due to both blocks being solved 'early' (less than the the difficulty number of submissions).

The 'Totals' section sums up the DGM and PPS payments over all the blocks and compares them. It also compares them to the block reward to see if the pool is losing or making money. For those two blocks the pool has been lucky. For the last number of blocks it's been a bit unlucky unfortunately so it's actually in the negative a fair amount (hence my reason for looking at lower variance reward options in case I need to change).

Under Estimated Payout is the current values for all miners mining the current block. It is sorted by total score (difficulty 1 submissions by the miner). It shows approximately what the payout for the miner would be using DGM if the block was found. It also shows the current PPS payment made to that miner and the score. At times DGM will be higher than PPS or lower depending on long it is taking to solve the block.

I don't show the miner user id for privacy reasons. Is there interest in that though? Or maybe a hash of the miner Id? It makes an interesting leaderboard.

Note that stats shown on this page are cached so are not 'immediate'. They are also generated by a seperate stats gathering module so may not represent the exact values you're getting in PPS but should be close. These stats are using floating point values rather than dealing with 'coins' so have some roundof errors. My DGM code may have bugs so another reason for this page is for me to test the implementation. The stats also show bitcoin values only.

Comments welcome! Any other data of interest?

I was on a pacific island for three weeks in April/May where they only have power about 12 hours a day - I can relate!

My problems stem from third world Mexico cutting power on and off during the day for god only knows what reason.

You may have noticed a few disconnects of 30 seconds or so today as I updated some of the pool software. Things should be back to normal now.

its just even when i check the site itself tho it doesnt say that im doing anything and on the miner it seems to be correct thats why im wondering if i did do something wrong even tho it should be working

According to share submissions we've hit 300 Ghash/s, mostly from one large miner. Last  time this happened someone had discovered a flaw in the pool but I don't see evidence of this at the moment since we're solving the alternate chain blocks at the expected rate. I'll  keep monitoring it to make sure nothing untoward is happening.