I like the method you use to reduce stales
Topic: bit pit - (LP, ESMPPS, 8-decimal payout, SSL, API, 0% fee, Almost 0% Stales!) - page 18. (Read 80690 times)
July 11, 2011, 09:47:10 AM
July 10, 2011, 08:46:32 PM
A quick update for everyone.
We deployed today, and among some other tasks we also added support for the X-Roll-NTime header.
Not that we were having problems with miner idles, but this should in-general be a good thing.
We deployed today, and among some other tasks we also added support for the X-Roll-NTime header.
Not that we were having problems with miner idles, but this should in-general be a good thing.
July 10, 2011, 04:34:20 PM
Well I brought everything I have over now in an attempt to beat this round into submission. 
It's only 5 cards but hey...we do what we can huh?
It's only 5 cards but hey...we do what we can huh?
Thanks!
Every little bit of hashing power counts. Before we know it, round 6 will be behind us.
July 10, 2011, 03:36:25 PM
Well I brought everything I have over now in an attempt to beat this round into submission.
It's only 5 cards but hey...we do what we can huh?
It's only 5 cards but hey...we do what we can huh?
July 10, 2011, 02:57:54 PM
1bitc0inplz, can you explain what a "Score this round" mean?
P.S. Longest round I ever seen... 7kk shares omfg
P.S. Longest round I ever seen... 7kk shares omfg
Score this round is a hold over from when we used to be a score based pool (the first round). Since then we've been proportional, but we left that on the screen as it shows you what percentage of the top earner's estimated reward is equal to your estimated reward.
Considering this has been the source of numerous confusions, we will be removing it shortly.
In regards to the long round, indeed.
From looking at other pool's public stats, I've seen some that have had 8 million share rounds. Yikes!
We just drew the unlucky straw on this one, however this (hopefully) means the end of the round is coming soon.
July 10, 2011, 02:53:39 PM
1bitc0inplz, can you explain what a "Score this round" mean?
P.S. Longest round I ever seen... 7kk shares omfg
P.S. Longest round I ever seen... 7kk shares omfg
July 10, 2011, 12:25:45 PM
I am also trying this pool, and found it sounds and looks good
Thanks! We appreciate you joining.
July 10, 2011, 09:29:34 AM
I am also trying this pool, and found it sounds and looks good
July 10, 2011, 08:31:08 AM
Code:
Pool │Long│ Job │ Failed │Accepted│Rejected│Accepted│Stale shares│Share upload│Average│Effi-
name │Poll│requests│job requests│ jobs │ jobs │ shares │ (rejected) │ retries │MHash/s│ciency
Bitp.it│Yes │ 2005│ 23 (1.1%)│ 1841│ 140│ 1116│ 3 (0.3%)│ 0 (0.0%)│ 66.92│ 97.4%
July 10, 2011, 03:22:25 AM
3/1100 rejected with a 5770 
July 10, 2011, 03:17:17 AM
At this nght I counted 1/2563*100 = 0,039% rejected on my HD 5830 / 318Mhps. Good videocard, good pool 
July 10, 2011, 01:31:33 AM
Keep up the great work everyone. We've seen some good steady growth today.
http://www.remixito.com/images/actu/propagande/we-can-do-it/together-we-can.jpg July 09, 2011, 10:53:33 PM
Keep up the great work everyone. We've seen some good steady growth today.
July 09, 2011, 01:49:14 PM
I have something about 10/42898 = 0.06%! And looks like this rate steadily decreasing since your last change. Good work!
Thanks!
Just a quick bit of news for everyone. I don't know how many of you are using the phatk kernel for your miner, but I just noticed they had an update: http://forum.bitcoin.org/index.php?topic=25860.0
My 5830 went from 295 MH/s to 306 MH/s just by updating the version of the phatk kernel phoenix was using.
I highly recommend everyone to check this out!
July 09, 2011, 12:00:28 PM
2 / 2304 * 100 = 0.08%
I have something about 10/42898 = 0.06%! And looks like this rate steadily decreasing since your last change. Good work! July 09, 2011, 11:01:37 AM
I have one more request - can you please make "stales" column in workers statistic? And if it will be separately for current round and since last reset (or for all time), this would be even better. I think pool which positioned as pool with lowest stales rate must have this info in statistic
Great idea!
We just started tracking this on a per worker basis in the middle of round 6. So, we will need to wait until round 7 begins to actually implement this. But, this is something I'd like to see as well.
Speaking of stales... I restarted my GPUs after we did our update a few hours ago. Here is a screen shot:
2 / 2304 * 100 = 0.08%
July 09, 2011, 07:04:33 AM
I have one more request - can you please make "stales" column in workers statistic? And if it will be separately for current round and since last reset (or for all time), this would be even better. I think pool which positioned as pool with lowest stales rate must have this info in statistic
+1
July 09, 2011, 03:46:45 AM
I have one more request - can you please make "stales" column in workers statistic? And if it will be separately for current round and since last reset (or for all time), this would be even better. I think pool which positioned as pool with lowest stales rate must have this info in statistic
July 09, 2011, 03:08:17 AM
Just a quick update for everyone. We just did a quick deploy with some fixes that will reduce stales even further.
Specifically we noticed a few miners were experiencing a higher than normal amount of stales following a longpool response. We isolated a specific corner case in our implementation and have significantly reduced our stale rate.
Nice!Specifically we noticed a few miners were experiencing a higher than normal amount of stales following a longpool response. We isolated a specific corner case in our implementation and have significantly reduced our stale rate.
Tomorrow I'll go in vacation, and when I will be back I wanna see this block completed!
@1bitc0inplz, don't worry, the 11 I will still post it since I'll have a notebook with wifi
July 09, 2011, 01:36:56 AM
Just a quick update for everyone. We just did a quick deploy with some fixes that will reduce stales even further.
Specifically we noticed a few miners were experiencing a higher than normal amount of stales following a longpool response. We isolated a specific corner case in our implementation and have significantly reduced our stale rate.
Specifically we noticed a few miners were experiencing a higher than normal amount of stales following a longpool response. We isolated a specific corner case in our implementation and have significantly reduced our stale rate.
Jump to: