cpuminer-opt v3.1.16 adds m7m algo.
https://drive.google.com/file/d/0B0lVSGQYLJIZSHJ3dGMxZUJfeTQ/view?usp=sharing
Topic: [ANN]: cpuminer-opt v3.8.8.1, open source optimized multi-algo CPU miner - page 174. (Read 444040 times)
April 20, 2016, 04:27:53 PM
April 20, 2016, 03:55:38 PM
Can you run this command on your AMD processors and show me the output?
Code:
gcc -march=native -Q --help=target | fgrep march
Here you are:
Code:
root@beast:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
This curious. I presume that shows which arch is used by native.
On my skylake I get core2-avx and on my
configure fails with -march=skylake on my skylake.
Intel Core i7-4790K CPU @ 4.40GHz
Code:
root@storm:~$ gcc -march=native -Q --help=target | fgrep march
-march= core-avx2
AMD Sempron 145
Code:
root@wolverine:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
Correction, my sandy bridge shows corei7-avx.
April 20, 2016, 03:38:19 PM
Can you run this command on your AMD processors and show me the output?
Code:
gcc -march=native -Q --help=target | fgrep march
Here you are:
Code:
root@beast:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
This curious. I presume that shows which arch is used by native.
On my skylake I get core2-avx and on my haswell I get corei7-avx.
configure fails with -march=skylake on my skylake.
Intel Core i7-4790K CPU @ 4.40GHz
Code:
root@storm:~$ gcc -march=native -Q --help=target | fgrep march
-march= core-avx2
AMD Sempron 145
Code:
root@wolverine:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
April 20, 2016, 03:30:30 PM
April 20, 2016, 03:30:05 PM
Can you run this command on your AMD processors and show me the output?
Code:
gcc -march=native -Q --help=target | fgrep march
Here you are:
Code:
root@beast:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
This curious. I presume that shows which arch is used by native.
On my skylake I get core2-avx and on my haswell I get corei7-avx.
configure fails with -march=skylake on my skylake.
April 20, 2016, 03:22:27 PM
windows support?
April 20, 2016, 03:17:53 PM
Can you run this command on your AMD processors and show me the output?
Code:
gcc -march=native -Q --help=target | fgrep march
Here you are:
Code:
root@beast:~$ gcc -march=native -Q --help=target | fgrep march
-march= amdfam10
April 20, 2016, 02:08:20 PM
Code:
gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.8/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.8.4-2ubuntu1~14.04.1' --with-bugurl=file:///usr/share/doc/gcc-4.8/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.8 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.8 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-libmudflap --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.8-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --disable-werror --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)
Can you run this command on your AMD processors and show me the output?
Code:
gcc -march=native -Q --help=target | fgrep march
Below are examples of how output should look like.
Output on my core2:
Code:
hmage@wraith:~$ gcc -march=native -Q --help=target | fgrep march
-march= core2
On haswell:
Code:
hmage@dhmd:~$ gcc -march=native -Q --help=target | fgrep march
-march= haswell
On ivybridge:
Code:
hmage@education:~$ gcc -march=native -Q --help=target | fgrep march
-march= ivybridge
List of supported architectures on GCC 4.8.4 is here -- https://gcc.gnu.org/onlinedocs/gcc-4.8.4/gcc/i386-and-x86-64-Options.html
April 20, 2016, 04:26:09 AM
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
Build successful on both AMD CPU's without -DNO_AES_NI
So, what was the setup that was giving the errors? If it works with and without NO_AES_NI?
What was the version of gcc?
3.19.0-58-generic kernel
AMD Phenom II X4 940 @ 3.6GHz with 4GB RAM
Code:
gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.8/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.8.4-2ubuntu1~14.04.1' --with-bugurl=file:///usr/share/doc/gcc-4.8/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.8 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.8 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-libmudflap --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.8-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.8-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --disable-werror --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)
April 19, 2016, 10:09:05 PM
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
Build successful on both AMD CPU's without -DNO_AES_NI
So, what was the setup that was giving the errors? If it works with and without NO_AES_NI?
What was the version of gcc?
I don't understand your question.
If you mean my request above to test it was just to confirm it works on real non-aes HW, which I don't have.
If you are referring to the compile problems in 3.1.14 it was because of a misunderstanding on the maning of __AES__.
If you're referring to you problems with -march=native I have no idea.
gcc v4.8.4
Did any of these answer your question?
April 19, 2016, 04:33:21 PM
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
Build successful on both AMD CPU's without -DNO_AES_NI
So, what was the setup that was giving the errors? If it works with and without NO_AES_NI?
What was the version of gcc?
