Looking for more information about reduced hashrate miners

Discussion of development releases of Prohashing.
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Thu Jul 05, 2018 1:57 pm

Steve Sokolowski wrote:
CSZiggy wrote:When my brother points his miner to this pool he never goes over 580 as a hashrate even at 100% efficiency. But on litecoinpool its over 610 for the 24-hour average.

Could this be reason my brother and I have very often seen higher payouts on litecoinpool.org It could very well be that PH does actually pay out the highest payout per hashrate. But with Litecoinpool.org seeing all of the hashrate being applied it just ends up paying out higher each day?


I guess I only ever thought about leaving coins on the table. Never thought to count the hashrate and make sure it was being credited equally and not being left on the table.
Part of this problem could be caused by work restarts that reduce hashrate because the miners require time to spin up to the next coin they are assigned. Litecoins have few work restarts because their block time is so long.

Last week, I modified the coin assignments so that we assigned a lower percentage of hashrate to the other coins, so that we don't get period of lots of blocks being found for those coins within a few seconds of each other.

However, the difference between 580 and 610 MH/s is less than the 12% advantage we had over straight LTC this week. The system automatically determines if it makes sense to assign miners to coins that would reduce their hashrate and only does so if the increase in profitability at the lower hashrate exceeds the reduced hashrate. While I'll continue to investigate this issue, if the reduction in hashrate is caused by work restarts, it isn't actually lowering profitability enough to cause other pools to be advantageous.

That said, I did find just now that we are not recording all stale shares to the database. I'm going to make some changes to record all stale shares. As with the stale blocks, this change will help us get a better idea of whether there are a large number of rejected shares due to some reason that we aren't aware of.
I actually find that the payout per MH stated on the prohashing site is a lot closer than 12% difference based on the calculator on the litecoinpool.org site. It is actually about 4-5% and if hash rates are being underutilized on prohashing it could easily wipe out that benefit and make litecoinpool.org equal to, if not, more profitable. You can't depend on your calculation of profitibablity for just mining litecoin because your fees are 5% and litecoinpool.org is much less and they don't have work restarts as you mentioned. I like prohashing, but I closely monitor profitability for me between them and will point my power where I feel like will be benificial. Right now the value comes in the fact I can be paid out in the coins I want in prohashing with profitability being about the same as litecoinpool.org.
User avatar
Steve Sokolowski
Posts: 3826
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Re: Looking for more information about reduced hashrate miners

Post by Steve Sokolowski » Thu Jul 05, 2018 4:32 pm

jeffms2003 wrote:
Steve Sokolowski wrote:
CSZiggy wrote:When my brother points his miner to this pool he never goes over 580 as a hashrate even at 100% efficiency. But on litecoinpool its over 610 for the 24-hour average.

Could this be reason my brother and I have very often seen higher payouts on litecoinpool.org It could very well be that PH does actually pay out the highest payout per hashrate. But with Litecoinpool.org seeing all of the hashrate being applied it just ends up paying out higher each day?


I guess I only ever thought about leaving coins on the table. Never thought to count the hashrate and make sure it was being credited equally and not being left on the table.
Part of this problem could be caused by work restarts that reduce hashrate because the miners require time to spin up to the next coin they are assigned. Litecoins have few work restarts because their block time is so long.

Last week, I modified the coin assignments so that we assigned a lower percentage of hashrate to the other coins, so that we don't get period of lots of blocks being found for those coins within a few seconds of each other.

However, the difference between 580 and 610 MH/s is less than the 12% advantage we had over straight LTC this week. The system automatically determines if it makes sense to assign miners to coins that would reduce their hashrate and only does so if the increase in profitability at the lower hashrate exceeds the reduced hashrate. While I'll continue to investigate this issue, if the reduction in hashrate is caused by work restarts, it isn't actually lowering profitability enough to cause other pools to be advantageous.

That said, I did find just now that we are not recording all stale shares to the database. I'm going to make some changes to record all stale shares. As with the stale blocks, this change will help us get a better idea of whether there are a large number of rejected shares due to some reason that we aren't aware of.
I actually find that the payout per MH stated on the prohashing site is a lot closer than 12% difference based on the calculator on the litecoinpool.org site. It is actually about 4-5% and if hash rates are being underutilized on prohashing it could easily wipe out that benefit and make litecoinpool.org equal to, if not, more profitable. You can't depend on your calculation of profitibablity for just mining litecoin because your fees are 5% and litecoinpool.org is much less and they don't have work restarts as you mentioned. I like prohashing, but I closely monitor profitability for me between them and will point my power where I feel like will be benificial. Right now the value comes in the fact I can be paid out in the coins I want in prohashing with profitability being about the same as litecoinpool.org.
The 12% being reported is per MH/s/day, so if you can get a higher hashrate at another pool due to reduced work restarts, then the comparison would need to be adjusted for the higher hashrate at that pool.

I'm curious to see what percentage of shares show up as rejected after tonight's release. That will provide more information about the hashrate issue.
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Thu Jul 05, 2018 4:57 pm

Steve Sokolowski wrote:
jeffms2003 wrote:
Steve Sokolowski wrote:
Part of this problem could be caused by work restarts that reduce hashrate because the miners require time to spin up to the next coin they are assigned. Litecoins have few work restarts because their block time is so long.

Last week, I modified the coin assignments so that we assigned a lower percentage of hashrate to the other coins, so that we don't get period of lots of blocks being found for those coins within a few seconds of each other.

However, the difference between 580 and 610 MH/s is less than the 12% advantage we had over straight LTC this week. The system automatically determines if it makes sense to assign miners to coins that would reduce their hashrate and only does so if the increase in profitability at the lower hashrate exceeds the reduced hashrate. While I'll continue to investigate this issue, if the reduction in hashrate is caused by work restarts, it isn't actually lowering profitability enough to cause other pools to be advantageous.

