inflatin
Newbie
Offline
Activity: 14
Merit: 0
|
|
February 22, 2014, 02:14:09 AM Last edit: February 22, 2014, 03:33:04 AM by inflatin |
|
Please change vardiff back lower... at least to 128! I have 25 workers all around 40 kHash/sec. Noticed a drop in shares since the 512 hardcode. edit: after looking this over my shares are pretty much the same. I do have more variance though. see http://imgur.com/ZIk5lHL
|
|
|
|
phzi
|
|
February 22, 2014, 03:20:22 AM Last edit: February 22, 2014, 03:32:32 AM by phzi |
|
Happy to hear the server load has improved without vardiff. I eventually realized that variance doesn't hurt my payouts at all, so this is a welcomed change.
My suggestion to appease the less knowledgable who are complaining about variance: provide 2 hashrate messurements - 5min window and then a full N period window. Look at share count over the entire N period, calculate hashrate based on time between rounds and total shares. Then lay it out like:
Instant Hashrate: xxxMH/s (will fluctuate) Average Hashrate: xxxMH/s (takes 10 shifts to be accurate)
Keep up the solid work.
|
|
|
|
poolwaffle (OP)
|
|
February 22, 2014, 09:18:21 AM |
|
Please change vardiff back lower... at least to 128! I have 25 workers all around 40 kHash/sec. Noticed a drop in shares since the 512 hardcode. edit: after looking this over my shares are pretty much the same. I do have more variance though. see http://imgur.com/ZIk5lHLThis is the exact definition of variance. Happy to hear the server load has improved without vardiff. I eventually realized that variance doesn't hurt my payouts at all, so this is a welcomed change.
My suggestion to appease the less knowledgable who are complaining about variance: provide 2 hashrate messurements - 5min window and then a full N period window. Look at share count over the entire N period, calculate hashrate based on time between rounds and total shares. Then lay it out like:
Instant Hashrate: xxxMH/s (will fluctuate) Average Hashrate: xxxMH/s (takes 10 shifts to be accurate)
Keep up the solid work.
Yep, we'll end up moving to a longer timeperiod displayed (30min most likely). The issue is (like an idiot), I have stuff spinkled throughout the codebase that expects miner hashrates to be saved in the DB as 5min values. So if I change it to precalculate 30min values, somewhere in the code, we'll be seeing values way off. Just gotta dig through and find them all before I can switch
|
|
|
|
asdosoasd
Member
Offline
Activity: 92
Merit: 10
|
|
February 22, 2014, 09:23:42 AM |
|
hi poolwaffle, why the unconverted is more than converted?
|
|
|
|
ANJULE
|
|
February 22, 2014, 09:32:11 AM |
|
Because new coins were added that cant be auto-changed to BTC.
I think this diff=512 is good, just have to be a bit more patient with the stats. Even more so if doing changes to your cgminer.conf.
|
CoinWare.eu - Just started an online-store that accepts BTC! For now only few items, but more to come! Shipping allover the World by Express-mail. Europe: 3-6workdays| USA: 4-7workdays | Rest of the World: 4-11 workdays
|
|
|
Qxw
|
|
February 22, 2014, 09:33:29 AM |
|
What is going on with eu server. Reject rate rised from average 0.5% to 10-20% and even more. Lot of stales, lot of duplicates, lot of job not found... etc... Problem is in server end, not my rigs (only 8MH/s now on waffle) Specially it looks that this is with some coin. Previously I have not find this with waffle.
After 12 hour watch, rejects <0.1% with my one test rig with eu.waffle. Changes what have made looks really perfect. Nice to see that wafflepool is in good hands. also vardiff off is very good. Fixed 512 is ok and very welcome change also with this amount of miners. It looks like many pepoles do not understand this diff / vardiff case at all. There is lot of misunderstooding in this. Partially due to lot (and more) disinformations around of forums. But I hope wafflepool admin keep his head calm. Poples who know - I mean know - mostly they do not rant about this at all. Who do not believe this, he need refresh lessons of statistics. This image is one nice example. More short time variations but nothing else and no any negative effect. Just do not spend your time and stress and disturb your mind with looking every minute and hour your hasrates. Well adjusted good machines do not need human watching all over time. Leave them do work and just "Do not disturb your employees when they are working." (This image is not from my test rig! This is from MisterWill link https://bitcointalk.org/index.php?topic=433634.msg5287138#msg5287138 ) What is loosed after change to fixed 512 diff. Just nothing or... On the contrary - it may produce more (profit) for all of us. Even for low hashrate miners.
