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

[kano]

No doubt my usbfail.c test would show that libusb not working either ....

[-ck]

Unofficial Mac binaries updated to 3.4.3 at http://spaceman.ca/cgminer.

Just tried them on a Mac Mini with Snow Leopard and a single Block Erupter USB, give the USB timeout errors:

Code:

 [2013-09-17 14:00:28] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:30] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:32] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:34] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:36] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:37] Stratum from pool 0 detected new block
 [2013-09-17 14:00:38] AMU0: TIMEOUT GetResults took 1998ms but was 100ms

Was there ever a version this didn't happen on? If not, we're kinda screwed until a better libusb for osx comes out since this is a libusb bug which is mostly addressed in the linux and windows versions, yet you're compiling against the same libusb version now (which is included statically in the cgminer source tree now).

[-ck]

No worries - I've done this before for you, happy to do it again.

Let me know when you have 3.4.3 ready

New debug executables uploaded.

[The Mafia]

New debug executables uploaded.

All done and started, now we just have to play the waiting game.

I haven't seen it crash for at least 2-3 days on 3.4.3

[-ck]

New debug executables uploaded.

All done and started, now we just have to play the waiting game.

I haven't seen it crash for at least 2-3 days on 3.4.3

Appreciated. That's no problem as no new release is scheduled any time soon unless I find some hotfix bug worth fixing.

I'm currently swamped with other work and probably need a small break from regular cgminer coding anyway.

[vekt0r7]

New debug executables uploaded.

All done and started, now we just have to play the waiting game.

I haven't seen it crash for at least 2-3 days on 3.4.3

Appreciated. That's no problem as no new release is scheduled any time soon unless I find some hotfix bug worth fixing.

I'm currently swamped with other work and probably need a small break from regular cgminer coding anyway.

Really?
How about this? Are you able to reproduce this?

Running 3.4.3 but notice now (it seems this issue goes back a few versions) that the accepted blocks field now shows the latest difficulty (x number of blocks found it seems) instead of total block count? Anyone got a fix for this?

Not sure what coin you're mining, but on the bitcoin network, cgminer now shows the number of "diff 1 shares" found for the accepted field.  So, if you're pool has a higher share difficulty, the counter will increase by whatever the difficulty of the associated job was each time a share is found.

That's not a bug.  If you really want the number of actual shares found (even though they are not as useful in a world that is moving towards variable difficulty mining), you can get them from the API.

I would like to point out that there is A BUG ASSOCIATED WITH THAT FEATURE.
And that is, that as soon as the "Number of accepted Shares" reaches 10 Million, CGMiner crashes/stops working. I mine on an LTC-Pool with high difficulty and I have to restart CGMiner on a regular basis to make sure the miner just doesn't stop working because it has reached 10 Million diff 1 shares again.

Is there an "official" way to report bugs or is ckolivas going to read this?

[-ck]

Really?
How about this? Are you able to reproduce this?

Running 3.4.3 but notice now (it seems this issue goes back a few versions) that the accepted blocks field now shows the latest difficulty (x number of blocks found it seems) instead of total block count? Anyone got a fix for this?

Not sure what coin you're mining, but on the bitcoin network, cgminer now shows the number of "diff 1 shares" found for the accepted field.  So, if you're pool has a higher share difficulty, the counter will increase by whatever the difficulty of the associated job was each time a share is found.

That's not a bug.  If you really want the number of actual shares found (even though they are not as useful in a world that is moving towards variable difficulty mining), you can get them from the API.

I would like to point out that there is A BUG ASSOCIATED WITH THAT FEATURE.
And that is, that as soon as the "Number of accepted Shares" reaches 10 Million, CGMiner crashes/stops working. I mine on an LTC-Pool with high difficulty and I have to restart CGMiner on a regular basis to make sure the miner just doesn't stop working because it has reached 10 Million diff 1 shares again.

Is there an "official" way to report bugs or is ckolivas going to read this?

I read it. I read all bug reports - that doesn't mean I instantly respond. It was probably coincidence and an unrelated crash since I went to 50 million shares without a problem. Try a debug build.

[vekt0r7]

Really?
How about this? Are you able to reproduce this?

Running 3.4.3 but notice now (it seems this issue goes back a few versions) that the accepted blocks field now shows the latest difficulty (x number of blocks found it seems) instead of total block count? Anyone got a fix for this?