That said, I did find just now that we are not recording all stale shares to the database. I'm going to make some changes to record all stale shares. As with the stale blocks, this change will help us get a better idea of whether there are a large number of rejected shares due to some reason that we aren't aware of.
I actually find that the payout per MH stated on the prohashing site is a lot closer than 12% difference based on the calculator on the litecoinpool.org site. It is actually about 4-5% and if hash rates are being underutilized on prohashing it could easily wipe out that benefit and make litecoinpool.org equal to, if not, more profitable. You can't depend on your calculation of profitibablity for just mining litecoin because your fees are 5% and litecoinpool.org is much less and they don't have work restarts as you mentioned. I like prohashing, but I closely monitor profitability for me between them and will point my power where I feel like will be benificial. Right now the value comes in the fact I can be paid out in the coins I want in prohashing with profitability being about the same as litecoinpool.org.
The 12% being reported is per MH/s/day, so if you can get a higher hashrate at another pool due to reduced work restarts, then the comparison would need to be adjusted for the higher hashrate at that pool.

I'm curious to see what percentage of shares show up as rejected after tonight's release. That will provide more information about the hashrate issue.
I understand how you calculated the difference, what I'm saying is I did the same thing based on the MH/s I know my miners are performing and the calculator on litecoinpool.org and came up with a 4-5% variance in Prohashing's favor which could be negated if there are really restart issues dragging down my hash rate performance on prohashing.
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Thu Jul 05, 2018 9:50 pm

Here is my problem illustrated in a chart form. The prohashing rates are based off of the expected payout charts from yesterday when market prices were actually a bit higher across the board. My alternate pool calculations are based off of slightly lower prices for today. I know for a fact that the alternate sites reward me in line with my actual has rate reported on my machines because I've tested them in the past and. Will likely do in the near future to validate.

I can't paste an image of the spreadsheet, so here is a google document to the spreadsheet: https://docs.google.com/spreadsheets/d/ ... sp=sharing
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Fri Jul 06, 2018 12:30 am

Just as an update, my earnings for the day were $37.48 and based on the published expected payouts and my actual hash rates I should have received $43, which means I was shorted about 13%. I'm more than happy to show a spreadsheet that would prove out that I'm not making anything near what you guys are saying if it would help.
User avatar
CSZiggy
Posts: 691
Joined: Wed Jan 31, 2018 2:44 pm

Re: Looking for more information about reduced hashrate miners

Post by CSZiggy » Fri Jul 06, 2018 10:08 am

Well, last night the difficulty on the litecoin network dropped again which will increase my profits yet some more.

On PH I was getting .028 litecoins per day on the L3+
On Litecoinpool I was getting .031

With the difficulty drop, should get .033-.035 daily for the next few days.

Still have 0 clue how you can claim PH pool pays out LTC + 12% when straight LTC mining on litecoinpool.org actually pays out higher.
I really think you need some type of disclaimer:
"LTC + 12%*" *=straight LTC mining on this pool only, not on other pools that do straight LTC mining.
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Fri Jul 06, 2018 10:23 am

I'm about to move all my hash power to other pools to test and see if PH really does offer more. To me it is slightly less profit than mining straight BTC, LTC, ZEC with the benefit of getting auto paid in the coins you want. There are occasional days where PH is definitely more profitable, but I value consistency and accurate reward of my hash power, so I'll just have to weigh and decide after I test a few options.
jeffms2003
Posts: 43
Joined: Fri May 11, 2018 11:40 am

Re: Looking for more information about reduced hashrate miners

Post by jeffms2003 » Sat Jul 07, 2018 12:41 am

After about 12 hours of testing, I'm on track to make about 5% more due to the fact that my hash rates weren't reflecting actual performance. The other pools are registering exactly in line with what my machines read. It has to do with the fact that miners are constantly changing coins mined and the delays caused by constantly switching. I wonder if it would make any sense to set a minimum time limit mining a coin once assigned.
User avatar
CSZiggy
Posts: 691
Joined: Wed Jan 31, 2018 2:44 pm

Re: Looking for more information about reduced hashrate miners

Post by CSZiggy » Sat Jul 07, 2018 1:05 am

>>> I wonder if it would make any sense to set a minimum time limit mining a coin once assigned.

Would only help if the coin you are on is never found, or a work restart is never needed to be sent to find the next one.
They have a coin for that, its called Litecoin.
User avatar
Steve Sokolowski
Posts: 3826
Joined: Wed Aug 27, 2014 3:27 pm
Location: State College, PA

Re: Looking for more information about reduced hashrate miners

Post by Steve Sokolowski » Sat Jul 07, 2018 9:11 am

jeffms2003 wrote:After about 12 hours of testing, I'm on track to make about 5% more due to the fact that my hash rates weren't reflecting actual performance. The other pools are registering exactly in line with what my machines read. It has to do with the fact that miners are constantly changing coins mined and the delays caused by constantly switching. I wonder if it would make any sense to set a minimum time limit mining a coin once assigned.
I think that there's more to it than this. While we don't have enough data yet for it to be statistically significant, the release last night has yielded 8 hours of rejected share data, and it shows that there are a significant number of "duplicate shares."

Some types of rejected shares, like "both the primary and merge-mined coin are stale," are unavoidable due to network latency. But "duplicate share" should never occur. It means that there are either a lot of miners who have misconfigured mining rigs, or the mining server has a bug that sends the same work when a coin is switched back to.

I'm going to look into where these duplicate shares are coming from today to see if there is a way we can eliminate them.
Post Reply