I have installed opencv and Cuda.I have made some opencv code and now I wanna try to make it run faster using CUDA.The problem is that the opencv GPU module does not satisfy me. How can I make a CUDA code to include and use opencv libraries and functions exactly like my .cpp files?I include the needed .h files (cv.h highgui.h) that I moved to the "include" CUDA folder in my CUDA code but when I try to compile it using nvcc it says
Undefined symbols for architecture i386:
"_cvLoadImage", referenced from:
_main in tmpxft_000177b6_00000000-14_cuda.o
ld: symbol(s) not found for architecture i386
collect2: ld returned 1 exit status
which means that it does not understand the functions so the libraries are not included the right way as far as I can see.How can I make it work?
As far as I know, you can't just include OpenCV functions direclty in CUDA code. CUDA has its own way of programming and code made for CPU can't run on GPU directly.
I would suggest you to look at the documentation to understand how works CUDA.
If the gpu build of OpenCV does not satisfy you maybe you should make your own implementation of those functions.
OR
you just want to use openCV on CPU and also use CUDA for some other computation ?
In the case, I think you have to create a cpp file which wraps your CUDA functions.
You will be able to use your wrapped functions as usual in your C or C++ code.
I think, the "Getting Started with GPU-accelerated Computer Vision using OpenCV and CUDA" webinar materials should help. You can find them here: link
You better not try to mix between opencv headers and nvcc compiler, this is a bad practice.
Opencv is built in library which was build using CMAKE (usually in windows under visual studio)
When you include an opencv header in nvcc compiler, you have to include all the "presents" you get from opencv, e.g: external dll, headers and extern function expected in opencv code.
What I find is the most appropriate solution in your case, would be to write a code which compiles in VS, refer to you CUDA code and Opencv's CUDA back and forth (you can always use primitives and extern functions).
Related
I need to use the GSL library in my program on LPCXpresso 4367(ARM CORTEX M4). I tried to follow the library linking procedure for LPC xpresso but the MCU linker is giving me these errors:
MCUXpressoIDE_10.3.0_2200\workspace\test1\Debug/../src/test1.c:53: undefined reference to 'gsl_linalg_LU_decomp'
MCUXpressoIDE_10.3.0_2200\workspace\test1\Debug/../src/test1.c:56: undefined reference to 'gsl_matrix_alloc'
MCUXpressoIDE_10.3.0_2200\workspace\test1\Debug/../src/test1.c:57: undefined reference to 'gsl_linalg_LU_invert'
and so on for other functions as well.
I have the libgsl.a and libgslcblas.a precompiled libraries for windows which works perfectly on codeblocks on windows with GCC compiler.
I read that I need to crosscompile library for the arm-none-eabi-gcc toolchain. But can someone please provide me the procedure as well?
the libgsl.a and libgslcblas.a precompiled libraries for windows
Those won't do for ARM.
In order to work on another platform, these libs need to be compiled from source code with the proper compiler (and settings - Cortex-M4F requires Thumb2 instruction set).
As the libraries are precompiled for Windows they don't work for ARM (as it is said in the other answer)
You need to cross compile the libraries first. If you install the GSL libraries following this procedure, you only need to change the parameters in the ./config according to your platform, for example I used:
./config --host=arm-linux-gnueabihf --prefix=/home/yourname/gsl_arm
Inside the .zip file with the gsl-2.5 files, there is a file called INSTALL. There you can find more details on the options for cross compiling.
Make sure to make clean before if you have already compiled the library for different settings. After cross-compiling the library when you run make check on the terminal you will probably get errors, but still it works. Continue with make install and you are ready to use it.
I have a situation where I recently added a bunch of pcap functionality to a shared library that I've written to do some packet sniffing/injecting stuff.
I installed the pcap stuff, i.e.
sudo apt-get install libpcap-dev
wrote all the code, then tried to build it with my makefile
All compiled and linked no problem
But when I looked more closely I noticed I hadn't specified -lpcap as a library dependency
But the linker hasn't complained about any undefined references..
So the question is how is the linker finding those pcap functions that I've called in my library code?
I was aware that if you use socket.h functions they're contained within the standard C runtime lib so you don't need to explicitly specify a library dependency. Is it the same for pcap?
Is there a way of querying where the linker has found its functions, i.e. where it found the pcap functions in this instance? I'd like to understand whats going on, rather than just being glad it worked..
To clarify, can you find the specific library file (.so or .a) that the linker has used when linking to a function that my code references?
possibly a verbose option to see exactly what the linker is doing?
I've been trying to run the code from https://github.com/dungtn/mpi-floyd/blob/master/floyd2d.c in my system. I'm using CodeBlocks IDE and MS-Mpi. When I try to compile the code, it says undefined reference to MPI_file_seek#12. Does this mean MS mpi does not support this function or why does this happen?
