OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.1

[spiccioli]

Yes, I'm persistent.

I've found that with a -g 1 it does not happen anymore (it's been running for half a hour now), but I loose a couple Mh/s.

spiccioli.

ps. I had to install curl since it's not installed in linuxcoin by default.

BTW, cgminer found 128 as optimal worksize, but using -w 256 gives me 3-5Mh/s more juice, steady

So now I'm forcing it to -w 256 ( and -g 1 -I 8 -w 256 -Q 2 ).

It is running at 361 on my 880/270 5850 and 444 on my 980/280 5870

spiccioli.

[1714057294]

1714057294

[1714057294]

1714057294

[1714057294]

1714057294

[1714057294]

1714057294

[1714057294]

1714057294

[spiccioli]

Yes, I'm persistent.

I've found that with a -g 1 it does not happen anymore (it's been running for half a hour now), but I loose a couple Mh/s.

spiccioli.

ps. I had to install curl since it's not installed in linuxcoin by default.

Do you have linuxcoin final? Everything should just work with it.
[/quote]

No, I'm sticking with v0.2b, since it's working ok for me and I don't need all the network boot stuff.

spiccioli.

[Meatball]

Anyone having problems with SDK 2.4/2.5 and CGMiner 1.5.1 on Windows?  With SDK 2.1 I fire up fine, but if I try 2.4 it crashes immediately on launch.

[Viceroy]

Anyone having problems with SDK 2.4/2.5 and CGMiner 1.5.1 on Windows?  With SDK 2.1 I fire up fine, but if I try 2.4 it crashes immediately on launch.

Windows version is on hold while we await the developer who is working on it to show up.

[dostortugas]

Yes, I'm persistent.

I've found that with a -g 1 it does not happen anymore (it's been running for half a hour now), but I loose a couple Mh/s.

spiccioli.

ps. I had to install curl since it's not installed in linuxcoin by default.

BTW, cgminer found 128 as optimal worksize, but using -w 256 gives me 3-5Mh/s more juice, steady

So now I'm forcing it to -w 256 ( and -g 1 -I 8 -w 256 -Q 2 ).

It is running at 361 on my 880/270 5850 and 444 on my 980/280 5870

spiccioli.

MH rate gets irrelevant at the point where shares/minute come to play, I can get +10MH with -I 14 but -2 shares/minute, 128 or 256 doesn't make a difference in terms of shares/per minute, got one rig with 3 5850s and one with 1 plus a 6950... after a couple of hours the 3x does ~1080MHs/s ~15 shares/min @"-w 256" and the other with "-w 128" does 729MHs/s ~10 shares both with "-I 8". It might do a difference in the hashrate but the shares/minute seem to be stable at the same rate after a couple of hours.
cards compared directly are about the same "U" but different hashrates...

1+1!=3

[sirky]

Anyone having problems with SDK 2.4/2.5 and CGMiner 1.5.1 on Windows?  With SDK 2.1 I fire up fine, but if I try 2.4 it crashes immediately on launch.

On one of my boxes it works perfectly. On another I get the 0 byte whatever error.

So, sort of.

[Meatball]

Yeah, I've tried 2 separate boxes and getting the 0 byte error as well.

[zaytsev]

Looks like the 1.5.3 configure script inadvertently dropped support for alternate cURL library path support via --with-libcurl:

Yes, because it's the crappy way to do it. The right way is to export a variable to specify the PKG_CONFIG_PATH to the correct *.pc file. See my RPMs for reference...

Seems to be a way that works on more systems  and with local built binaries is the right way to do it.  Dependance on a packaging system  is the crappy way to do it.

No offense, but do read up on pkgconfig a little bit before commenting. For how long have you been involved in packaging? I've told what's the correct way of doing it with pkconfig and locally built binaries.

[zaytsev]

I don't want to talk about the reasons why I cpu-mine. It leads to nothing but arguments.

All I want to know is how to get cgminer to work without an openCl-device.

1.5.3 works for me for CPU mining, did you specify -g 0?

[Ali]

Yes, I did but I guess you're not on Windows since there's no Windows binary for 1.5.3 yet.

Also remember that I really don't have any cl-device in my system.

[twobits]

Looks like the 1.5.3 configure script inadvertently dropped support for alternate cURL library path support via --with-libcurl:

Yes, because it's the crappy way to do it. The right way is to export a variable to specify the PKG_CONFIG_PATH to the correct *.pc file. See my RPMs for reference...

Seems to be a way that works on more systems  and with local built binaries is the right way to do it.  Dependance on a packaging system  is the crappy way to do it.

