[ Bonus PPS Pool - 130% PPS - OPEN ]

[DutchBrat]

Clipse:

The pool ran pretty good out of the gate, it seems though the longer it runs the worse it gets

Over the last hour or so I have been running at 5% - 6% rejects while earlier today it was steady between 2.5% and 3%

Brat

Edit: I don't see any pool 0 not providing work fast enough errors anymore, but the rejects after a LP change are back up to no good 

[1714172623]

1714172623

[1714172623]

1714172623

[1714172623]

1714172623

[1714172623]

1714172623

[1714172623]

1714172623

[MrTeal]

Running with the --net-delay argument in cgminer I'm seeing around 5-10% stales as well, usually in huge groupings. I'll get no stales or maybe one at each long poll, but then I'll have a long poll where it will punt out a string of 8 of them.

On the plus side, I've had no issues with disconnects or work not being provided fast enough which is a much more serious problem IMO. I've restarted two miners without net-delay to see if that might bring back the connection issues, or lower the rejects on that miner. I'll let you know what I see.

[DutchBrat]

Running with the --net-delay argument in cgminer I'm seeing around 5-10% stales as well, usually in huge groupings. I'll get no stales or maybe one at each long poll, but then I'll have a long poll where it will punt out a string of 8 of them.

On the plus side, I've had no issues with disconnects or work not being provided fast enough which is a much more serious problem IMO. I've restarted two miners without net-delay to see if that might bring back the connection issues, or lower the rejects on that miner. I'll let you know what I see.

That's a good idea, I will do the same, and you are right, Connection issues are much worse on your hash rate than the reject rate... at one point I was running 10% - 12% under my avg hash-rate.... So good job Clipse on solving that problem !

[rudrigorc2]

running flawless for 12hours  2.4% stale
10142 accepted shares

[DutchBrat]

running flawless for 12hours  2.4% stale
10142 accepted shares

Are you running CGMiner ?

I'm at 27665 Accepted 943 Rejected for a 3.41% reject rate

[rudrigorc2]

cgminer 2.3.3 -  5970 and single icarus @ win7-64

I use for both -Q 4

The numbers now are

5970: A:7230 R:246 SS:21 - 11hours 52min
icarus: A: 3466 R:99 SS:38 - 11hours 48min

I have a feeling they can still improve!


note the rejected by the 5970 is same as yours: 3.4%

[DutchBrat]

-Q 4

Interesting... I am going to run one with -Q4 --net-delay and one without --net-delay then compare them to those without -Q 4

We will get there in the end !

EDIT: I envy you guys, running at 7%+ reject rate now with/without --net-delay and/or -Q 4

I'll wait patiently until Clipse tackles this problem

[Zubilica]

Bamt - Pheonix - 1.648% . Going strong for now .

[Clipse]

I know for a fact that stales are higher than they should be thats why Im still running at 112% to cover atleast some of the not-so-nice stales.

The biggest issues that got resolved werent actually stale issues but server crashing/choking/stop feeding work etc and now I am going to get this stale issue under control so we can have a proper pool going

[Zubilica]

10997 total accepted shares
76 total invalid shares
0.686% reject ratio

It seem more stable now. I,m still waiting for 24-48h results.

[rudrigorc2]

reject here is dropping too.

1.5% the last 2 hours

[Clipse]

What Ive noticed, rejects occur mainly on gpus that is busy with a long hash ie. longpoll doesnt seem to correctly allways drop the old work that the gpu might be busy with.

[DutchBrat]

It is strange how it affects us all differently

rodrigorcm reject rate has been dropping over the last few hours, while mine is still steady between 5% - 6% (what CGMiner is reporting)

We are both mining on the same CGMiner version, I even switched one of my miners over to include -Q 4 but that is making no difference whatsoever on my end, --net-delay or a combination of those is also not helping me....

All very strange

In some rounds (in between LPs) it takes almost a minute before I get a share accepted 5-6 rejects in a row, so presumably it takes a minute before LP drops my work and my GPU hashes on the correct nonce....

Weird !!!

[MrTeal]

What Ive noticed, rejects occur mainly on gpus that is busy with a long hash ie. longpoll doesnt seem to correctly allways drop the old work that the gpu might be busy with.

Would the queue depth? Running cgminer with a queue of 2 it seems like it keeps working on the old work for quite awhile. This is my office desktop that's just running a single 6870 at about 300MH/s. It seems to be an excessive amount of rejects in one go, but it only happens infrequently. This would have happened at 19:00 UTC.

[2012-04-19 13:00:56] Accepted 00000000.cfcd80e0.104c0a23 GPU 0 thread 0 pool 0
[2012-04-19 13:01:05] LONGPOLL detected new block on network, waiting on fresh work
[2012-04-19 13:01:08] Rejected 00000000.45665b79.c108b63d GPU 0 thread 0 pool 0
[2012-04-19 13:01:24] Rejected 00000000.9ecfb9ed.6fa7198c GPU 0 thread 1 pool 0
[2012-04-19 13:01:51] Rejected 00000000.90610cd4.47e441ba GPU 0 thread 0 pool 0
[2012-04-19 13:01:52] Rejected 00000000.480ad701.1d873186 GPU 0 thread 0 pool 0
[2012-04-19 13:02:11] LONGPOLL detected new block on network, waiting on fresh work
[2012-04-19 13:02:12] Rejected 00000000.96d36cb2.e8aeea3a GPU 0 thread 0 pool 0
[2012-04-19 13:02:16] Rejected 00000000.057d0d22.74acf4a4 GPU 0 thread 1 pool 2
[2012-04-19 13:02:29] Rejected 00000000.d9de5e22.7769897f GPU 0 thread 0 pool 0
[2012-04-19 13:02:30] Rejected 00000000.d26cf9a7.866b3b22 GPU 0 thread 1 pool 2
[2012-04-19 13:02:31] Rejected 00000000.de395ccc.71ea79ec GPU 0 thread 0 pool 0
[2012-04-19 13:02:32] Rejected 00000000.b388c75a.6c83d933 GPU 0 thread 1 pool 2
[2012-04-19 13:02:52] Rejected 00000000.b82844cc.402565aa GPU 0 thread 0 pool 0
[2012-04-19 13:03:30] Accepted 00000000.5bd40d89.ad756376 GPU 0 thread 1 pool 0
[2012-04-19 13:03:33] Accepted 00000000.35d270d2.0a190f78 GPU 0 thread 1 pool 0
[2012-04-19 13:03:38] Accepted 00000000.122a4e56.0abf1b39 GPU 0 thread 0 pool 0