This usually happens if you are trying to link 32-bit code with 64-bit libraries. The fact that the unresolved symbol has #12 in its name means that the compiler is expecting that MPI_File_seek is an stdcall function. stdcall is mainly used for DLL functions and only on x86 (x64 uses a different calling convention similar to fastcall). If you are linking against the 64-bit import library of MS-MPI, the decorated symbol won't be found in the library and such an error will occur.
Double check what version of MS-MPI you have and also your project settings and make sure that both have the same "bitness".
Change the project settings in Code::Blocks to a C project (rather than C++ project, what you have currently). It may be easier to create a brand new C project and import the file there. Double check that Code::Blocks in running gcc and not g++ to compile your code (floyd2d.c).
If it still doesn't work, please post the full compiler and linker output of Code::Blocks, including the commands run and their output messages.
Let me start out by saying that I'm not a C developer and I know very little about actually writing real world C code. I've been doing some research to find a xUnit framework that I can use to write tests for C code and based on what I've found it seems like Unity is the one that I want to go with. It seems simple enough, but I really just don't know what to do after I download the zip file from Unity's website. It doesn't seem to have the normal configure/make/make install, and if it did, I'm not sure that is what I should be using anyway. It does, however, ship with some rake tasks, but none of those seemed to be any kind of "install" task. As a last resort I tried to just copy the 3 source files in with my code (which I really hope is not the right thing to do), but when I try that I get an error trying to compile my c file with gcc, but I think this should be working. Here is my set up:
src/
mycode.c
unity.c
unity.h
unity_internals.h
Here is the source for mycode.c
/* mycode.c */
#include "unity.h"
void test_sample(void)
{
TEST_ASSERT_EQUAL_INT(0, 0);
}
When I run gcc mycode.c I get:
Undefined symbols for architecture x86_64:
"_main", referenced from:
start in crt1.10.6.o
"_UnityAssertEqualNumber", referenced from:
_test_sample in ccyHByv6.o
ld: symbol(s) not found for architecture x86_64
collect2: ld returned 1 exit status
(I get a similar error when I try to compile unity.c with gcc). Which I assume means that the code that ships with unity requires a different compiler than what I have which is:
i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.9.00)
or that maybe unity is not compatible with a 64 bit processor... (I'm running on Mac OS 10.7.3 with a 2.4 GHz Intel Core 2 Duo processor - another thing that may or may not be relavent is that I've got XCode Version 4.3 (4E109) and also Command Line Tools for XCode) At this point I'm just grasping at straws and I'm in way over my head.
My question is, what is the correct process to go through to take a 3rd party C library, such as Unity, and make it available to my C code? Do I need to install something like in Python or Ruby or add something to my path like in Java or something else? Shouldn't just dropping unity's code in with mine work? Am I doing something wrong or is Unity or both? I really just want to be able to test drive C code using Unity. Any help would be greatly appreciated. Thanks in advance!
First, try 'gcc *.c -o mytest'. This will compile all of the C source files into object files, and then link them together into the binary 'mytest'. Keep in mind that all C source files have to be compiled to object files before they can be linked together. (A library is just a bunch of packaged object files.)
If you had a unity library installed in /usr/lib, you could do something like 'gcc mycode.c -lunity -o mytest'. If you had a unity library sitting in the current directory, you might do 'gcc mycode.c ./unity.a -o mytest'. This tells the compiler to look for a file named 'unity.a' in the current directory. Some libraries build .so files ('shared object' files, similar to DLLs in Windows). Replacing 'unity.a' with 'unity.so' should work if that is the case. (I'm assuming a Unix/Linux environment here.)
As an alternative to Unity, look at Google Test, which can be used with C code. I know it is supported on the Mac as well. The primary benefit is a large and active community. More information on Google Test from another SO question: Is Google Test OK for testing C code?
I figured out my problem. It turns out that unity requires you to define a setup and a teardown function and if you do not, you will get errors similar to the one that I was running into.
I want to write a simple SDL OpenGL app, Codeblocks is the IDE I use.
When I create a new OpenGL project, it compiles fine, but if I try to use a function from the SDL header, le wild "undefined reference error" occurs. The same goes for the other direction, if I create a new SDL project, I can use the SDL functions without problems but I get a "undefined reference error" for the OpenGL functions...
NOTES:
I Use Ubuntu 11.10
I have installed the SDL and the OpenGL packages
You need to add the correct library. Headers just give the compiler sort of a index. But you need to tell the linker which libraries to actually pull in. You should find the linker options at the build settings. You need the following libraries for SDL + OpenGL
libGL.so ( -lGL linker switch )
libSDL.so ( -lSDL linker switch)
You may also require libGLU.so if you're using glu… functions.
Asking pkg-config is the preferable thing for obtain the particular flags and options needed for compilation and linkage against SDL and Mesa's GL+GLU. (Some GL implemenetations may not be shipping .pc files, but they should still be used where available.)