An (even more) optimized version of cpuminer (pooler's cpuminer, CPU-only)

[DeathAndTaxes]

PS: Seems to me I shouldn't publish my miner that works at 2.5 rate (2500 h/s for 1 GHz of cpu). Just to prevent its usage for a so-called "dark pool"...

Well the larger implication is that once shared and upgraded it no longer provide anyone any advantage. 

So say you made a miner that was 200% hashrate on same hardware.  Double the money right?  Sure until you share it. Eventually everyone will be doubling their hashrate, so global hashrate will double, so difficulty will double and you are right back making the same revenue as when you were using your 100% miner.

Still sharing info is always good for the network (although not for your personal profits).   Hypothetically says an attacker already had a 200% hashrate miner.  Obviously he isn't going to share.   Now with 1000 CPU he has the same hashing power as 2000 avg defender CPU.  In effect the network effective strength is half of what is appears to be.  However if you share your 200% hashrate miner it neutralizes that potential advantage. 

As the miner's get more and more efficient (approaching theoretical max performance of CPU) the confidence grows that nobody has an "unfair" advantage.

[1713976132]

1713976132

[1713976132]

1713976132

[1713976132]

1713976132

[kjlimo]

I still haven't seen much "doubling" in difficulty or hash rate yet...

[iongchun]

Hey, wait, you are on a Mac! I thought you were under Linux, sorry. Unfortunately gcc 4.2 seems to have some problems with macros in assembly code.
In order to get it to compile, SockPuppet had to expand all the macros by hand. Unfortunately I cannot just use his version in the main package, because the code becomes very hard to read. But I think you can contact him if you want the modified sources.

Not only with the stock gcc 4.2, I tried to compile with gcc 4.6 from MacPorts, the result is the same.
Maybe it is a problem of as, the GNU-based assembler of Mac OS X.

[pooler]

Hey, wait, you are on a Mac! I thought you were under Linux, sorry. Unfortunately gcc 4.2 seems to have some problems with macros in assembly code.
In order to get it to compile, SockPuppet had to expand all the macros by hand. Unfortunately I cannot just use his version in the main package, because the code becomes very hard to read. But I think you can contact him if you want the modified sources.

Not only with the stock gcc 4.2, I tried to compile with gcc 4.6 from MacPorts, the result is the same.
Maybe it is a problem of as, the GNU-based assembler of Mac OS X.

You are probably right. What version of binutils ('as' is part of binutils) is giving you problems?

[iongchun]

Not only with the stock gcc 4.2, I tried to compile with gcc 4.6 from MacPorts, the result is the same.
Maybe it is a problem of as, the GNU-based assembler of Mac OS X.

You are probably right. What version of binutils ('as' is part of binutils) is giving you problems?

On Mac OS X 10.6.8:

bash-3.2$ as -v
Apple Inc version cctools-795~45, GNU assembler version 1.38

I don't know if Apple ships the whole GNU binutils, but apparently they customize it for their need.

[pooler]

Not only with the stock gcc 4.2, I tried to compile with gcc 4.6 from MacPorts, the result is the same.
Maybe it is a problem of as, the GNU-based assembler of Mac OS X.

You are probably right. What version of binutils ('as' is part of binutils) is giving you problems?

On Mac OS X 10.6.8:

bash-3.2$ as -v
Apple Inc version cctools-795~45, GNU assembler version 1.38

I don't know if Apple ships the whole GNU binutils, but apparently they customize it for their need.

Hey, that does look old! Current version is 2.22. I had a rapid look at the NEWS file for gas, and it seems that it only supports macros starting from version 2.6.

[ehmdjii]

hello, i am trying to compile this on ubuntu 32bit but get this error:

Code:

checking for working alloca.h... yes
checking for alloca... yes
checking for json_loads in -ljansson... no
checking for pthread_create in -lpthread... yes
./configure: line 4973: syntax error near unexpected token `LIBCURL_CHECK_CONFIG'
./configure: line 4973: `LIBCURL_CHECK_CONFIG(, 7.10.1, ,'

gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3)

[pooler]

hello, i am trying to compile this on ubuntu 32bit but get this error:

Code:

checking for working alloca.h... yes
checking for alloca... yes
checking for json_loads in -ljansson... no
checking for pthread_create in -lpthread... yes
./configure: line 4973: syntax error near unexpected token `LIBCURL_CHECK_CONFIG'
./configure: line 4973: `LIBCURL_CHECK_CONFIG(, 7.10.1, ,'

gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3)

This should solve the problem:
https://bitcointalk.org/index.php?topic=4823.msg588208#msg588208

[ehmdjii]

This should solve the problem:
https://bitcointalk.org/index.php?topic=4823.msg588208#msg588208

thanks, ive seen this thread already but i cant seem to find this file on my ubuntu. libcurl3 and libcurl4 are installed.

[pooler]

This should solve the problem:
https://bitcointalk.org/index.php?topic=4823.msg588208#msg588208

thanks, ive seen this thread already but i cant seem to find this file on my ubuntu. libcurl3 and libcurl4 are installed.