[DutchBrat]

What Ive noticed, rejects occur mainly on gpus that is busy with a long hash ie. longpoll doesnt seem to correctly allways drop the old work that the gpu might be busy with.

Would the queue depth? Running cgminer with a queue of 2 it seems like it keeps working on the old work for quite awhile. This is my office desktop that's just running a single 6870 at about 300MH/s. It seems to be an excessive amount of rejects in one go, but it only happens infrequently. This would have happened at 19:00 UTC.

[2012-04-19 13:00:56] Accepted 00000000.cfcd80e0.104c0a23 GPU 0 thread 0 pool 0
[2012-04-19 13:01:05] LONGPOLL detected new block on network, waiting on fresh work
[2012-04-19 13:01:08] Rejected 00000000.45665b79.c108b63d GPU 0 thread 0 pool 0
[2012-04-19 13:01:24] Rejected 00000000.9ecfb9ed.6fa7198c GPU 0 thread 1 pool 0
[2012-04-19 13:01:51] Rejected 00000000.90610cd4.47e441ba GPU 0 thread 0 pool 0
[2012-04-19 13:01:52] Rejected 00000000.480ad701.1d873186 GPU 0 thread 0 pool 0
[2012-04-19 13:02:11] LONGPOLL detected new block on network, waiting on fresh work
[2012-04-19 13:02:12] Rejected 00000000.96d36cb2.e8aeea3a GPU 0 thread 0 pool 0
[2012-04-19 13:02:16] Rejected 00000000.057d0d22.74acf4a4 GPU 0 thread 1 pool 2
[2012-04-19 13:02:29] Rejected 00000000.d9de5e22.7769897f GPU 0 thread 0 pool 0
[2012-04-19 13:02:30] Rejected 00000000.d26cf9a7.866b3b22 GPU 0 thread 1 pool 2
[2012-04-19 13:02:31] Rejected 00000000.de395ccc.71ea79ec GPU 0 thread 0 pool 0
[2012-04-19 13:02:32] Rejected 00000000.b388c75a.6c83d933 GPU 0 thread 1 pool 2
[2012-04-19 13:02:52] Rejected 00000000.b82844cc.402565aa GPU 0 thread 0 pool 0
[2012-04-19 13:03:30] Accepted 00000000.5bd40d89.ad756376 GPU 0 thread 1 pool 0
[2012-04-19 13:03:33] Accepted 00000000.35d270d2.0a190f78 GPU 0 thread 1 pool 0
[2012-04-19 13:03:38] Accepted 00000000.122a4e56.0abf1b39 GPU 0 thread 0 pool 0

Not sure if I understand how this works, but if LP fails to clear the queue and provide new nonce, then the less work is in the queue the better... GPU will idle because it has no work

As far as I see it, you can only tell CGMiner the minimum amount of work it has to keep in the queue, so it won't idle quickly, not a maximum amount of work

If I see it correctly, a queue of 0 getworks would be optimal, since the only reject that could occur is the hash the GPU is working on the moment a new block is found and LP changes before the share is delivered.... but it's not possible to set the queue to 0, only a minimum of 0

I will try 1 miner with -Q 0, see if that improves anything, I doubt it though, as default is 1

[rudrigorc2]

this reject sequency only happens when a new block is found, i did not noticed this happening earlier today...

[Beaflag VonRathburg]

4 machines, roughly 5 GH. I've been sitting between 1.9-4.5% rejected since the server swap. I was just sitting watching one of my machines to check on things and noticed one, waiting on work situation where it then rejected 9/11 shares before going stable. Followed, by a new block being found, a restart of LP, and 0 rejected shares. Instead, 5/6 of the shares went to my backup pool. Obviously, the latter is preferred, but I haven't been watching it long enough to give a full report.

[Clipse]

Something just went ape-sheet, Im having a look, sigh.

EDIT: Reset it , need to get to bed hopefully it doesnt fkout while I sleep but Im logging to see what is causing it if it doesnt occur again. I will send Thursday payments as soon as I wake up had some stuff to deal with and extremely tired atm so cant deal with that right now.

[DutchBrat]

I don't know what changed, but for the last 2.5 - 3 hours I have been running below 2% rejects on avg and that is mainly caused by 1 Lp that went haywire and caused 40% of my rejects

More than happy !!! Great JOB Clipse !  

EDIT: 2 horrible LPs in a row; went from 7300 accepted and 100 rejects to 7700 accepted and 200 rejects.... still is odd.... runs like a charm for 3 hours then 2 LPs that don't seem to reset work for a minute or so and then it is fine again.... well it is still a huge improvement over yesterday !

[Clipse]

Its definitely better after last reset, I got 1700 accepted and 5 rejected running a single gpu test over night (since my last message here)

Slowly but surely Ill get this sorted.