Not sure what coin you're mining, but on the bitcoin network, cgminer now shows the number of "diff 1 shares" found for the accepted field.  So, if you're pool has a higher share difficulty, the counter will increase by whatever the difficulty of the associated job was each time a share is found.

That's not a bug.  If you really want the number of actual shares found (even though they are not as useful in a world that is moving towards variable difficulty mining), you can get them from the API.

I would like to point out that there is A BUG ASSOCIATED WITH THAT FEATURE.
And that is, that as soon as the "Number of accepted Shares" reaches 10 Million, CGMiner crashes/stops working. I mine on an LTC-Pool with high difficulty and I have to restart CGMiner on a regular basis to make sure the miner just doesn't stop working because it has reached 10 Million diff 1 shares again.

Is there an "official" way to report bugs or is ckolivas going to read this?

I read it. I read all bug reports - that doesn't mean I instantly respond. It was probably coincidence and an unrelated crash since I went to 50 million shares without a problem. Try a debug build.

Thank you very much!

[erk]

Unofficial Mac binaries updated to 3.4.3 at http://spaceman.ca/cgminer.

Just tried them on a Mac Mini with Snow Leopard and a single Block Erupter USB, give the USB timeout errors:

Code:

 [2013-09-17 14:00:28] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:30] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:32] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:34] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:36] AMU0: TIMEOUT GetResults took 1999ms but was 100ms
 [2013-09-17 14:00:37] Stratum from pool 0 detected new block
 [2013-09-17 14:00:38] AMU0: TIMEOUT GetResults took 1998ms but was 100ms

Was there ever a version this didn't happen on? If not, we're kinda screwed until a better libusb for osx comes out since this is a libusb bug which is mostly addressed in the linux and windows versions, yet you're compiling against the same libusb version now (which is included statically in the cgminer source tree now).

I have a stack of libusb libraries, no idea where they come from. There are even ones with the binary cgminer from http://spaceman.ca/cgminer though I can't see them linked with otool -L

root % otool -L cgminer
cgminer:
   /System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation (compatibility version 150.0.0, current version 550.44.0)
   /System/Library/Frameworks/IOKit.framework/Versions/A/IOKit (compatibility version 1.0.0, current version 275.0.0)
   @executable_path/dependencies/lib/libcurl.4.dylib (compatibility version 8.0.0, current version 8.0.0)
   /usr/lib/libssl.0.9.8.dylib (compatibility version 0.9.8, current version 0.9.
   /usr/lib/libcrypto.0.9.8.dylib (compatibility version 0.9.8, current version 0.9.
   /System/Library/Frameworks/LDAP.framework/Versions/A/LDAP (compatibility version 1.0.0, current version 2.2.0)
   /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.3)
   /System/Library/Frameworks/OpenCL.framework/Versions/A/OpenCL (compatibility version 1.0.0, current version 1.0.0)
   /usr/lib/libobjc.A.dylib (compatibility version 1.0.0, current version 227.0.0)
   /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 125.2.11)
   /usr/lib/libncurses.5.4.dylib (compatibility version 5.4.0, current version 5.4.0)

Perhaps I need some other command to see which USB library is being used.

Here is a list of the libusb stuff on my Mac

Code:

/usr/lib/libusb-1.0.0.dylib
/usr/lib/libusb-1.0.a
/usr/lib/libusb-1.0.la
/opt/local/bin/libusb-config
/opt/local/i386-mingw32/lib/libusbcamd.a
/opt/local/i386-mingw32/lib/libusbcamd2.a
/opt/local/include/libusb-1.0
/opt/local/include/libusb-1.0/libusb.h
/opt/local/lib/libusb-0.1.4.dylib
/opt/local/lib/libusb-1.0.0.dylib
/opt/local/lib/libusb-1.0.a
/opt/local/lib/libusb-1.0.dylib
/opt/local/lib/libusb-1.0.la
/opt/local/lib/libusb.a
/opt/local/lib/libusb.dylib
/opt/local/lib/libusb.la
/opt/local/lib/pkgconfig/libusb-1.0.pc
/opt/local/lib/pkgconfig/libusb.pc
/opt/local/var/macports/software/libusb
/opt/local/var/macports/software/libusb/libusb-1.0.8_0.darwin_10.x86_64.tbz2

I haven't been able to get the cgminer github source to compile myself, not from lack of trying! So I am dependent on the pre-compiled binaries I can find.

A fresh attempt yeilds:

Code:

root % ./autogen.sh 
readlink: illegal option -- f
usage: readlink [-n] [file ...]
usage: dirname path
Running autoreconf -if...
./autogen.sh: line 12: libtoolize: command not found
configure.ac:24: installing './compile'
configure.ac:17: installing './config.guess'
configure.ac:17: installing './config.sub'
configure.ac:22: installing './install-sh'
configure.ac:53: error: required file './ltmain.sh' not found
configure.ac:22: installing './missing'
Makefile.am:21: warning: 'INCLUDES' is the old name for 'AM_CPPFLAGS' (or '*_CPPFLAGS')
Makefile.am: installing './depcomp'
autoreconf: 'configure.ac' or 'configure.in' is required
Configuring...
./autogen.sh: line 21: /configure: No such file or directory

I know the Mac uses glibtoolize not libtoolize so if I patch autogen.sh to use glibtoolize it then gives:

Code:

root % ./autogen.sh 
readlink: illegal option -- f
usage: readlink [-n] [file ...]
usage: dirname path
Running autoreconf -if...
glibtoolize: putting auxiliary files in `.'.
glibtoolize: copying file `./config.guess'
glibtoolize: copying file `./config.sub'
glibtoolize: copying file `./install-sh'
glibtoolize: copying file `./ltmain.sh'
glibtoolize: putting macros in AC_CONFIG_MACRO_DIR, `m4'.
glibtoolize: copying file `m4/libtool.m4'
glibtoolize: copying file `m4/ltoptions.m4'
glibtoolize: copying file `m4/ltsugar.m4'
glibtoolize: copying file `m4/ltversion.m4'
glibtoolize: copying file `m4/lt~obsolete.m4'
Makefile.am:21: warning: 'INCLUDES' is the old name for 'AM_CPPFLAGS' (or '*_CPPFLAGS')
autoreconf: 'configure.ac' or 'configure.in' is required
Configuring...
./autogen.sh: line 21: /configure: No such file or directory
root % 

A  configure file is actually created and runs to completion with
CFLAGS="-O2" ./configure --enable-icarus


However when I try to make I get:

Code:

root % make
make  all-recursive
Making all in lib
  GEN      arg-nonnull.h
  GEN      c++defs.h
  GEN      warn-on-use.h
  GEN      signal.h
  GEN      stdint.h
  GEN      string.h
make  all-recursive
  CC       dummy.o
  CC       memmem.o
  AR       libgnu.a
/opt/local/bin/ranlib: file: libgnu.a(dummy.o) has no symbols
ranlib: file: libgnu.a(dummy.o) has no symbols
Making all in compat
Making all in jansson
  CC       dump.o
  CC       error.o
  CC       hashtable.o
  CC       load.o
  CC       memory.o
  CC       pack_unpack.o
  CC       strbuffer.o
  CC       strconv.o
  CC       utf.o
  CC       value.o
  AR       libjansson.a
Making all in libusb-1.0
make[3]: *** No rule to make target `all'.  Stop.
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
root % 

And that's where I get stuck. Any suggestions on how to proceed?

I am on MacOSX 10.6.8 (Snow Leopard) with the Mac Ports collection installed.

[-ck]

I am on MacOSX 10.6.8 (Snow Leopard) with the Mac Ports collection installed.

SNIP

Only the release tarball will compile on osx. What I'm saying, though, is if you have a 3.4.3 release, it IS using the latest libusb - statically compiled - regardless of who compiled it for you unless they've gone in and hacked the code to dynamically link the library because there is no way for it to use a dynamic library any more.

[vulgartrendkill]

I am on MacOSX 10.6.8 (Snow Leopard) with the Mac Ports collection installed.

SNIP

Only the release tarball will compile on osx. What I'm saying, though, is if you have a 3.4.3 release, it IS using the latest libusb - statically compiled - regardless of who compiled it for you unless they've gone in and hacked the code to dynamically link the library because there is no way for it to use a dynamic library any more.

I've tried and tried using the latest releases compiling etc. I use an anker usb 10 port hub usb 3.0, so maybe thats whats wrong?

[techman05]

the hub not necessarily cgminer issue . My cg miner compile on Linux what successfully compiled but so far mines buffering assuming my erroupter when in use with the pi needs more power since it works with my pc.

Only place I've heard 3.0 hubs don't work is the raspberry pi for whatever compatibility issue occuring.

