Topic: [POOL][Scrypt][Scrypt-N][X11] Profit switching pool - wafflepool.com - page 223. (Read 465668 times)

I'd like to add that I was completely wrong about my issues. I thought it was my slow CPU miners, but it can't be because as everyone else said, the diff should have just increased variance, but I switched to Middlecoin (which is locked at 1024 diff) and I'm getting no REJECTED shares. So sorry for that confusion, but I only seem to have this problem on Wafflepool, which is too bad. I've tried multiple Wafflepool servers, but just get continuous REJECTED shares. I'll keep trying, I really want to mine here. Thanks to all for the information and suggestions.

We're actually testing it today on a more broad scale than we'd like (making it as fine-grained as we want is challenging on a technical side).  Right now, about 1/5th of our hashrate (EU cluster) has a more open set of coins that it can deal with (essentially 5+ difficulty), where the others are only dealing with 20+ for now.  We want to split it further than that, but again, its not something that happens overnight (unlike the hashrate boost).

Thanks for what your doing poolwaffle!  Work with wafflepool has been a great experience for me so far!

+1 for a donate button, or a way to contribute more. I think you're doing a great job. My hash rate is > 20 MH/s and I'd gladly donate .5-1% above your fee. Many of us came from MC and that fee was 3.88%.

MC's biggest problem was lack of communication. As long as you keep talking to us, there should be NO complaining - only suggestions, questions, and thanks!

It's Bitcoin, not BitCHcoin.    It's the wild west and we need more frontiersmen and less whining.

שָׁלוֹם

Are you working on/going to implement soon some sort of "coin flipping" distribution for all the wafflepowah at your command over more than one coin at a given time?

He politely requested to be removed from that list via email, was nothing to do with what I wanted, just that he requested it.  Didn't seem like a problem (also is literally 1 line of code), if anyone else would specifically like to be removed from the list, feel free to email me, I'd be happy to make it happen.  Same goes, if people think that list is annoying/useless (I'm not sure there actually is much use to it), I'm perfectly OK with removing it entirely, so we just have per-address stat pages.


As for all of the talk of 512 difficulty for CPU miners.  Again, it really doesn't change anything.  You won't get any higher reject rate (or at least you shouldn't unless something is wrong on your miner) than with vardiff.  You will see many different blocks causing work restarts, but these work restarts take a few microseconds or so (especially on CPU - where they take much less time than GPU) and you won't get a share in during each block, but your stats will be (over a long enough time period) exactly the same if the difficulty is hardlocked at 1, or hardlocked at 512 (or vardiff).

The only logical point I've heard here is the issue with PPLNS timeframes and hardlocked 512.  But again, even this, overall you should end up with the exact same amount of payout over a long enough time.  There might be PPLNS rounds where you have 0 shares in the full 10 rounds that are paid on a block, and you'll earn less than you would have on vardiff.  But there will also be sets of 10 rounds where you have more than your normal amount of shares, and will be paid significantly more than you should have.  There is some variance on the side of which blocks were found at which times, and you might have different per-round earnings than another person with a slow hashrate, but again, over a long enough time, you will have the exact same average earnings.

I keep using the words "over a long enough time", and I haven't explained what that means, partially because there really isn't a good answer (without me doing a bunch of math).  It essentially means that over infinite time, things will be equal.  Obviously we don't have infinite time.  The longer the time frame, the closer your averages should be, the shorter the timeframe, the more the variance could show.

For example, if I ask you to flip a coin 2 times and tell you I'll only pay you if you flip a tails, a lot of people are going to have 2 tails, a lot of people would have 2 heads.  This would be 2 extremely different results (in terms here, some people would be paid 0, some people would be paid 200% of what they expected).  Now, if I tell you to flip it 100 times each, you'll notice some people still have lower than average numbers (maybe 30 tails), some have higher than average (70 tails), some have just about correct (45-55).  Over a long enough time (number of flips here), everyone will come very close to 50/50 (on average).  People with faster miners just flip coins faster, people on CPUs flip coins slower.

Sounds good. Thanks for the reply.

Can you confirm that you were paid yesterday?


**EDIT** Just want to say that this morning's payout just came through for me.

Thank you, wafflepool! Loving it.

Check your stats. Payouts could be late again as I see a lot of unconverted again.

Just curious if anyone else has received a payout within the last few hours.

