[UNRESPONSIVE POOL OPERATOR] 50BTC.com

[Nemo1024]

Going full throttle. Sure the Web-site may occasionally report a worker as down, but they are actually up and counting towards your total. Only 2 invalids out of 46677 shares for the last 24 hours. Historically I had 13561 invalids out of 1907178 total shares (0.711%).

I have a set-up with 3 backup pools. The first 2 point to 50BTC, while the last one to OzCoin:
pool.50btc.com:3333
pool80.50btc.com
eu.ozco.in:3333

That way, if Stratum is down, I will attempt port 80 connect before abandoning the pool in favour of a reserve.

[John (John K.)]

Anyone else tonight seem to lose some speed? And also have invalids increase?

+1. I've been seeing sporadic speed decreases since 12 hours ago, and it's deteriorated to the point where the speed shows as 0.  My invalids have went to the moon too.

[John (John K.)]

Going full throttle. Sure the Web-site may occasionally report a worker as down, but they are actually up and counting towards your total. Only 2 invalids out of 46677 shares for the last 24 hours. Historically I had 13561 invalids out of 1907178 total shares (0.711%).

I have a set-up with 3 backup pools. The first 2 point to 50BTC, while the last one to OzCoin:
pool.50btc.com:3333
pool80.50btc.com
eu.ozco.in:3333

That way, if Stratum is down, I will attempt port 80 connect before abandoning the pool in favour of a reserve.

Hmm...19941 (3331 invalids) for me in the last 24 hours according to the site.

[Nemo1024]

What is your share's difficulty setting?
I tried setting it to 2 for 7970 cards (per their advice for miners with >500MH/s power), but I noticed that it negatively affected my performance, so now I am back to level 1 for all cards.

EDIT: One card had a spike of 33 (and counting) invalids over the last hour, so my  24-hour stats are now 46463 (35) cumulative for all cards.

[John (John K.)]

What is your share's difficulty setting?
I tried setting it to 2 for 7970 cards (per their advice for miners with >500MH/s power), but I noticed that it negatively affected my performance, so now I am back to level 1 for all cards.

PS: One card had a spike of 33 (and counting) invalids over the last hour, so my  24-hour stats are now 46463 (35) cumulative for all cards.

I ran my 9~ GH/s farm at 8 diff. I'll try modifying it to 4 and see later then. 

[notlist3d]

Ive tried a few things.... my speed ultimately goes down.   Never gets up to proper speed.  And eventually invalids just go crazy.

Growing pains from it moving from 17 to 19+ so quick?  If anyone finds solutions please help with ideas.

[notlist3d]

What is your share's difficulty setting?
I tried setting it to 2 for 7970 cards (per their advice for miners with >500MH/s power), but I noticed that it negatively affected my performance, so now I am back to level 1 for all cards.

EDIT: One card had a spike of 33 (and counting) invalids over the last hour, so my  24-hour stats are now 46463 (35) cumulative for all cards.

Can you please share your settings you are using? 
 
Are you using a backup pool or anything of thiers? Also what port just 3333?

[Nemo1024]

I ran my 9~ GH/s farm at 8 diff. I'll try modifying it to 4 and see later then.  

Does your farm consist of 1 unit á 9GH/s or 18 units (cards) á 500MH/s?

One of my machines has 2x7970 churning out 1.3GH/s, yet I still treat it as 2 units, producing ~600MH/s + ~700MH/s, assigned to two different workers on the pool, each set to diff. 1 shares.

I now restarted cgminer, associated with that card which started to spike with invalids (cgminer also reported an increasing number of R's), and it is back to normal - the increase stopped. I guess I will slightly reduce the restartEvery setting in my Bitcoin Miners in Tray front-end so that restarts and therefore refreshes of the miners' states would occur more often.

Can you please share your settings you are using? 
 
Are you using a backup pool or anything of thiers? Also what port just 3333?

I use cgminer with a front-end of my own writing.

The settings are (this machine has 6670 + 7970, and this setting is for device 0):

-T -l 2 -I d,9 -d 0 --gpu-engine 920,1150 --gpu-memclock 925,1375 --gpu-fan 50,84 -o http://pool.50btc.com:3333 -u username_machine6670 -p password -o http://pool80.50btc.com -u username_machine6670 -p password -o http://eu.ozco.in:3333 -u username.machine6670 -p password

[John (John K.)]

I ran my 9~ GH/s farm at 8 diff. I'll try modifying it to 4 and see later then.  

Does your farm consist of 1 unit á 9GH/s or 18 units (cards) á 500MH/s?

One of my machines has 2x7970 churning out 1.3GH/s, yet I still treat it as 2 units, producing ~600MH/s + ~700MH/s, assigned to two different workers on the pool, each set to diff. 1 shares.

I now restarted cgminer, associated with that card which started to spike with invalids (cgminer also reported an increasing number of R's), and it is back to normal - the increase stopped. I guess I will slightly reduce the restartEvery setting in my Bitcoin Miners in Tray front-end so that restarts and therefore refreshes of the miners' states would occur more often.

Can you please share your settings you are using? 
 
Are you using a backup pool or anything of thiers? Also what port just 3333?

I use cgminer with a front-end of my own writing.