Do you have libcurl4-gnutls-dev (or something similar) installed?

[ehmdjii]

Code:

apt-file search libcurl.m4
libcurl4-gnutls-dev: /usr/share/aclocal/libcurl.m4
libcurl4-nss-dev: /usr/share/aclocal/libcurl.m4
libcurl4-openssl-dev: /usr/share/aclocal/libcurl.m4
pandora-build: /usr/share/aclocal/pandora_have_libcurl.m4

thank you!

[ripper234]

I'm trying to run the optimized miner on Windows 7 64 bit, and it just crashes immediately without any error messages.

I see this in the system (application) event log:

Faulting application name: minerd.exe, version: 0.0.0.0, time stamp: 0x4eef5c94
Faulting module name: pthreadGC2.dll, version: 2.8.0.0, time stamp: 0x4be6d174
Exception code: 0xc0000005
Fault offset: 0x00000000000036a3
Faulting process id: 0x149c
Faulting application start time: 0x01ccc39bb160b781
Faulting application path: C:\Program Files (x86)\Litecoin\new_miner\minerd.exe
Faulting module path: C:\Program Files (x86)\Litecoin\new_miner\pthreadGC2.dll
Report Id: ef0fea24-2f8e-11e1-9e5e-14dae96c0870


So far I have been successfully using the miner-i7.exe release by ArtForz. Any idea why this crashes? I have Soluto installed, so that might be the reason why I'm not seeing any error message ... although I did try to exit it and that didn't help.

[pooler]

I'm trying to run the optimized miner on Windows 7 64 bit, and it just crashes immediately without any error messages.

I see this in the system (application) event log:

Faulting application name: minerd.exe, version: 0.0.0.0, time stamp: 0x4eef5c94
Faulting module name: pthreadGC2.dll, version: 2.8.0.0, time stamp: 0x4be6d174
Exception code: 0xc0000005
Fault offset: 0x00000000000036a3
Faulting process id: 0x149c
Faulting application start time: 0x01ccc39bb160b781
Faulting application path: C:\Program Files (x86)\Litecoin\new_miner\minerd.exe
Faulting module path: C:\Program Files (x86)\Litecoin\new_miner\pthreadGC2.dll
Report Id: ef0fea24-2f8e-11e1-9e5e-14dae96c0870

So far I have been successfully using the miner-i7.exe release by ArtForz. Any idea why this crashes? I have Soluto installed, so that might be the reason why I'm not seeing any error message ... although I did try to exit it and that didn't help.

Two questions:
1. Are you using the 12/19 or the 12/21 version?
2. Did you specify any parameters when invoking the exeutable?

[pooler]

The latest release (12/24) of the miner brings a little speed increase (about 1%) for all architectures, and an additional boost for 64-bit Intel CPUs (you must be using a 64-bit OS).

SAC was able to produce a binary for Mac OS X 10.6/7 (link now available in the first post), but we still don't have a binary for Windows (diki is on an extended vacation). If anybody has some experience with mingw-w64 and wants to give this a try, please let me know.

[mrx]

Code:

[2011-12-26 20:35:56] thread 0: 28662 hashes, 5.71144 khash/s
[2011-12-26 20:35:56] thread 5: 28420 hashes, 5.71091 khash/s
[2011-12-26 20:35:57] thread 1: 28132 hashes, 5.72506 khash/s
[2011-12-26 20:35:57] thread 6: 28644 hashes, 5.63936 khash/s
[2011-12-26 20:35:58] thread 2: 28580 hashes, 5.71380 khash/s
[2011-12-26 20:35:58] thread 7: 27310 hashes, 5.67194 khash/s
[2011-12-26 20:35:59] thread 3: 28146 hashes, 5.72380 khash/s
[2011-12-26 20:36:00] thread 4: 28300 hashes, 5.64322 khash/s

(@ ubuntu server 64-bit with xeon)
that's amazing.

[ripper234]

I'm trying to run the optimized miner on Windows 7 64 bit, and it just crashes immediately without any error messages.

Two questions:
1. Are you using the 12/19 or the 12/21 version?
2. Did you specify any parameters when invoking the exeutable?

1. I tried the 21th version.
2. It crashed without parameters. I tried adding "/?" or "--help", but that still crashed.
Should I specify parameters?

[exahash]

Thanks again Pooler!  This new version brings my old Pentium D 920's up to 2.22 kh/thread (50% improvement) and my newer xeon's all produce about 1/3 more per thread.

[rTech]

Binaries for Windows, could someone make new builds as it seems they are outdated

[mrx]

Binaries for Windows, could someone make new builds as it seems they are outdated

new to mingw32, tried to make a windows version but couldn't make libcurl working now

[kjlimo]

Binaries for Windows, could someone make new builds as it seems they are outdated

heh, this stuff is tough to keep up with. 

But it's still amazing that the hashing power for the cpu chains isn't increasing much.

I have a minerd.exe as of Dec 19, 2011 for windows.  Does that mean this one has been updated again?