April 18, 2016, 05:46:06 PM
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
Build successful on both AMD CPU's without -DNO_AES_NI
Excelllent, thanks.
April 18, 2016, 05:45:02 PM
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
Build successful on both AMD CPU's without -DNO_AES_NI
April 18, 2016, 05:38:19 PM
Also, just to confirm, did you successfully compile without the -DNO_AES_NI flag?
I compiled successfully few times on each cpu with this configure
Code:
./configure CFLAGS="-DNO_AES_NI -O3 -march=btver1" --with-curl --with-crypto
Please try with
Code:
./configure CFLAGS="-O3 -march=btver1" --with-curl --with-crypto
April 18, 2016, 05:31:48 PM
Also, just to confirm, did you successfully compile without the -DNO_AES_NI flag?
I compiled successfully few times on each cpu with this configure
Code:
./configure CFLAGS="-DNO_AES_NI -O3 -march=btver1" --with-curl --with-crypto
The build.sh error at the end is this:
Code:
In file included from algo/echo/aes_ni/hash.c:19:0:
algo/echo/aes_ni/hash.c: At top level:
./miner.h:479:20: warning: ‘algo_names’ defined but not used [-Wunused-variable]
static const char *algo_names[] = {
^
In file included from algo/echo/aes_ni/hash.c:20:0:
algo/echo/aes_ni/hash_api.h:47:23: warning: ‘initial_echo512_ctx’ defined but not used [-Wunused-variable]
static hashState_echo initial_echo512_ctx =
^
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Error 1
make[2]: *** Waiting for unfinished jobs....
mv -f algo/groestl/.deps/cpuminer-groestl.Tpo algo/groestl/.deps/cpuminer-groestl.Po
In file included from algo/groestl/myr-groestl.c:1:0:
./miner.h:479:20: warning: ‘algo_names’ defined but not used [-Wunused-variable]
static const char *algo_names[] = {
^
mv -f algo/.deps/cpuminer-fresh.Tpo algo/.deps/cpuminer-fresh.Po
mv -f algo/groestl/.deps/cpuminer-myr-groestl.Tpo algo/groestl/.deps/cpuminer-myr-groestl.Po
make[2]: Leaving directory `/home/*****/cpuminer-opt-3.1.15-phenom'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/*****/cpuminer-opt-3.1.15-phenom'
make: *** [all] Error 2
strip: 'cpuminer': No such file
And this one
Code:
algo/echo/aes_ni/hash.c:385:4: note: in expansion of macro ‘TRANSFORM’
TRANSFORM(_state[i][j], _k_opt, t1, t2);
^
In file included from algo/echo/aes_ni/hash.c:19:0:
algo/echo/aes_ni/hash.c: At top level:
./miner.h:479:20: warning: ‘algo_names’ defined but not used [-Wunused-variable]
static const char *algo_names[] = {
^
In file included from algo/echo/aes_ni/hash.c:20:0:
algo/echo/aes_ni/hash_api.h:47:23: warning: ‘initial_echo512_ctx’ defined but not used [-Wunused-variable]
static hashState_echo initial_echo512_ctx =
^
mv -f algo/argon2/ar2/.deps/cpuminer-ar2-scrypt-jane.Tpo algo/argon2/ar2/.deps/cpuminer-ar2-scrypt-jane.Po
gcc -std=gnu99 -DHAVE_CONFIG_H -I. -Iyes/include -Iyes/include -fno-strict-aliasing -I. -Iyes/include -Iyes/include -Wno-pointer-sign -Wno-pointer-to-int-cast -O3 -march=native -Wall -Iyes/include -Iyes/include -MT algo/groestl/sse2/cpuminer-grso-asm.o -MD -MP -MF algo/groestl/sse2/.deps/cpuminer-grso-asm.Tpo -c -o algo/groestl/sse2/cpuminer-grso-asm.o `test -f 'algo/groestl/sse2/grso-asm.c' || echo './'`algo/groestl/sse2/grso-asm.c
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Error 1
make[2]: *** Waiting for unfinished jobs....
mv -f algo/groestl/.deps/cpuminer-myr-groestl.Tpo algo/groestl/.deps/cpuminer-myr-groestl.Po
mv -f algo/groestl/sse2/.deps/cpuminer-grso.Tpo algo/groestl/sse2/.deps/cpuminer-grso.Po
mv -f algo/groestl/sse2/.deps/cpuminer-grso-asm.Tpo algo/groestl/sse2/.deps/cpuminer-grso-asm.Po
make[2]: Leaving directory `/home/*****/cpuminer-opt-3.1.15-sempron'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/*****/cpuminer-opt-3.1.15-sempron'
make: *** [all] Error 2
strip: 'cpuminer': No such file
April 18, 2016, 05:19:46 PM
Following the topic I get the impression that the new changes detects cpu, OS, algo (for AES-NI support)
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
That is correct. You need all three to get the best performance.
