-ck
Legendary
Offline
Activity: 4284
Merit: 1645
Ruu \o/
|
|
November 08, 2014, 08:41:45 PM Last edit: November 08, 2014, 11:06:55 PM by ckolivas |
|
Wow... Are we lucky or what...
That is some crazy performance - Kano or CK - Have you ever though of trying some "Multi Arm Bandit" type of logic on the parts of the coin to process? - We use in on our Ad platform, and have some really good performance specs from it.
Yes we have been lucky but there is nothing that will guarantee we can maintain that. We will have bad luck in the future, it is guaranteed, but things always even out over time and that's no reason not to enjoy the runs of luck. While the performance of the pool is very good, what you're attributing to performance is luck and the mechanism you speak of doesn't exist. There is no hardware or software mechanism to affect luck. That's basing development on a cognitive bias known as "clustering illusion" and I suggest you read up about it. Please don't try to convince two of the most informed mining developers otherwise as you'll be wasting your time and taking the discussion offtopic for this thread.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 08, 2014, 09:50:21 PM |
|
Nope can't get the I menu either, can get help on the right, and home works on the left too.
seems just the middle doesn't work
Someone has sent a pull that I'll check on all the browsers I've checked on. Then I'll put it live. If that doesn't positively affect the iDevices then I'll do what I should have done with the 'i' - add them to help (yeah it's a hack, just like the 'i' change is a hack, but of course it will allow you to click on them)
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 08, 2014, 11:40:42 PM |
|
Sent out the payments for the last 3 matured blocks: 329041 329073 and 329075 Transactions have confirmed already also.
|
|
|
|
Trademan1
Member
Offline
Activity: 84
Merit: 10
|
|
November 09, 2014, 12:09:08 AM Last edit: November 09, 2014, 12:43:53 AM by Trademan1 |
|
Gday Kano and CK, ive just pointed my 2 S3s at your pool.
All the GHs adds up eh. Ive also had about 50GHs made up of a few USBs running at the solopool for a while now, just in case my lotto number comes up. Yanevano.
The S3s settled down showing pretty much what they should on your pool within 1 hr so its all looking pretty good.
The site looks good too, its got all the info we need, I just need to work out a way of not having to refresh the website to see the updated shares etc.
Anyway, well done with the pool, and lets hope it keeps growing and getting better.
Cheers.
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 09, 2014, 09:34:57 AM |
|
Edit: oh, also, I paid out the last block, 329091, shortly after it matured I implemented the majority of the Marker changes on the pool over the last hour. https://bitcointalk.org/index.php?topic=790323.msg9417792#msg9417792Still a few more of the changes to do but I've done all the major design enhancements required. The older blocks that weren't showing their stats before on the web page, now have their data they need that I deprived them of. The next change regarding Markers will be a minor change to the payout. The payout is (and will stay) N based on the network difficulty when a block is found. However, even now it isn't exactly N, its a bit more. Since CKDB doesn't store shares singly, but rather as sharesummaries per ~30s, the N value used if N+the rest of the shares in the starting summary Thus this currently averages ~15s more shares above N, per payout. With the markersummaries, the sharesummaries are summarised into markersummaries that will be (probably) 50 sharesummaries each - i.e. ~25 minutes of shares per markersummary. So the N value used will be, on average, ~12.5 minutes more shares per payout (on average ~12 minutes more shares per payout than currently used) I'll post again in here (whenever I get to doing that) before I make that change.
|
|
|
|
Axiste
|
|
November 09, 2014, 10:24:49 AM |
|
ok, menus still no good, now i can only click on help when im not logged in. Once logged in its the same as the others. Seems to be related to them not being available unless logged in as a user.
|
If my advice has helped you out, feel free to throw some Satoshi's my way! BTC: 1Nq7hrRDamdnNiTBaFxEn5DuZYS9kD3tbJ Running a Full Node
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 09, 2014, 09:43:12 PM |
|
ok, menus still no good, now i can only click on help when im not logged in. Once logged in its the same as the others. Seems to be related to them not being available unless logged in as a user.
You can only access the Home and the Help->Payouts page unless you are logged in. Edit: and we just found another block under 100%
|
|
|
|
Xtra7973
Newbie
Offline
Activity: 47
Merit: 0
|
|
November 09, 2014, 10:05:35 PM |
|
And I just got all my miners up and running, damn internet problems. Just found another block, lets hope my internet demons stay away this time.
|
|
|
|
norgan
|
|
November 09, 2014, 10:08:35 PM |
|
What's that now 7 blocks in 24 hours?
|
|
|
|
Trademan1
Member
Offline
Activity: 84
Merit: 10
|
|
November 09, 2014, 10:24:49 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
|
|
|
|
norgan
|
|
November 09, 2014, 10:41:47 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
ah yeah that's true. it just feels like 7 in 24 hours haha
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 09, 2014, 10:58:34 PM Last edit: November 09, 2014, 11:26:39 PM by kano |
|
What's that now 7 blocks in 24 hours?
