I am trying to build all CUDA samples by running make in the sample's base folder. One of the samples require mpi.h, but the system did not have it, which causes an error:
make[1]: Entering directory '$HOME/cuda_samples/samples/0_Simple/simpleMPI'
/bin/mpicxx -I../../common/inc -o simpleMPI_mpi.o -c simpleMPI.cpp
simpleMPI.cpp:25:10: fatal error: mpi.h: No such file or directory
25 | #include <mpi.h>
| ^~~~~~~
compilation terminated.
make[1]: *** [Makefile:371: simpleMPI_mpi.o] Error 1
Since I don't have root privilege, I downloaded a deb file for libopenmpi-dev package (using apt-get download command) and extracted it to somewhere in my user space (using dpkg -x command). However, as we can see, mpicxx tries to find mpi.h in ../../common/inc, which is not where I installed libopenmpi-dev in my user space (I did not notice that untill I installed the package. My bad). So I need to somehow tell mpicxx to find mpi.h in another directory. I know there is a -I option to tell make where additional include directories are, but this option does not apply to mpicxx. How to pass directory information from make's command line to mpicxx is beyond my knowledge. Can you please teach me what option I should use in make's command line to specify include directory used by mpicxx? Of course I can manually copy the installed libopenmpi-dev package to ../../common/inc to accommodate original settings in CUDA sample, but I would like to do something cool and learn something new, so I ask here. Thank you in advance for teaching me.
Environment:
Remote Linux with core version 5.8.0. I am not a super user.
CUDA version: 11.2
CPU: Intel Core i9-10900K
gcc version: (Ubuntu 10.2.0-13ubuntu1) 10.2.0
make version: GNU Make 4.3, Built for x86_64-pc-linux-gnu
MPI version: 4.0.3
The include directory in Makefile is held in a variable INCLUDES, together with -I. So, if we can somehow transfer the include directory of mpi installed in my user space to this variable, we are done. So, the question is reduced to how to transfer a user-defined value from make's command line into Makefile's variable and override it if it has be defined, as is clearly asked in the question.
Fortunately make provides this command line option: VAR=value, so the option to answer my question is
make INCLUDES=-I/path/to/mpi/include/in/my/user/space
Related
I am trying to simulate a simple Hello world ARM example on my desktop computer. I tried both qemu and gem5. Both gives a similar error. They cannot find ld-linux-aarch64.so.1. Actually I cannot find it either. If I could find it, I will show it with -L (in qemu) or --redirects (in gem5).
The file is:
armhello: ELF 64-bit LSB shared object, ARM aarch64, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-aarch64.so.1, BuildID[sha1]=23a21b7a545ac510923b6b3713d2bbee092f820a, for GNU/Linux 3.7.0, not stripped
It is compiled with: aarch64-linux-gnu-gcc
I am trying to run it in qemu with:
qemu-aarch64 armhello
I got this error:
/lib/ld-linux-aarch64.so.1: No such file or directory
I try to run it in gem5 with: (simpleARM.py points to my executable (named as armhello))
build/ARM/gem5.opt configs/tutorial/simpleARM.py
I got this error:
panic: panic condition fd < 0 occurred: Failed to open file /lib/ld-linux-aarch64.so.1.
How can I solve this?
Note: I know it works when compiled --static. But I need to run more complex binaries that are dynamically linked and I cannot change those. This is just an example.
For gem5 you can use --redirects and --interp-dir: How to run a dynamically linked executable syscall emulation mode se.py in gem5?
For qemu you need -L: Using dynamic linker with qemu-arm
same problem on x86_64 machine docker build with an arm64 docker image:
FROM multiarch/qemu-user-static:x86_64-aarch64 as qemu
FROM alpine
COPY --from=qemu /usr/bin/qemu-aarch64-static /usr/bin/
# add this line to resolve
RUN apk add libc6-compat
I just came across a weird situation after installing cuda... I literally followed every single step suggested by the nvidia website: http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-linux/#axzz3H0tm46yY.... everything seems to be fine, even the samples work... however when I try to run the "hello world" program presented in Cuda by Example documentation, the terminal displays:
The program 'nvcc' is currently not installed. You can install it by typing:
sudo apt-get install nvidia-cuda-toolkit
This is quite strange since nvcc should already be installed during the procedure suggested by the nvidia website....
Does anyone have any idea?
The PATH variable needs to include your cuda /bin directory (by default it is /usr/local/cuda-6.5/bin)
On Ubuntu you can edit the .bashrc file on your $HOME directory, adding the following;
export PATH=/usr/local/cuda-6.5/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-6.5/lib64:$LD_LIBRARY_PATH
If you are running a 32-bit Ubuntu version, then the paths are:
export PATH=/usr/local/cuda-6.5/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-6.5/lib:$LD_LIBRARY_PATH
After editing the bashrc file, source it (or simply restart the terminal):
source ~/.bashrc
Of course, if you have changed the default location of your CUDA directory you must change the PATHS accordingly.
PLUS: there are another options to tell the system where to find a library, a program, etc. like using enviromental modules, specify the complete path to the libraries and programs manually are another options. There are many methods out there!
I installed the driver and toolkit for CUDA 5 in 64-bit RHEL 6.3 successfully.
However, when I tried compiling the CUDA 5 examples, I got the error message:
make[1]: Leaving directory `/root/NVIDIA_CUDA-5.0_Samples/0_Simple/cppIntegration'
which: no mpicxx
How can I fix this for the CUDA 5 examples to compile?