The settings are (this machine has 6670 + 7970, and this setting is for device 0):

-T -l 2 -I d,9 -d 0 --gpu-engine 920,1150 --gpu-memclock 925,1375 --gpu-fan 50,84 -o http://pool.50btc.com:3333 -u username_machine6670 -p password -o http://pool80.50btc.com -u username_machine6670 -p password -o http://eu.ozco.in:3333 -u username.machine6670 -p password

It consists of 11 FPGA's at around 840 MH/s each, connected to the pool via a single stratum proxy. Hmm... I changed the diff to 4, and the numbers seem correct again. My 24 H speed and 24 H earned seems way off though...

[Nemo1024]

It consists of 11 FPGA's at around 840 MH/s each, connected to the pool via a single stratum proxy. Hmm... I changed the diff to 4, and the numbers seem correct again. My 24 H speed and 24 H earned seems way off though...

Wait for 24 hours to see how the change impacts performance long-term. In the meantime, the 60-minute shares should give a fair indication if the change helped.

EDIT: I would treat each board as a separate miner, follow this table

Worker's speed           Difficulty
<500 Mhash/s                 1
500- 1000Mhash/s             2
1000-2000Mhash/s             4
2000-4000Mhash/s             8
4000-10000Mhash/s           16
10000-40000Mhash/s          32
>40000Mhash/s               64

and would keep the difficulty at 2, even though they are all behind the same proxy. Work shares get distributed between the boards, and it might be so that they do not manage to process higher-difficulty shares in time, turning up a large number of invalids...

It is my guess, though. It would be nice if a pool admin elaborated on share difficulty in multi-card/multi-board scenarios.

[John (John K.)]

It consists of 11 FPGA's at around 840 MH/s each, connected to the pool via a single stratum proxy. Hmm... I changed the diff to 4, and the numbers seem correct again. My 24 H speed and 24 H earned seems way off though...

Wait for 24 hours to see how the change impacts performance long-term. In the meantime, the 60-minute shares should give a fair indication if the change helped.

EDIT: I would treat each board as a separate miner, follow this table

Worker's speed           Difficulty
<500 Mhash/s                 1
500- 1000Mhash/s             2
1000-2000Mhash/s             4
2000-4000Mhash/s             8
4000-10000Mhash/s           16
10000-40000Mhash/s          32
>40000Mhash/s               64

and would keep the difficulty at 2, even though they are all behind the same proxy. Work shares get distributed between the boards, and it might be so that they do not manage to process higher-difficulty shares in time, turning up a large number of invalids...

It is my guess, though. It would be nice if a pool admin elaborated on share difficulty in multi-card/multi-board scenarios.

I ran them nicely now at 4. Wasn't sure if that's the solution here, but it's working anyway. 
Thanks!

[John (John K.)]

Okay, here we go again....

Worker is dead. 

[John (John K.)]

30 min shares (invalid): 821 (715) 

[organofcorti]

Try this for an explanation of how and why share difficulty affects your income, and what settings you should use.

[Tigggger]

Okay, here we go again....

Worker is dead. 

Have you tried connecting to the pool nearest you, rather than the general url. I'm in UK so have pool-de, then pool-ru then pool-us

24 Hours
150,000 shares, 9 invalids, < 0.1% Rejects and payments are almost exactly what they should be for hashrate (and it's been like this for weeks, no problems except web page)

[John (John K.)]

Okay, here we go again....

Worker is dead. 

Have you tried connecting to the pool nearest you, rather than the general url. I'm in UK so have pool-de, then pool-ru then pool-us

24 Hours
150,000 shares, 9 invalids, < 0.1% Rejects and payments are almost exactly what they should be for hashrate (and it's been like this for weeks, no problems except web page)

Whoa, I didn't know there was specific pools for different regions. I'll try to connect to the nearest pool for me. 

Try this for an explanation of how and why share difficulty affects your income, and what settings you should use.

*John is dizzy from the math*  I'm running at around 8.5GH/s for now, and I want to submit say 13 shares per minute, I'll need to choose diff 8, right? What is the optimal share amount per minute?

Thanks everyone!

[organofcorti]

[ ... ]

Try this for an explanation of how and why share difficulty affects your income, and what settings you should use.

*John is dizzy from the math*  I'm running at around 8.5GH/s for now, and I want to submit say 13 shares per minute, I'll need to choose diff 8, right?

Yep, which will amount to about +/- 9% hourly variation.

What is the optimal share amount per minute?

It all depends on you. If there's no need to reduce your bandwidth at your end, then by choosing a higher difficulty you're increasing your variance to help the pool reduce its bandwidth usage.

Pick an amount that you're happy with - in the long run (say a few days) even at D32 you wont notice much variance. It's only a problem if you're trying to match the pools' estimate of your hashrate with your local estimate.

[John (John K.)]

[ ... ]

Try this for an explanation of how and why share difficulty affects your income, and what settings you should use.

*John is dizzy from the math*  I'm running at around 8.5GH/s for now, and I want to submit say 13 shares per minute, I'll need to choose diff 8, right?

Yep, which will amount to about +/- 9% hourly variation.

What is the optimal share amount per minute?

It all depends on you. If there's no need to reduce your bandwidth at your end, then by choosing a higher difficulty you're increasing your variance to help the pool reduce its bandwidth usage.

Pick an amount that you're happy with - in the long run (say a few days) even at D32 you wont notice much variance. It's only a problem if you're trying to match the pools' estimate of your hashrate with your local estimate.

Ah, okay, thanks for the advice. 

[hekiman]

since 15:50 dashboard shows 0 MH but worker is connected an gets work???

[Nemo1024]

since 15:50 dashboard shows 0 MH but worker is connected an gets work???

That happens from time to time. Check that your balance increases at the expected rate. The worker stats and graphs should come back on track after a while.