No offense, but do read up on pkgconfig a little bit before commenting. For how long have you been involved in packaging? I've told what's the correct way of doing it with pkconfig and locally built binaries.

Why?  You seem to miss the point completly.  I don't have pkgconfig.  I don't want pkgconfig, so why should I read up on it?  I have been building software from source for a few decades now.   Since usenet was the way to distribute the sources for most.   The all the worlds linux folks are as annoying today as the all the worlds a vax foks where then.

[zaytsev]

Why?  You seem to miss the point completly.  I don't have pkgconfig.  I don't want pkgconfig, so why should I read up on it?  I have been building software from source for a few decades now.   Since usenet was the way to distribute the sources for most.   The all the worlds linux folks are as annoying today as the all the worlds a vax foks where then.

If you have been building software for that long, it should have taught you that building complex software with many dependencies without an unified system for discovering library paths, compile, link flags etc. (which is exactly what pkg-config is) is a living hell of hooking up each and every library to the build system via crappy third-party autotools macros for which no decent unified and maintained archive exists. And in this particular instance, these macros were even unable to resolve version requirements for which reason (I guess) they were finally nuked in favor of pkg-config.

It could be that your system doesn't ship pkg-config (which I would doubt, unless it's Windows), but since you've been building software since vax days, it would certainly take you near-zero effort to build pkg-config in addition to cgminer. I bet there are even third-party repositories with pre-build binaries somewhere around.

[twobits]

Why?  You seem to miss the point completly.  I don't have pkgconfig.  I don't want pkgconfig, so why should I read up on it?  I have been building software from source for a few decades now.   Since usenet was the way to distribute the sources for most.   The all the worlds linux folks are as annoying today as the all the worlds a vax foks where then.

If you have been building software for that long, it should have taught you that building complex software with many dependencies without an unified system for discovering library paths, compile, link flags etc. (which is exactly what pkg-config is) is a living hell of hooking up each and every library to the build system via crappy third-party autotools macros for which no decent unified and maintained archive exists. And in this particular instance, these macros were even unable to resolve version requirements for which reason (I guess) they were finally nuked in favor of pkg-config.

It could be that your system doesn't ship pkg-config (which I would doubt, unless it's Windows), but since you've been building software since vax days, it would certainly take you near-zero effort to build pkg-config in addition to cgminer. I bet there are even third-party repositories with pre-build binaries somewhere around.

The point is I don't want to install pkg-config. I know exactly where say libcurl is,  as I of course build and install that from source.   If there is another dependency I would do the same.  At least with makefiles I could also just edit them easily myself quickly.  Autotools already makes that a longer cycle to fix when things go wrong.  What does it hurt to at least give the option for me to tell it exactly where to get the library it wants to link to?  Put in the README/INSTALL the version dependencies (though the library itself should offer some  way to check this, and let it abort if it is a bad version).

What is wrong with  leaving in the manual override option?  What is crappy about leaving people  the option to not install more abstraction layers on things they don't want?  If I want enough abstraction layers I may as well just go use windows and autoupdates.

[Meatball]

Just added another card to my PC so now im running 2 in same machine.
CPU usage went from 0-1% (using 1 card) up to 90% (using both cards)
Running 1 process for each card brings it down a bit (15-20% cpu per process) but its still higher then when there was just the one card in the PC.

Had a look though a few of the posts to see if there was a fix and running it with affinity to core 0 at least limits its and brings power usage down a bit but its still using more power then it should be.

Windows 7 32bit.
SDK 2.4
Catylist 11.6 running dummy plugs.

There any other way to set affinity and lock down the usage to a single core?  Might help with the 100% bug.

[zaytsev]

What is wrong with  leaving in the manual override option?  What is crappy about leaving people  the option to not install more abstraction layers on things they don't want?  If I want enough abstraction layers I may as well just go use windows and autoupdates.

Did you actually read ./configure --help at some point? Is it your vax experience that did not prevent you from blaming pkg-config that is preventing you from exporting LIBCURL_LIBS and LIBCURL_CFLAGS before running ./configure if so you wish? --with-libcurl option was broken and hence removed. If you don't want to use pkg-config, fine, override the flags and there you go. I have just checked it by uninstalling pkg-config and it works fine.

[xcooling]

There any other way to set affinity and lock down the usage to a single core?  Might help with the 100% bug.

Use the following in your  .bat file to launch cgminer, will force cpu affinity to a single core.. which will max only that specific core out to 100%.