In order to build the simpleMPI example, you need some kind of MPI installed on your system. You can get around this and build most of the samples by doing:
make -k
this will attempt to go past errors in the make process and build all targets that can be built.
If you prefer, you can delete this directory:
/root/NVIDIA_CUDA-5.0_Samples/0_Simple/simpleMPI
perhaps with the following command, as root:
rm -Rf /root/NVIDIA_CUDA-5.0_Samples/0_Simple/simpleMPI
and relaunch your make. Personally I think the make -k option is simpler.
(the message about cppIntegration is just the last target that got successfully built)
Developing locally in MAMP and need Sphinx to work with MAMP's MySQL. Basically hitting a dead end during $ sudo make.
MAMP installs without some necessary resources for Sphinx integration, mainly a mysql lib directory and an include directory filled with C header source files. These were downloaded and installed successfully (using CMake) into the following directories:
/Applications/MAMP/Library/include/mysql
/Applications/MAMP/Library/lib/mysql
After unzipping Sphinx and running:
sudo ./configure --prefix=/usr/local/sphinx --with-libstemmer --with-mysql=/Applications/MAMP/Library
I got:
******************************************************************************
ERROR: cannot find MySQL include files.
Check that you do have MySQL include files installed.
The package name is typically 'mysql-devel'.
If include files are installed on your system, but you are still getting
this message, you should do one of the following:
1) either specify includes location explicitly, using --with-mysql-includes;
2) or specify MySQL installation root location explicitly, using --with-mysql;
3) or make sure that the path to 'mysql_config' program is listed in
Changing the ./configure command to:
sudo ./configure --prefix=/usr/local/sphinx--with-libstemmer --with-mysql-includes /Applications/MAMP/Library/include --with-mysql-libs /Applications/MAMP/Library/lib
throws the following on the outset, but regardless results in a successful configuration:
configure: WARNING: you should use --build, --host, --target
configure: WARNING: invalid host type: /Applications/MAMP/Library/include
configure: WARNING: you should use --build, --host, --target
configure: WARNING: invalid host type: /Applications/MAMP/Library/lib
The log also has the following line, (which I think is relevant in a sec):
checking MySQL include files... -Iyes
Now, continuing, $ sudo make throws the following:
Making all in src
/bin/sh svnxrev.sh ..
make all-am
g++ -DHAVE_CONFIG_H -I. -I../config -DSYSCONFDIR="\"/usr/local/sphinx--with-libstemmer/etc\"" -DDATADIR="\"/usr/local/sphinx--with-libstemmer/var/data\"" -I/usr/local/include -Iyes -Wall -g -D_FILE_OFFSET_BITS=64 -O3 -DNDEBUG -MT sphinx.o -MD -MP -MF .deps/sphinx.Tpo -c -o sphinx.o sphinx.cpp
In file included from sphinx.cpp:16:
sphinx.h:64:19: error: mysql.h: No such file or directory
... //whole bunch of errors follow, resulting from the above
So, I know mysql.h exists in my includes file, I think I have all necessary binaries as well, and the configuration seems to see the include files so I'm tearing out my hair a bit. I'm hoping this is a simple path problem, or a syntax error on my part with the ./configure attribute since this is my first time compiling and installing from the command line.
Thanks.
Sphinx's ./configure uses an unknown default value for --with-mysql, which is enabled by default, which could make you think calling it is not necessary.
To get Sphinx to work with MAMP's varyingly distributed MySQL files, make sure to use each of the following attributes in your ./configure command, specifying direct paths for each:
--with-mysql= // root mysql app
--with-includes= // path to mysql header includes
--with-mysql-libs= // path to libmysqlclient.dylib files
The final command that worked for me:
./configure --prefix=/usr/local/sphinx --with-libstemmer --with-mysql=/Applications/MAMP/Library --with-mysql-includes=/Applications/MAMP/Library/include/mysql --with-mysql-libs=/Applications/MAMP/Library/lib/mysql
--prefix is your preferred install directory in usr/local and --libstemmer adds Snowball's extended stemming functionality (if you downloaded it).
If this runs without error, just make and make install and you're good to go.
Also note that there's a configure error log in the uncompressed Sphinx directory that's easy to overlook. Didn't help in this scenario, but could be really useful to others if you're having problems.
I have compiled a program in fedora using the mysql dev library (include mysql.h in header file). I need to compile in on FreeBSD. I do not want to download from source and compile but rather would like to download from ports or something equivalent to facilitate removing it if need be.
Does anyone know the equivalent of the libmysql-dev in FreeBSD. I have not found it in ports?
You do not need a special package for this. The standard mysql client package/port already includes the libraries and the header files you will need. The header files will end up, by
default, in /usr/local/include/mysql/ directory, while
the libraries will go to /usr/local/lib/mysql/.
So - just install a client from ports or packages, and you are set.
mysql.h is included in the mysql-client-xx port.
assuming you have a configure script or makefile you should set the LDFLAGS and CPPFLAGS environment variables to:
export LDFLAGS='-l /usr/local/lib'
export CPPFLAGS='-I /usr/local/include'
(or setenv, if using csh)
then ./configure and make as normal.
if you are compiling from the command line "gcc -o myprog mysource.c" just add the -I and -l options to your command and it should compile fine.
try: find /usr/local -iname 'mysql' to see files (headers, shared objects and binaries) you in fact have installed on your system.