|
BTC, BCH, BTG, UTC
|
|
|
asdosoasd
Member
Offline
Activity: 92
Merit: 10
|
|
February 22, 2014, 09:46:31 AM |
|
i have diff=1024 and it works very well!! rejects are 0.3% and 1.45 Mhs
|
|
|
|
bigblind
|
|
February 22, 2014, 10:00:03 AM |
|
but careful please with too many unexchanged new coins as one might dump the price if you flood the market with them alltogether ^^
|
|
|
|
poolwaffle (OP)
|
|
February 22, 2014, 10:13:57 AM |
|
but careful please with too many unexchanged new coins as one might dump the price if you flood the market with them alltogether ^^
We calculate our price based on the unexchanged amount (depth), as well as our hashrate compared to the network (assuming the other percent of the network will dump their coins as well). This gives us a conservative estimate as far as depth goes. Some of the issue currently is that our Cryptsy API isn't working properly (getting unauthorized errors for trading). Nothing has changed on our side, so we've got a ticket into them to see whats up...
|
|
|
|
kosmonauti
Newbie
Offline
Activity: 7
Merit: 0
|
|
February 22, 2014, 10:22:10 AM |
|
24H = 54% LTC+DOGE http://wafflepool.com/statsIt is no wonder that the weak profits. I'll go to some other place where the admins know what they're doing. Soon here that falls below 1gig, again, I see. Is there a pilot in the plane. edit: Here's a list of amateur http://www.coinwarz.com/cryptocurrency
|
|
|
|
poolwaffle (OP)
|
|
February 22, 2014, 10:33:32 AM |
|
And as explained multiple times, to multiple people. That list is blatently wrong in some regards (block rewards), in some regards slightly wrong (delayed), and in some regards correct (coins we can't mine - low difficulty). I'm extremely confident in our profitability calculator, its not like its hard to multiply two numbers.
|
|
|
|
azebro
Newbie
Offline
Activity: 52
Merit: 0
|
|
February 22, 2014, 10:37:28 AM |
|
Something is wrong with stats. I've now put all worker names and on 10MH+ can see only 2mh (!!!) Shall I remove the worker names? i'm using BTC_workername poolwaffle, can you pls explain the issue with workers? as a test I set up 8 workers for my rigs and hash reported dropped to 2mh took away workers and went to normal. Thanks, Can you see each worker reporting separately, or you just see single worker @ 2 MH? I cannot see the each worker, can I? (correct me if i'm wrong) Total for btc address dropped.
|
|
|
|
ANJULE
|
|
February 22, 2014, 10:45:04 AM |
|
No you can't. If I remember correctly it was for variable diff. only. So that each of your rigs could have different variable diff. Correct me if I'm wrong...
|
CoinWare.eu - Just started an online-store that accepts BTC! For now only few items, but more to come! Shipping allover the World by Express-mail. Europe: 3-6workdays| USA: 4-7workdays | Rest of the World: 4-11 workdays
|
|
|
GunPowder
|
|
February 22, 2014, 10:54:56 AM |
|
Is there any reason to mine with separate worker names, after we've got d=512 constantly? When workers are numbered my total hash rate in the list of miners is wrong.
P.S. Seems to me today is the good day
|
|
|
|
azebro
Newbie
Offline
Activity: 52
Merit: 0
|
|
February 22, 2014, 11:05:16 AM |
|
Going through in order poolwaffle, can you pls explain the issue with workers? as a test I set up 8 workers for my rigs and hash reported dropped to 2mh took away workers and went to normal.
I don't know of any issues with workers like that. Can you email me what you were using for worker names? Perhaps its an invalid character or something? Can anyone explain what requested work restarts are?
I mostly use old cards like 5850-s in my rigs. I have around 2.5-2.8MH. I tried out middlecoin but my profit was really bad. Since they changed to 1024 diff it was worse. I had lot of work restarts and around 5% rejects.
Then I found waffle with vardiff. Worked like a charm. Around 1% rejected and no restarts. Since diff is changed to 512 I'm getting a lot of restarts again and reject went up to 2,5%. My profit went down.
Anyone can explain why is that?
Restarts are a normal thing. It means there is new work (compared to the old work you were working on), this could be because there are new transactions we want to include in a block, or we switched coins. You're most likely getting the exact same number of them as before, just that with fewer shares submitted (512 difficulty instead of whatever you were before), you see them more often. They're perfectly normal, and are to be expected. As for the 5% rejects, that seems high, what is the reasons returned for them? I use the following pattern: BTC_off1 BTC_off2 BTC_garage1 BTC_garage2 and so on just incrementing numbers. When I set that no hashrate reported dropped dramatically. Please advise.
|
|
|
|
oktay50000
|
|
February 22, 2014, 11:29:59 AM |
|
now pool working great thanks waffle efficiency incresed to almost 70% from 35% and profit per mh returning
|
BTC : bc1qqz9hvv806w2zs42mx4rn576whxmr202yxp00e9
feel free to buy me a bear
|
|
|
poolwaffle (OP)
|
|
February 22, 2014, 11:31:11 AM |
|
Going through in order poolwaffle, can you pls explain the issue with workers? as a test I set up 8 workers for my rigs and hash reported dropped to 2mh took away workers and went to normal.
I don't know of any issues with workers like that. Can you email me what you were using for worker names? Perhaps its an invalid character or something? Can anyone explain what requested work restarts are?
I mostly use old cards like 5850-s in my rigs. I have around 2.5-2.8MH. I tried out middlecoin but my profit was really bad. Since they changed to 1024 diff it was worse. I had lot of work restarts and around 5% rejects.
Then I found waffle with vardiff. Worked like a charm. Around 1% rejected and no restarts. Since diff is changed to 512 I'm getting a lot of restarts again and reject went up to 2,5%. My profit went down.
Anyone can explain why is that?
Restarts are a normal thing. It means there is new work (compared to the old work you were working on), this could be because there are new transactions we want to include in a block, or we switched coins. You're most likely getting the exact same number of them as before, just that with fewer shares submitted (512 difficulty instead of whatever you were before), you see them more often. They're perfectly normal, and are to be expected. As for the 5% rejects, that seems high, what is the reasons returned for them? I use the following pattern: BTC_off1 BTC_off2 BTC_garage1 BTC_garage2 and so on just incrementing numbers. When I set that no hashrate reported dropped dramatically. Please advise. Again, please email me the actual worker names you were using (all of them if possible) so I can check on them in the database.
|
|
|
|
richmke
|
|
February 22, 2014, 11:40:46 AM |
|
Is there a way to see individual worker stats? If not, what is the purpose of the worker option? If so, how do you see individual worker stats?
|
|
|
|
ANJULE
|
|
February 22, 2014, 11:47:13 AM |
|
See above and earlier post before asking the same question every 5mins, please. Just to answer it again: No, it was for separate rigs for their separate variable diff, which no longer exists because hard diff 512
|
CoinWare.eu - Just started an online-store that accepts BTC! For now only few items, but more to come! Shipping allover the World by Express-mail. Europe: 3-6workdays| USA: 4-7workdays | Rest of the World: 4-11 workdays
|
|
|
azebro
Newbie
Offline
Activity: 52
Merit: 0
|
|
February 22, 2014, 11:58:40 AM |
|
Going through in order poolwaffle, can you pls explain the issue with workers? as a test I set up 8 workers for my rigs and hash reported dropped to 2mh took away workers and went to normal.
I don't know of any issues with workers like that. Can you email me what you were using for worker names? Perhaps its an invalid character or something? Can anyone explain what requested work restarts are?
I mostly use old cards like 5850-s in my rigs. I have around 2.5-2.8MH. I tried out middlecoin but my profit was really bad. Since they changed to 1024 diff it was worse. I had lot of work restarts and around 5% rejects.
Then I found waffle with vardiff. Worked like a charm. Around 1% rejected and no restarts. Since diff is changed to 512 I'm getting a lot of restarts again and reject went up to 2,5%. My profit went down.
Anyone can explain why is that?
Restarts are a normal thing. It means there is new work (compared to the old work you were working on), this could be because there are new transactions we want to include in a block, or we switched coins. You're most likely getting the exact same number of them as before, just that with fewer shares submitted (512 difficulty instead of whatever you were before), you see them more often. They're perfectly normal, and are to be expected. As for the 5% rejects, that seems high, what is the reasons returned for them? I use the following pattern: BTC_off1 BTC_off2 BTC_garage1 BTC_garage2 and so on just incrementing numbers. When I set that no hashrate reported dropped dramatically. Please advise. Again, please email me the actual worker names you were using (all of them if possible) so I can check on them in the database. Please check those to start with: 1Y9My747MgN9jVovucztEKSgLengPrsm6_off1 1Y9My747MgN9jVovucztEKSgLengPrsm6_off2 1Y9My747MgN9jVovucztEKSgLengPrsm6_garage1 1Y9My747MgN9jVovucztEKSgLengPrsm6_garage2
|
|
|
|
|