Just started mining Sunday evening in the central U.S. Probably didn't have .01 BTC at the time of yesterday's payout, and didn't receive anything then, and haven't received anything this morning (meaning in the 11:00 - 13:00 GMT time span today). That's completely fine - just wanting to make sure I have everything set up correctly.

My stats do show perfectly fine.

Thanks.

This is similar to what I'm seeing. Even with 100 cpu miners, I maybe get in 1 share every 2 blocks and all the old work is rejected as the other CPUs take too much time to finish their share and then when submitting, it's too old. At least that's how I understand it works.

So it seems in the end it is a mix of the high difficulty + such a short block time on certain coins. I imagine that if Wafflepool starts mining on an altcoin with a long block time, my rejects would go away.

I was really hoping that it was the stratum proxy lagging, as I hadn't checked that before, but my tests show a 50ms delay from miner to stratum proxy to external proxy to Wafflepool. I'm not sure that that kind of delay is enough to cause this issue though. When I was mining Litecoins before I switched to this pool, I was always getting a difficulty < 200. Any time it went above 300, I started to have the issue of less accepted shares because of the time it takes the cpu to calculate such a high difficulty share. If you say CPU mining is perfectly viable at 512 share difficulty, then I'll keep working on my config to see where my setup is going wrong. I am really hoping to be able to stay at Wafflepool. Thanks for the info.

In my opinion the only way to achieve this (with static diff.) is to have a bigger n. But a bigger n would be bad for the pool, because it would penalize faithful miners against pool hoppers. The only reasonable solution (again IMHO) is to have the ability to reduce the diff.

could anybody tell me, how many traffic a rig with five r9 280x (3.6mH/s) mining at wp uses? (the os is smos linux/litecoin bamt) will it work properly with 3g or even just edge? which data plan will i need for a month?

Qwert brings up a good point. If you have very stable hashing rate, then PPLNS is fine. I find that my hash rate varies by +/- 10%, and diff 512 can make that worse. In some pools I was in, my payout rates were 97% because of my natural variance, when it should be over 100% (benefit from other people coming and going).

Is there a way to modify PPLNS so that the variance is not as big a killer?

Both will work.

FYI: Since worker stats are not available yet, I have separate BTC addresses for each worker. That way I can monitor each one.

If I have 2 workers, I need to put them with 2 different users or I can put with same?

example:

worker 1:-u 1GqgsR6Ut66BnasLyc2k9NCdWqUBJmtELE_worker1
worker 2:-u 1GqgsR6Ut66BnasLyc2k9NCdWqUBJmtELE_worker2

or

worker 1:-u 1GqgsR6Ut66BnasLyc2k9NCdWqUBJmtELE
worker 2:-u 1GqgsR6Ut66BnasLyc2k9NCdWqUBJmtELE

Thanks!

I have 2 cpu miners too with 38KH/s. One of the pointed HC, to keep my presence. (They are almost nothing compared to my 9 older GPUs)
The other is on WP.
Last 18 hours: [2014-02-25 12:06:27] accepted: 80/84 (95.24%), 36.13 khash/s (yay!!!)

Hi!

How can I calculate from accepted shares and avarage KH that my miner works fine? Like it should be like this many accepted and I got that many.

Poolwaffle!

OK, we understand that vardiff uses cpu power. But why fixed 512? You know that this kills cpu miners. What would change if you would allow to set it lower and the default would be 512?
And those who are saying that higher diff only makes higher variance: I can have tens of machines running at 20-30 khash/s each. And yes my (say) weekly average share will be the same. But think about PPLNS! I lose the advantage of the LN part! If i only find shares every 5-10 rounds then i will run like PPS.
And please don't misunderstand me, i highly appreciate and adore your work, and i also understand the need to change as the pool grows, but its frustrating to see my profitability plunge.

Edit: It's easy to misunderstand the PPLNS part. The sentence would be correct as: If i only find shares every 5-10 rounds then i won't benefit of PPLNS's "luck equalizer" effect.

This is not true... you will see higher variance in rejects, and possibly spikes of higher rejects, but over time it evens out no matter what the share diff.

This is simply obvious - your chance to discover a block is proportional to your hashrate... shares are no different.


PoolWaffle has clearly stated he's working on it - this requires a massive re-work of the profitability switcher.