I haven't found specifically whether either of your CPUs has AES_NI but I suspect not. Try compiling with
-march=core2. It seems to have worked for others.
Both AMD's are without AES-NI support.
-march=core2 works on both, but the compiled binary is about 15% slower than with -march=btver1.
AFAIK its related to SSE implementations on AMD CPU's. So with -march=btver1 which is AMD specific I got the best results.
I tried several -march options, starting with core2 and measured the hashrate on the miner and the pool to choose that one for me.
P.S. Logs are on the way
The logs you sent had no errors, but that's ok I found the nugget in the first line of the previous compile session:
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Error 1
It's clear it was trying to compile AES code.
Also, just to confirm, did you successfully compile without the -DNO_AES_NI flag?
I've drafted the following note to be added to the README file:
Also, just to confirm, did you successfully compile without the -DNO_AES_NI flag?
April 18, 2016, 04:55:28 PM
Following the topic I get the impression that the new changes detects cpu, OS, algo (for AES-NI support)
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
That is correct. You need all three to get the best performance.
I haven't found specifically whether either of your CPUs has AES_NI but I suspect not. Try compiling with
-march=core2. It seems to have worked for others.
Both AMD's are without AES-NI support.
-march=core2 works on both, but the compiled binary is about 15% slower than with -march=btver1.
AFAIK its related to SSE implementations on AMD CPU's. So with -march=btver1 which is AMD specific I got the best results.
I tried several -march options, starting with core2 and measured the hashrate on the miner and the pool to choose that one for me.
P.S. Logs are on the way
April 18, 2016, 04:48:09 PM
Following the topic I get the impression that the new changes detects cpu, OS, algo (for AES-NI support)
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
That is correct. You need all three to get the best performance.
I haven't found specifically whether either of your CPUs has AES_NI but I suspect not. Try compiling with
-march=core2. It seems to have worked for others.
April 18, 2016, 04:42:05 PM
Following the topic I get the impression that the new changes detects cpu, OS, algo (for AES-NI support)
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
and if the cpu, OS, algo doesnt support AES-NI it falls back to SSE2.
I will send you PM in a few minutes with the compile logs for both CPU's - Phenom II X4 940 and Sempron 145
April 18, 2016, 04:37:21 PM
New in v3.1.15
- unified build procedure fixed
- build.sh now works for CPUs with and without AES_NI
- it is no longer necessary to add "-DNO_AES_NI" CFLAG to the
configure command when building for CPUs without AES_NI.
- The system will automatically compile for the correct architecture
Thanks!- unified build procedure fixed
- build.sh now works for CPUs with and without AES_NI
- it is no longer necessary to add "-DNO_AES_NI" CFLAG to the
configure command when building for CPUs without AES_NI.
- The system will automatically compile for the correct architecture
Strange thing - when I compile with
Code:
./autogen.sh && ./configure CFLAGS="-DNO_AES_NI -O3 -march=btver1" --with-curl --with-crypto && make
When I use ./build.sh I get an error in compile.
Will try on AMD Phenom II X4 940 and see can I reproduce it.
Edit: ./build.sh fails also on AMD Phenom II X4 940
Code:
make[2]: *** [algo/echo/aes_ni/cpuminer-hash.o] Error 1
make[2]: *** Waiting for unfinished jobs....
mv -f algo/groestl/sse2/.deps/cpuminer-grso-asm.Tpo algo/groestl/sse2/.deps/cpuminer-grso-asm.Po
mv -f algo/argon2/ar2/.deps/cpuminer-opt.Tpo algo/argon2/ar2/.deps/cpuminer-opt.Po
mv -f algo/argon2/ar2/.deps/cpuminer-ar2-scrypt-jane.Tpo algo/argon2/ar2/.deps/cpuminer-ar2-scrypt-jane.Po
make[2]: Leaving directory `/home/urban/cpuminer-opt-3.1.15'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/*****/cpuminer-opt-3.1.15'
make: *** [all] Error 2
The only real difference I see is -march. build.sh always uses native. -DNO_AES_NI is now redundant and should
not be necessary unless -march=native is choosing an incompatible architecture. If that's the case I don't know
what I can do other than document the workaround.
It would be nice to see the compile errors to give me an idea what it was trying to do.
I don't have a non-aesni CPU to test on so it would be nice if you could confirm the capabilities reported when the miner
is started is correct.
Edit: the real errors are further back but didn't cause the compile to fail immediately. Do they compile if you specify
a specific arch but without -DNO_AES_NI?
Jump to: