[ANN] ccminer 2.3 - opensource - GPL (tpruvot)

[flowjob]

on pool switch a device reset is made... that can explain the start

edit: no, only on algo switch, which is not the case here..

Well i checked the behavior 10 out of 10 times the locked thread is unlocked by the pool switch

so im gonna use that as monitoring action instead of killing the miner, it's more effective to get maximum hashrate but still i can't understand this bug, i tried out 10 different drivers version, always the same

Thanks for having looked into it anyway.

Hi. I have the same problem and a couple of days ago i "discovered" that doing a

Code:

strace -f -p $PIDofMiner

after starting ccminer and let it run for couple of seconds "solved" the problem. Very strange behavior...

[1714776507]

1714776507

[1714776507]

1714776507

[1714776507]

1714776507

[1714776507]

1714776507

[m1n1ngP4d4w4n]

on pool switch a device reset is made... that can explain the start

edit: no, only on algo switch, which is not the case here..

Well i checked the behavior 10 out of 10 times the locked thread is unlocked by the pool switch

so im gonna use that as monitoring action instead of killing the miner, it's more effective to get maximum hashrate but still i can't understand this bug, i tried out 10 different drivers version, always the same

Thanks for having looked into it anyway.

Hi. I have the same problem and a couple of days ago i "discovered" that doing a

Code:

strace -f -p $PIDofMiner

after starting ccminer and let it run for couple of seconds "solved" the problem. Very strange behavior...

Hummm very interesting indeed ! Happy to know im not alone in the world lol

[bughatti]

Team, I am trying to use the API, I have used the search function and found no help on this thread.  I have tried -b --api-bind 0.0.0.0:4068 and also -b --api-bind=0.0.0.0:4068   I have tried with 127.0.0.1  and also without the -b.  I am using ccminer 1.8.4  How do I get the api page to show up

[m1n1ngP4d4w4n]

Team, I am trying to use the API, I have used the search function and found no help on this thread.  I have tried -b --api-bind 0.0.0.0:4068 and also -b --api-bind=0.0.0.0:4068   I have tried with 127.0.0.1  and also without the -b.  I am using ccminer 1.8.4  How do I get the api page to show up

use -b OR --api-bind= not both

Code:

--api-bind 0.0.0.0:4068

or

Code:

-b 0.0.0.0:4068

Then you can use any raw TCP tool to get the API

Code:

echo help |nc your_mining_rig_ip or localhost 4068
summary
threads
pool
histo
hwinfo
meminfo
scanlog
seturl
quit

Code:

echo summary |nc your_mining_rig_ip or localhost 4068
NAME=ccminer;VER=1.8.4;API=1.8;ALGO=myr-gr;GPUS=6;KHS=390606.38;SOLV=0;ACC=386;REJ=8;ACCMN=55.407;DIFF=377.171377;NETKHS=0;POOLS=1;WAIT=0;UPTIME=418;TS=1487402397

there you go.

nc is netcat, it's a tcp raw tool, it exist on linux, and there are windows port also

[m1n1ngP4d4w4n]

For people struggling with the lbry algo and card being stuck (maybe im the only one aside from that linux guy earlier but whatever), here is the solution i came up with, it's not super elegant, but it work 100% of the time.

Do a conf for the algo instead of a command line.

LBC.conf

Code:

{
	"_note": "2 pool failover & switching enabled",

	"pools": [{
		"name": "YIIMP-LBC",
		"url": "stratum+tcp://yiimp.ccminer.org:3334",
		"user": "mywallet",
		"pass": "stats",
		"max-diff": 0.0,
		"scantime": 40,
		"disabled": false,
		"time-limit": 10
	}, {
		"name": "MINE-GPU4U-LBC",
		"url": "stratum+tcp://mine.gpu4u.eu:3334",
		"user": "mywallet",
		"pass": "stats",
		"max-diff": 0.0,
		"scantime": 40,
		"disabled": false
	}],

	"algo": "lbry",
	"intensity": 23.5,

	"api-bind": "127.0.0.1:4444",
	"api-remote": true,
	"statsavg": 20,

	"quiet": true,
	"debug": false
}

As you can see i put in setup a pool that only last for 10sec, after it will go to the 2nd one, you can also of course add pools for failsafe mesures. Since switching pool automatically unlock the locked GPU threads, 10 sec after miner start, it will change pool, and unlock threads, work 100% of the time.

@Epsylon3 : i really wish for a json API, and also logging features, would make my monitoring more powerful and easy to make. Those | & ; separated datas are really hard to parse when you begin to drill deep into other API function like threads/hwinfo and so on. Hope my wish will be heard Thanks.

[Epsylon3]

the "|" separator is made to be compatible with telnet, same protocol is used in cgminer and sgminer if you query them via sockets. I will not change that

[m1n1ngP4d4w4n]

the "|" separator is made to be compatible with telnet, same protocol is used in cgminer and sgminer if you query them via sockets. I will not change that

Okay, thanks anyway.

[m1n1ngP4d4w4n]

Do you know of any good app on android to monitor CCminer (i tried a few SG/CG miner compatible one (same API if i understood you correctly ? but they weren't very good and haven't been updated for years)

Also are you going reconsider adding Zcoin back to yiimp now that the network/wallet seems to be alot more stables ? would love to have this coin on yiimp.

Thanks

[Epsylon3]

might give a re-try on zcoin

[m1n1ngP4d4w4n]

Nice 

No ideas for the android app ?

[sp_]

24 hours estimates differences between zpool and yiimp.
20-30% lower estimates. Isn't zpool using the same software as yiimp? Why are the estimates so different..?

