Bad profit

[Minerman]

Aha, I guess I dont need to spend more time and money testing then if you have the same results.
To bad the three breakthroughs did nothing to improve the profits....

[gaanthony]

I think if my miners got the actual hash rate here that I get at the other pools that my profit would be on par if not better. But when the miner is showing 420MHs/sec on it's GUI with minimal fluctuation and the Earnings live view here is showing 342MHs/sec for same miner for long period of time there is an issue somewhere in the system.

[CritterDog]

I noticed yesterday I did a short test just checking the re-start time which I was not even going to mention because it was only about a 30 min test. But on my A2 screen it showed that I had found 7 blocks but on the Prohash pool it only showed I had found 2 blocks.

[Steve Sokolowski]

I think if my miners got the actual hash rate here that I get at the other pools that my profit would be on par if not better. But when the miner is showing 420MHs/sec on it's GUI with minimal fluctuation and the Earnings live view here is showing 342MHs/sec for same miner for long period of time there is an issue somewhere in the system.

This is a result of the way that the GUI is displaying hashrate. I'm going to make a change to make it closer to what you expect it to read.

The displayed hashrate (before I make the change) is the theoretical hashrate that you would obtain with this coin if losses due to the work restart penalty were accounted for. It's the hashrate you see on your miner plus a period of zero mining, divided by the total time for both.

But that's not the actual hashrate at any given time, so I'm going to change it to show your actual hashrate, which should be very close to what the mining software displays. Expect this to be changed by tomorrow unless you hear otherwise.

[Steve Sokolowski]

I noticed yesterday I did a short test just checking the re-start time which I was not even going to mention because it was only about a 30 min test. But on my A2 screen it showed that I had found 7 blocks but on the Prohash pool it only showed I had found 2 blocks.

This is interesting. Which coin were you mining that it failed to show blocks for? Or, if you don't know that, which coin did you find the two blocks for?

If this is an easy coin, it is possible that your miner found blocks after someone from the pool had already found that block, due to network latency, and therefore your block was discarded. But I would need to know the coin to look into that.

[CritterDog]

I noticed yesterday I did a short test just checking the re-start time which I was not even going to mention because it was only about a 30 min test. But on my A2 screen it showed that I had found 7 blocks but on the Prohash pool it only showed I had found 2 blocks.

This is interesting. Which coin were you mining that it failed to show blocks for? Or, if you don't know that, which coin did you find the two blocks for?

If this is an easy coin, it is possible that your miner found blocks after someone from the pool had already found that block, due to network latency, and therefore your block was discarded. But I would need to know the coin to look into that.

I don't remember witch coin it was that it showed I had found the 2 blocks. I do remember the 2 were both on the same coin and I think it started with a E like maybe extream coin or maybe exitcoin but to be honest I really do not remember but it was a pretty good payout coin and pretty sure it started with E if that helps.. On the ones I did not get my miner software does not say what the name is.. It only keeps track of the total blocks found. And I notice my A2 will not find many blocks unless I set the difficulty very low like d=512.. I noticed my smaller gblade miners 4 of them seem to find more blocks then the A2 does. But if I lower the difficulty of the A2 then it starts finding blocks.. If I leave it set on d=4096 it never seems to get any blocks.. But I never really tested this.. In order to be sure I would have to run them seperate times and count blocks with the smaller miners compared to the A2

[Steve Sokolowski]

I noticed yesterday I did a short test just checking the re-start time which I was not even going to mention because it was only about a 30 min test. But on my A2 screen it showed that I had found 7 blocks but on the Prohash pool it only showed I had found 2 blocks.

This is interesting. Which coin were you mining that it failed to show blocks for? Or, if you don't know that, which coin did you find the two blocks for?

If this is an easy coin, it is possible that your miner found blocks after someone from the pool had already found that block, due to network latency, and therefore your block was discarded. But I would need to know the coin to look into that.

I don't remember witch coin it was that it showed I had found the 2 blocks. I do remember the 2 were both on the same coin and I think it started with a E like maybe extream coin or maybe exitcoin but to be honest I really do not remember but it was a pretty good payout coin and pretty sure it started with E if that helps.. On the ones I did not get my miner software does not say what the name is.. It only keeps track of the total blocks found. And I notice my A2 will not find many blocks unless I set the difficulty very low like d=512.. I noticed my smaller gblade miners 4 of them seem to find more blocks then the A2 does. But if I lower the difficulty of the A2 then it starts finding blocks.. If I leave it set on d=4096 it never seems to get any blocks.. But I never really tested this.. In order to be sure I would have to run them seperate times and count blocks with the smaller miners compared to the A2

The reason it finds so many blocks when the difficulty is low is because the pool will not reduce a miner's difficulty very low to mine easy coins. If your miner has a very high difficulty, then you will mine very difficult coins, resulting in few blocks being found.

I think the reason the blocks aren't being recorded is because other miners at our pool have found those blocks first. When that happens, we don't record them as orphans, because our own miners orphaned each other. In the future, it might be a good idea to record these so that users can see why that's going on.

Chris is going to look into work restart penalties tonight. We haven't been able to make sense of why profit is so high recently but some miners are complaining about low earnings. We think that the cause may be that some miners are getting a very high penalty, higher than what should be the case, and other miners have too low a penalty. Chris will investigate to see whether we should change the formula to better smooth out the curve.

[gaanthony]

The Live Worker hash rate display appears to be on par now with the miner GUI. Also looks like after 8 hours so far that the earnings have improved. Still a little over half a LTC diff for me as I get 9.8 to 10.2 LTC day at the other pool depending on LTC network conditions and difficulty but that would account for the 5% fee of mining here as 5% of 10 would be 0.5.
Also work restart penalty and delays values appear to have improved.
So in summary it appears there is some improvement.

[CritterDog]

I am running just one gblack right now rated hash 24MHS. I notice if I set the software d=to 1024 when the miner stabilizes after about 5 mins it says my re-starts are about 2 sec and I am lucky if i get one block over 30 mins.. But as the machine changes coins over time the re-starts slowly will improve but it takes a very long time. Now if I do a reboot on the same machine changed to d=512 it is much much better as it instantly has re-start numbers very low well under 1 sec and under the penalty and instantly starts finding blocks. I have noticed the higher you set the difficulty the re-starts also go higher and then you get a penalty. But if you set a low diff like 512 on a high power machine it comes up fine right off with great restart times and no penalty.

[gaanthony]

@CritterDog - I think you may be on to something with the d= setting. If I don't use a static d= the pool eventually changes it dynamically to 262144. I've run with it at d=8196 up to 262144 in the past not seeing much difference. I run with that as each of my miners are >400MHs/s.
Now I test different runs from d=256 up to d=8196 and d=4096 seems to be a sweet spot for my miners on this pool.
Long running test will tell but on track now to beat earnings at the other pool.