[OLD] Eligius: ASIC, no registration, no fee CPPSRB BTC + 105% PPS NMC, 877 #

[jamesg]

It's nice to have you here Giga. Stick around homie.

[helmax]

fuck eligius why stop SERVER 2

i lost 1 day of shares

[HellDiverUK]

i lost 1 day of shares

I doubt it...

[btcspender]

Looks like dedicated server for KNC november Jupiter is down (stratum+tcp://v2.srv.eligius.st:12234). Where all KNC miners with cgminer bug should be mining then? For now I switched to different pool.

Yeah I hope its back up soon. I am trying out slush at the moment seems to be working ok. I used to use ghash when eligius was down but who knows what is going on with them over this 51% uproar at the moment so I figure I'll just stay away.

fuck eligius why stop SERVER 2

i lost 1 day of shares

Thats unfortunate buddy, it will be back up soon hopefully. Just remember to be vigilant checking stats and this thread and knc thread I luckily caught it in about the first 30 mins

Edit: also dont forget backup pool

[wizkid057]

Can someone help me understand how and when nmc payouts are made?

Thanks,
James

Roughly daily.
 
As for the dev server for KnC workaround it is still up on stratum.mining.eligius.st:12234 ... v2 is having some connetivity issues.  Main stratum host is more redundant.

I'm still working on the KnC firmware fix.

Also, I'm aware that some stats are broken right now ( otably payout queue and possibly balance listings).  Pool is running normally. However I'm at the office and won't be back until about 4pm ET. I unfortunately left my physical devices needed to access Eligius servers (security) remotely at home this morning by accident, so I can't fix it from here today.

As always,  stats and pool are separate.  No earnings lost.

Thanks for mining!

-

[wizkid057]

Better reply now that I'm home.

Figures right after I catch up the payout queue a failsafe happens and puts it behind again. *sigh*  I'll try to do another tonight.  I like it cleaned up!

Anyhow, sorry for the confusion.  I've been trying to make it so that the stats maintain most of their functionality regardless of the status of the reward system code.  Seems I have a few kinks to work out on that front still.

I'm rebuilding the share log for the past half-day to make sure everything is correct.  I don't foresee any issues, but, you never know.  This is expected to take roughly 90-120 minutes, at which point the stats should automatically start updating again.

As always, all shares are accounted for.  The pool will not accept shares if they can't be stored and tallied, so, regardless of the status of the stats pages, if the pool accepts your shares then all is well and you will be paid as you should be for those shares.  Eligius is setup to be very redundant in this manner.

Thanks for mining!

-wk

[kendog77]

fuck eligius why stop SERVER 2

i lost 1 day of shares

You didn't setup a backup pool? Really? Rookie mistake.

[gravityz3r0]

Better reply now that I'm home.

Figures right after I catch up the payout queue a failsafe happens and puts it behind again. *sigh*  I'll try to do another tonight.  I like it cleaned up!

Anyhow, sorry for the confusion.  I've been trying to make it so that the stats maintain most of their functionality regardless of the status of the reward system code.  Seems I have a few kinks to work out on that front still.

I'm rebuilding the share log for the past half-day to make sure everything is correct.  I don't foresee any issues, but, you never know.  This is expected to take roughly 90-120 minutes, at which point the stats should automatically start updating again.

As always, all shares are accounted for.  The pool will not accept shares if they can't be stored and tallied, so, regardless of the status of the stats pages, if the pool accepts your shares then all is well and you will be paid as you should be for those shares.  Eligius is setup to be very redundant in this manner.

Thanks for mining!

-wk

Thanks for your hardwork and efforts wk,

For your information, your dev patch for Nov Jupiter is already almost as perfect as it can be. It's even better than any KnC firmwares they have churn out so far, and even better than your own dedicated KnC server. They should really hire you as their staff. Hehe!

Before applying this fix at stock KnC .99E firmware. I'm getting only 670GH/S hardware side, and coupled with 8-10% HW error, poolside hashrate always ended up around 600+ GH/S only.

Now with the same firmware, main pool (not KnC) but with your "temporary" fix, i get 690GH/S at hardware side, and although with the higher HW Reject of almost 2%, the HW Error itself is almost negligbile (~0.05%). Therefore with a total combination error of 2% (WU: 9530 stable), the pool side hash rate (12 hrs average) is amazingly stable at 675+ GH/S (690GH/S-2%) . What more can i ask for? Now, if only the rejects can be eliminated

http://s17.postimg.org/5gotykwi3/Untitled.jpg

I have never even gotten anywhere close to 690GH/S  hardware side at any other mining pool before. Most i'm getting is about 680GH/S (as shown in cgminer), and with the relatively same HW error of 1.5%, the poolside hashrate always ended up below 670GH/S no matter what i do.

I'm curious on what magic did you put into place to have the hardware running at 690GH/S when all other pools fails to do so? I'm pretty sure it's not a false indication because the pool side hashrate is reflecting the correct net hashrate after 2% deduction. I really don't think you can squeeze anything more out of this, can you? 2% HW is already considered low, unless you can bring it down further to 1% while keeping the 690GH/S as shown on cgminer


You are not obligated to provide a any fix for KnC and yet you did. Keep up the good job!