[Aurum]

Decided to just use the prebuilt Binaries instead of compiling it for Windows XP 32bit because I ran into too many problems.  I can get CGminer to load but it is only recognizing the card in the first PCI-E slot.  I have a dual card system with an Invidia gt520 in the first PCI-E slot and the XFX 5870 HD.  Cgminer will only mine on the Nvidia card although it gives and error of Multiple OpenCl devices and ADL devices being misdetected/mismatched.  I tried to use the -n function in a bat file but I get a brief flash of readout and then the program shuts down.  I have been able to mine BTC on OpenCL via GIUminer on the ATI card but for some reason it's giving me errors for scrypt mining with CGminer.  Any suggestions?

1.  Is there anyway to generate a Log file so I can read the error output message?

2.  Is there a command to put in the Bat file to ignore the primary card (nvidia) and only work off the secondary card (ati).  Since I can't get -n command to function it I don't even know if it is being detected?  It is being detected by cpuz and the system registry and I have been able to mine BTC with it but not Litecoin in scrypt?

Add a "pause" to your batch file, e.g.:
cgminer.exe -c gristle.conf -n
pause
exit
It will wait for you to hit any key before it exits cmd window.
In your gristle.conf file add the line:
"device" : "1",
and cgminer will only enable device 1.

[kano]

More correctly, read up about gpus (GPU-README) and --gpu-platform

[c1010010]

One of the most frustrating parts about mining is the vast number of versions of libraries, SDKs, etc that can be mixed and matched.

With the latest Linux 3.4.3 is there a suggested version of:
        AMD Catalyst Drivers (AMD_CATALYST_DRIVERS)
        AMD APP SDK
        Open CL
        ADL SDK

to use?

I tried to update to 3.4.3 last night and it nothing seems to compile or work correctly anymore.

Thoughts?  Advice?

[os2sam]

One of the most frustrating parts about mining...

And fun/interesting!

...is the vast number of versions of libraries, SDKs, etc that can be mixed and matched.

[techman05]

Probably won't help but you have to do it real slowly. I at first did this by wireless and had to do it again though luckily linux I could use the up key after restart on wired to do again.

This is what I followed to get all the files for at least eroupters from https://forums.butterflylabs.com/bitcoin-discussion/3238-super-tutorial-raspberry-pi-raspbian-cgminer-tightvnc-wifi-boot.html

Code:

 sudo aptitude install git autoconf libtool libcurl4-openssl-dev libncurses5-dev pkg-config yasm make libusb-1.0-0-dev  Will ask "Do you want to continue?", "Y" to say Yes and wait until finish

I thought on github it had the filenames to apt-get install for the other sdks for nvidia and amd. After that autogen.sh does the long stuff moving things to the right spots.

Good luck for now.

[c1010010]

One of the most frustrating parts about mining...

And fun/interesting!

...is the vast number of versions of libraries, SDKs, etc that can be mixed and matched.

No doubt. 

It is just that what I thought would be a simple upgrade ended bringing the whole rig down... no longer works at all.

AND, on top of it all, all the old sources I used to build the working machine are no longer available online. (Teaches me to rely on other people hosting stuff. :p)

[aa]

Hey, OP. I'm sure this isn't going to be on your priority list, but I'm also sure it'll be a very quick fix.

cgminer will be a big push to standardize hash measurements in writing (since nearly everyone uses it) and the current symbols used are not correct. Just like how we wouldn't write Kwh (kWh) and how we don't write KB/s to represent 1000 B/s (kB/s), the symbols used for hashing should be changed to kH/s and MH/s, GH/s, TH/s, etc.

The little table on the Wikipedia page for kilobyte explains it the same (check out the kilowatt page as well): http://en.wikipedia.org/wiki/Kilobyte

The longer form would likely look like kHash/s or MHash/s.

The PHP script, that I'm pretty sure is still shipped with cgminer, lists it as MHS. That should be updated as well.

Finally, I'm basing this off of using 1000x. If we are actually using 1024 (and I'm pretty sure we aren't), the only change will be an uppercase K.

Edit: The above post still applies, but I'm currently getting input on actually creating a symbol for cryptographic hashing. It seems the entire cryptographic sector doesn't have anything to specifically represent hashes or hashes per second.

[kentt]

I don't have the expertise to evaluate this, but I'm wondering about your thoughts on this:
https://github.com/adamvillena/cgminer/commit/ea532254829d8f57dd561eb223129b4f88a41c61#diff-0ff4390339955f5d81e4a259179a42cd