Code:

start /AFFINITY 0x1 /NORMAL cgminer.exe  -o http://someurl.com:8344

[PLaci1982]

Just added another card to my PC so now im running 2 in same machine.
CPU usage went from 0-1% (using 1 card) up to 90% (using both cards)
Running 1 process for each card brings it down a bit (15-20% cpu per process) but its still higher then when there was just the one card in the PC.

Had a look though a few of the posts to see if there was a fix and running it with affinity to core 0 at least limits its and brings power usage down a bit but its still using more power then it should be.

Windows 7 32bit.
SDK 2.4
Catylist 11.6 running dummy plugs.

There any other way to set affinity and lock down the usage to a single core?  Might help with the 100% bug.

For Win7, I recommend you this 2 websites:

http://technet.microsoft.com/en-us/library/ff384148%28WS.10%29.aspx
http://www.sevenforums.com/tutorials/419-run-different-user.html

[Meatball]

Nice, looks like that works.  If anyone is using Windows 7 and want's something similar to what I'm doing I added the batch file code below.  Basically I've created a single batch file that includes all the information needed to launch the cgminer processes for all of my miner's and stuck it in a network share.  I then set up a scheduled task on the mining machines set to run at login and launch the batch file in the network share.

This way I can make mass configuration changes all in one location instead of having to jump to individual boxes and modify each one.

Code:

REM Grab Local Machine Name and set to variable %host%

set host=%COMPUTERNAME%

REM Check variable %host% to determine and launch correct machine subroutine.

if %host% == Miner1 goto:Miner1
if %host% == Miner2 goto:Miner2
goto:eof

REM Individual machine launch subroutines - Timout value to allow for MSI Afterburner to fully load before miners start up.

:Miner1
echo Miner 1 Batch Startup...
TIMEOUT /T 75
cd /d C:\cgminer-1.5.1
start "Miner 1 CGMiner Status" /AFFINITY 0x1 /NORMAL cgminer -o http://mainserver:port -u userid -p password -I 8 -o http://backupserver:port -u userid -p password
goto:eof

:Miner2
echo Miner 2 Batch Startup...
TIMEOUT /T 75
cd /d C:\cgminer-1.5.1
start "Miner 2 CGMiner Status" /AFFINITY 0x1 /NORMAL cgminer -o http://mainserver:port -u userid -p password -I 8 -o http://backupserver:port -u userid -p password
goto:eof

:eof

I don't believe the timeout function is available on other versions of windows, but this seems to work fine on W7.  It can easily be expanded out for multiple boxes as well by just replicating the host variable checks and subroutines for more machines.

[sirky]

Nice, looks like that works.  If anyone is using Windows 7 and want's something similar to what I'm doing I added the batch file code below.  Basically I've created a single batch file that includes all the information needed to launch the cgminer processes for all of my miner's and stuck it in a network share.  I then set up a scheduled task on the mining machines set to run at login and launch the batch file in the network share.

This way I can make mass configuration changes all in one location instead of having to jump to individual boxes and modify each one.

Code:

REM Grab Local Machine Name and set to variable %host%

set host=%COMPUTERNAME%

REM Check variable %host% to determine and launch correct machine subroutine.

if %host% == Miner1 goto:Miner1
if %host% == Miner2 goto:Miner2
goto:eof

REM Individual machine launch subroutines - Timout value to allow for MSI Afterburner to fully load before miners start up.

:Miner1
echo Miner 1 Batch Startup...
TIMEOUT /T 75
cd /d C:\cgminer-1.5.1
start "Miner 1 CGMiner Status" /AFFINITY 0x1 /NORMAL cgminer -o http://mainserver:port -u userid -p password -I 8 -o http://backupserver:port -u userid -p password
goto:eof

:Miner2
echo Miner 2 Batch Startup...
TIMEOUT /T 75
cd /d C:\cgminer-1.5.1
start "Miner 2 CGMiner Status" /AFFINITY 0x1 /NORMAL cgminer -o http://mainserver:port -u userid -p password -I 8 -o http://backupserver:port -u userid -p password
goto:eof

:eof

I don't believe the timeout function is available on other versions of windows, but this seems to work fine on W7.  It can easily be expanded out for multiple boxes as well by just replicating the host variable checks and subroutines for more machines.

Did you fix the 0 byte issue?

[Meatball]

Did you fix the 0 byte issue?

No, it's happening on 3 machines, any of which have SDK 2.4 on them.  I'm trying to roll back to 2.1 on some of them, but they're giving me some issues.