[lenny_]

Thanks for your hardwork and efforts wk,

For your information, your dev patch for Nov Jupiter is already almost as perfect as it can be. It's even better than any KnC firmwares they have churn out so far, and even better than your own dedicated KnC server. They should really hire you as their staff. Hehe!

Where this patch can be found?

[gravityz3r0]

Thanks for your hardwork and efforts wk,

For your information, your dev patch for Nov Jupiter is already almost as perfect as it can be. It's even better than any KnC firmwares they have churn out so far, and even better than your own dedicated KnC server. They should really hire you as their staff. Hehe!

Where this patch can be found?

I found this from another KnC thread. This is the link to the file posted by wk himself :

https://bitcointalk.org/index.php?topic=313978.msg4229913#msg4229913 (post #235)

If you are wondering how to apply it to the Nov Jupiter, you need to have access to SSH and enter the following command :

Code:

wget http://vpn.wizkid057.com/nas/cgminer-binary-wizdev20131222
chmod +x cgminer-binary-wizdev20131222
screen -dm ./cgminer-binary-wizdev20131222 -c /config/cgminer.conf
screen -d -r

To restore, just reboot your machine or reflash stock KnC firmware. You will know it's working if you see the CGminer version is upgraded to 3.8.5 which is provided by wk patch.
Stock 0.99E run 3.8.1, and the newer 0.99.1E runs 3.9.0. I find his patch perform better on the stock 0.99E firmware.

[joeventura]

Are the October Jupiters affected by this problem?

[j03]

Are the October Jupiters affected by this problem?

AFAIK, no. I have an October unit with 2 upgrade modules and it's doing just fine on firmware 0.99. Someone correct me if I'm wrong.

[merv77]

Thanks for your hardwork and efforts wk,

For your information, your dev patch for Nov Jupiter is already almost as perfect as it can be. It's even better than any KnC firmwares they have churn out so far, and even better than your own dedicated KnC server. They should really hire you as their staff. Hehe!

Where this patch can be found?

I found this from another KnC thread. This is the link to the file posted by wk himself :

https://bitcointalk.org/index.php?topic=313978.msg4229913#msg4229913 (post #235)

If you are wondering how to apply it to the Nov Jupiter, you need to have access to SSH and enter the following command :

Code:

wget http://vpn.wizkid057.com/nas/cgminer-binary-wizdev20131222
chmod +x cgminer-binary-wizdev20131222
screen -dm ./cgminer-binary-wizdev20131222 -c /config/cgminer.conf
screen -d -r

To restore, just reboot your machine or reflash stock KnC firmware. You will know it's working if you see the CGminer version is upgraded to 3.8.5 which is provided by wk patch.
Stock 0.99E run 3.8.1, and the newer 0.99.1E runs 3.9.0. I find his patch perform better on the stock 0.99E firmware.

I've tried 'cgminer-binary-wizdev20131222' as described above in my October Jupiter and hash rate up from 550GH/s to 573GH/s, HW errors down from 1.8% to 0.7% and Rejects about 0.8%
(but not running very long yet, but I will re-post after running for 12 hours.  BTW I'm runnig 0.95 firmware)

but I haven't seen this before in cgminer when viewing with 'screen -r'... eg. KnC: HW Error XXXXXXXX - work_id: XXXXX   <-- is it normal to see many of these?

[wizkid057]

Yes. Its debugging output I added.

[merv77]

thanks wizkid..
rejects seem on the rise, but I I'm liking the extra hash rate 
I'll report back after 12 hour run.
Cheers

[CeeCee]

i have tried your cgminer version and for the last 5 minutes cgminer shows me an average plus of around 15GH/s on my two october units. don't tried your binary beforce because i was thinking it is only helpful for november units. lets see what the next 12h will show me on the stats page. did you suggest to use october units too on port 12234? and i have also a loooooot of these hw error debug messages.

[wizkid057]

My cgminer binary will get slightly higher rejects but much much lower HW errors.  The tradeoff seems to work out well, though.  The spewing of HW error messages is normal, just a debugging thing I added.  Every single HW error has info shown for it, which is why there are a lot of messages.

You can save a little sanity by running cgminer without the console or normal console output.  Just add -q -T to the end of starting command. (The one where you specific the location of cgminer.conf)
This still shows the HW error messages, but nothing else.

-wk

[lenny_]

On this version of your cgminer - cgminer-binary-wizdev20131222 - you disabled work flushing? For how long after work restart cores are still reporting old work (doing rejects)?
Is it possible to do opposite thing, and optimize flushing so Jupiters will work better on alternative coins with more frequent work restart signal, or with p2pool (30s work restart)?

[wizkid057]

On this version of your cgminer - cgminer-binary-wizdev20131222 - you disabled work flushing? For how long after work restart cores are still reporting old work (doing rejects)?
Is it possible to do opposite thing, and optimize flushing so Jupiters will work better on alternative coins with more frequent work restart signal, or with p2pool (30s work restart)?

I'm working on fixing flushing, but there may be a hardware issue at work preventing me from doing so.

Disabling flushing doesn't have a set length of time where cores still report old work.  Some immediately are finishing old work and start working on new work, others take a few seconds or more depending on if they recently started the old work or not.

-wk

[smooth]

Are payouts backed up or stuck or is this something specific to my account?