The blocks page is all OK again since I implemented the Markers yesterday I put a fix in there to make sure it was OK even when I used the new -w option to reduce ram while I was working on the Markers Before markers, CKDB (using a -w to greatly reduce the number of sharesummaries in ram) was using > 40GB ram After Markers basic implementation it's now down to about 10GB and will drop more once I implement the longer marker shifts. Yeah ckdb still has a memory image of the entire database. There's a bunch of todos in the code to drop history from ram once I no longer need it - but the marker changes were required before that and at the moment it's back again well under control since the back end has more than enough ram and performance. Edit: the % at the bottom is since the pool started and also correct i.e. it doesn't count the orphan block as a block, so it's 2.69% just adds (correctly) on to the total
|
|
|
|
philipma1957
Legendary
Online
Activity: 4298
Merit: 8793
'The right to privacy matters'
|
|
November 09, 2014, 11:32:26 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
ah yeah that's true. it just feels like 7 in 24 hours haha actually we did 7 in 27 1/2 hours first of the 7 was block made 2014-11-07 06:44am last of the 7 was 2014-11-08 10:18 am see below
|
|
|
|
norgan
|
|
November 09, 2014, 11:34:01 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
ah yeah that's true. it just feels like 7 in 24 hours haha actually we did 7 in 27 1/2 hours first of the 7 was block made 2014-11-07 06:44am last of the 7 was 2014-11-08 10:18 am Awesome, see I wasn't that far off after all. Go ckpool!
|
|
|
|
judypug1956
|
|
November 09, 2014, 11:36:31 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
ah yeah that's true. it just feels like 7 in 24 hours haha actually we did 7 in 27 1/2 hours first of the 7 was block made 2014-11-07 06:44am last of the 7 was 2014-11-08 10:18 am Awesome, see I wasn't that far off after all. Go ckpool! it was a tremendous run of luck 1ph at current diff should take 13 days to do what we did in 27.5 hours.
|
1956jUdYPFwiBSzt9AECdWj3KE4WV7taiM I can't do 1957philma.. for btc address the i are not allowed This is a secondary account for Philipma1957, don't do business with this account deal with philipma1957
|
|
|
philipma1957
Legendary
Online
Activity: 4298
Merit: 8793
'The right to privacy matters'
|
|
November 09, 2014, 11:39:29 PM |
|
BTW the last block was under 50% cdf as I understand it.
I believe 39603666252 is the current diff. If you take exactly that number you are at 50% cdf.
Last block was 86.1% of that number.
@kano or ckolivas is that correct? if we match the diff shares cdf is 50%
|
|
|
|
Trademan1
Member
Offline
Activity: 84
Merit: 10
|
|
November 09, 2014, 11:56:22 PM |
|
No, the last block took about 35hrs. There was 7 blocks over about 4 days though. The one before that took about 4 1/2 days.
ah yeah that's true. it just feels like 7 in 24 hours haha actually we did 7 in 27 1/2 hours first of the 7 was block made 2014-11-07 06:44am last of the 7 was 2014-11-08 10:18 am see below Depends which way you look at it. The block at 2014-11-07 06:44am took about 4 1/2 days .. actuallly Either way, not bad for only 1.2Ph.
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 10, 2014, 12:27:37 AM |
|
BTW the last block was under 50% cdf as I understand it.
I believe 39603666252 is the current diff. If you take exactly that number you are at 50% cdf.
Last block was 86.1% of that number.
@kano or ckolivas is that correct? if we match the diff shares cdf is 50%
The figure I have there on the blocks page is the % (the Diff amount next to it is) of network diff when the block was found. I'm pretty sure CDF isn't exactly 50% when you match 100% diff. Anyway, the colouring is black at 100%, red above 100% and green below 100% The brighter the green, the better, the brighter the red, the worse, though red hits it's brightness limit at 562% https://bitbucket.org/ckolivas/ckpool/src/aca254495ebfc20f9358995d3859b4a2ffd138dc/pool/page_blocks.php?at=master#cl-3
|
|
|
|
kano (OP)
Legendary
Offline
Activity: 4620
Merit: 1851
Linux since 1997 RedHat 4
|
|
November 10, 2014, 01:00:42 AM |
|
...
Depends which way you look at it.
The block at 2014-11-07 06:44am took about 4 1/2 days .. actuallly
Either way, not bad for only 1.2Ph.
Yeah, the time for a single block is from the previous block to the time of the block. Although you can say we found 7 blocks in a bit over a day, all the statistical calculations including the 1st block of those seven has to include the ~4.42days it took to find it. I'm still amazed that the 6 blocks averaged less than 10.4% diff for all 6 ...
|
|
|
|
hurricandave
Legendary
Offline
Activity: 966
Merit: 1003
|
|
November 10, 2014, 01:00:57 AM |
|
I don't know what could be so different about the color settings for safari vs. IE 11 but I can't make out any of the dark shaded % areas. Its like the offset colors for the digits isn't happening or lacks a proper brightness control to make them visible.
|
|
|
|
|