[Epsylon3]

the btc exchange fee i guess... else the curves are similar for me (a peak of almost 0.006 mbtc) at Europe and Canada hours

[joblo]

24 hours estimates differences between zpool and yiimp.
20-30% lower estimates. Isn't zpool using the same software as yiimp? Why are the estimates so different..?

Maybe he lowered the estimates by 20% to match the payouts.

[sp_]

I have a feature request on yiimp.

Can you add a sortable column in the benchmark tab MHASH/Watt. It should be interesting to see what setting that is giving the most hash out of little power.

I see that the H/W is available in the totals tab, but not in the per card statistics.

[Epsylon3]

Ok, and you please fix your intensity value, or at least the throughput value. Without that your data is excluded from averages

[sp_]

Yes, I will fix it later.

Is it possible to extract the current TDP setting of the card?

Would be interesting..

[m1n1ngP4d4w4n]

Yes, I will fix it later.

Is it possible to extract the current TDP setting of the card?

Would be interesting..

With nvidia smi you have a few queries that help with that

nvidia-smi.exe --query-gpu=index,timestamp,power.draw,clocks.sm,clocks.mem,clocks.gr --format=csv -l 1

this one show you an output like this

Code:

index, timestamp, power.draw [W], clocks.current.sm [MHz], clocks.current.memory [MHz], clocks.current.graphics [MHz]
0, 2017/02/27 17:18:05.101, 91.43 W, 90 MHz, 3299 MHz, 90 MHz
1, 2017/02/27 17:18:05.103, 113.66 W, 116 MHz, 3299 MHz, 116 MHz
2, 2017/02/27 17:18:05.108, 108.22 W, 52 MHz, 3299 MHz, 52 MHz
3, 2017/02/27 17:18:05.111, 112.54 W, 103 MHz, 3299 MHz, 103 MHz
4, 2017/02/27 17:18:05.117, 111.85 W, 116 MHz, 3299 MHz, 116 MHz
5, 2017/02/27 17:18:05.121, 113.02 W, 90 MHz, 3299 MHz, 90 MHz
0, 2017/02/27 17:18:06.125, 106.84 W, 65 MHz, 3299 MHz, 65 MHz
1, 2017/02/27 17:18:06.127, 107.54 W, 90 MHz, 3299 MHz, 90 MHz
2, 2017/02/27 17:18:06.130, 110.24 W, 65 MHz, 3299 MHz, 65 MHz
3, 2017/02/27 17:18:06.132, 113.33 W, 103 MHz, 3299 MHz, 103 MHz
4, 2017/02/27 17:18:06.135, 111.65 W, 116 MHz, 3299 MHz, 116 MHz

You can also calculate the current % of TDP for your card design with

./nvidia-smi -i 0 -q |grep -i power (where -i is your card ID)

it will give you as output

Code:

    Power Readings
        Power Management            : Supported
        Power Draw                  : 113.72 W
        Power Limit                 : 111.00 W
        Default Power Limit         : 185.00 W
        Enforced Power Limit        : 111.00 W
        Min Power Limit             : 92.50 W
        Max Power Limit             : 208.00 W

where you can see the power draw (actual) the power limit (your current tdp setting) the default power limit (the TDP @ 100%) the enforced power limit (where you set your TDP) , the minimal & maximal power limit (the 50% TDP and 112% TDP)

you can also format the output for easier parsing like so

Code:

nvidia-smi --query-gpu=index,power.draw.power.limit --format=csv

you can basically ask anything that nvidia smi show you using nvidia-smi --help-query-gpu to know the syntax like so.

Code:

nvidia-smi --query-gpu=index,power.draw,power.default_limit,enforced.power.limit,power.min_limit,power.max_limit --format=csv

will give you this output

Code:

index, power.draw [W], power.default_limit [W], enforced.power.limit [W], power.min_limit [W], power.max_limit [W]
0, 112.70 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W
1, 112.29 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W
2, 114.00 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W
3, 113.86 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W
4, 108.60 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W
5, 113.22 W, 185.00 W, 111.00 W, 92.50 W, 208.00 W

This is what i use to calculate maximum profit/tdp, and also i use it as a power monitoring (always useful to monitor process, hashrate, power)

[bathrobehero]

I use this when I'm testing stuff, which is very similar to m1n1ngP4d4w4n's:

Code:

echo off
:loop
cls
nvidia-smi --query-gpu=index,power.draw,enforced.power.limit,power.default_limit --format=csv
pathping 127.0.0.1 -n -q 1 -p 500 >nul
goto loop

which keeps looping something like this:

Code:

index, power.draw [W], enforced.power.limit [W], power.default_limit [W]
0, 163.68 W, 162.00 W, 180.00 W
1, 161.54 W, 162.00 W, 180.00 W
2, 163.15 W, 162.00 W, 180.00 W
3, 163.11 W, 162.00 W, 180.00 W
4, 163.90 W, 162.00 W, 180.00 W
5, 161.90 W, 162.00 W, 180.00 W

But do note that some cards (most 750 Ti) and some models doesn't use the correct power limit so the figures are wrong.

[Epsylon3]

i think the 750ti is only monitoring the gpu chip power, not the full card one.. (like the RX480)

[bathrobehero]

i think the 750ti is only monitoring the gpu chip power, not the full card one.. (like the RX480)

I think it's more about that smi only gets the power usage in a percentage format from the cards and calculates that using the default TDP (and modifiers) from the card's BIOS and most 750 Ti's BIOS' says the card can only use 38 or 46 watts depending on the model while all of them can pull 60 watts stock. So 50% usage on a 750 Ti will show as 19.25/38 or 23/46 watts watts instead of 30 watts. Some overclocked non-750 Ti models also have this issue but it's rare.