I'm looking to build a host program calling OpenCL code running on my GPU device. The cl source has the following form:
#include "skip_mwc.cl"
typedef struct{ uint x; uint c; } mwc64x_state_t;
//blah...
If I get rid of the #include directive and copy/paste the content of "skip_mwc.cl" directly into this source, I can partially "build" and at least get some errors, showing that my compiler (clang9 cl compiler) can at least recognize the kernels code. With the #include approach I get the following error:
Build log::
<program source>:9:10: fatal error: 'skip_mwc.cl' file not found
#include "skip_mwc.cl"
I have checked and the file is there in the search paths, so I'm inclined to believe that my Xcode IDE doesn't index .cl files properly to perform automatic file inclusion (as in .c or .cpp).
I really want to avoid having to copy/paste source from one file into the other. Any suggestions from someone familiar with Xcode, who has encountered this problem and managed to solve it, are very welcome and needed.
Thanks,
A
Two possible solutions:
Set the -I include_dir compiler option in clBuildProgram(), see also this answer.
Read both files from C++ with fstream and string-concatenate their content.
Also see the option of embedding the OpenCL code into the executable via stringification macro.
Related
I've tried to get an audio library statically linked to my program. I use this nimble package. To get it run, i had to build the soloud library as described here. For short after download i ran "genie --with-miniaudio-only --platform=x64 vs2017" in the "build" folder and got the source code to generate the dynamic and the static library. For now i can run the following demo program from the nimble package with the generated dll alongside:
import solouddotnim, times, os
var i, spin = 0
var sl : ptr Soloud
sl = Soloud_create()
discard Soloud_init(sl)
Soloud_setGlobalVolume(sl, 1)
var stream = WavStream_create()
discard WavStream_load(cast[ptr Wav](stream), "test.ogg")
let currentTime = epochTime()
let length = WavStream_getLength(stream)
discard Soloud_play(cast[ptr Soloud](sl), cast[ptr Wav](stream))
while epochTime() - currentTime <= length:
sleep(100)
Soloud_deinit(sl)
Soloud_destroy(sl)
Now to the static-link part. In the solouddotnim.nim file of the nimble package i use, i see this part:
when defined(windows):
const
libname* = "libsoloud.dll"
elif ...
So i simple changed the windows part to the following, re-installed the nimble-package and placed the "soloud_static_x64.lib" alongside to the "main.nim" of the testproject:
when defined(windows):
const
libname* = "soloud_static_x64.lib"
elif ...
But this doesent make it. (cant open "soloud_static_x64.lib" error when build)
Evereywhere where the constant "libname" is used there are the pragmas "cdecl", "importc" and "dynlib". For example:
proc Soloud_create*(): ptr Soloud {.cdecl, importc: "Soloud_create", dynlib: libname.}
So "dynlib" is telling nim to use a dll on windows. But was is the pragma for static libraries?
In the nim documentations i only found DynlibOverride to link to static libraries, but i dont understand the example and here is where i stuck. I've tried the followings:
nim c --dynlibOverride:libname --passL:soloud_static_x64.lib "examples\00-ogg\Example00_ogg.nim"
nim c --dynlibOverride:soloudtotnim --passL:soloud_static_x64.lib "examples\00-ogg\Example00_ogg.nim"
Firstly i dont know what parameter dynlibOverride expects and secondly both compiles, but dont work. It expects a dynamic library alongside the exe.
My last try was to remove all dynlib pragmas from the nimble package. But now i cant compile it.
undefined reference to `Soloud_create'
...
Error: execution of an external program failed: 'gcc.exe...
My knowlege ends here. Can someone help me?
Thanks in advance.
Edit:
I could not get any of your solutions work. I break down the problem as small as possible so everybody can reproduce this:
"foo.nim" contains this:
proc add*(a, b: int): int {.cdecl, exportc.} =
a + b
proc sub*(a, b: int): int {.cdecl, exportc.} =
a - b
The .lib is simply generated with this command: "nim c --app:staticlib foo.nim"
Now to use it i created a file "main.nim" with this content:
{.passL:"foo.lib".}
proc add*(a, b: int):int {.cdecl, importc.}
proc sub*(a, b: int):int {.cdecl, importc.}
echo add(10, 5)
echo sub(10, 5)
if i simply build it with "nim c -r main.nim", i get the following output and error:
P:\Nim\LearnCBinding>nim c -r main.nim
Hint: used config file 'C:\nim-1.5.1\config\nim.cfg' [Conf]
Hint: used config file 'C:\nim-1.5.1\config\config.nims' [Conf]
....CC: stdlib_io.nim
CC: stdlib_system.nim
CC: main.nim
Hint: [Link]
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x1f6): multiple definition of `PreMainInner'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x120): first defined here
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x20a): multiple definition of `PreMain'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x134): first defined here
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x240): multiple definition of `NimMainInner'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x16f): first defined here
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x254): multiple definition of `NimMain'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x183): first defined here
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x285): multiple definition of `main'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x1b4): first defined here
foo.lib(#mfoo.nim.c.o):#mfoo.nim.c:(.text+0x2da): multiple definition of `NimMainModule'
C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o:#mmain.nim.c:(.text+0x209): first defined here
collect2.exe: error: ld returned 1 exit status
Error: execution of an external program failed: 'C:\nim-1.5.1\dist\mingw64\bin\gcc.exe -o P:\Nim\LearnCBinding\main.exe C:\Users\Peter\nimcache\main_d\stdlib_io.nim.c.o C:\Users\Peter\nimcache\main_d\stdlib_system.nim.c.o C:\Users\Peter\nimcache\main_d\#mmain.nim.c.o foo.lib '
Because of the multiple definition error i also tried to build foo.lib with parameter "--noMain:on", but it doesnt make any difference.
Do you have the same problem? By the way i use the current version of Nim "nim-1.5.1" and reinstalled MingW with the finish.exe from nim.
I will try to help you with the following error you have:
undefined reference to `Soloud_create'
but i will assume that you have configured your environment so you can compile your nim programs with visual studio compiler (by adding --cc:vcc to your compile command)
this is because you already seem to have visual studio 2017 and you are compiling soloud static library with it. I think this is the best option when you are compiling with one compiler both: static library and executable that will use it.
open your static library (soloud_static_x64.lib) with some text/hex editor and search for "Soloud_create". i guess you will not find anything. so why is that? because for some reason author decided to not include "C interfacing" in a static library project. so it contains only C++ symbols and not pure C symbols that are needed for our solouddotnim.nim module.
let's try to find out what .cpp file we need for that. i noticed this information on official web site of Soloud - http://sol.gfxile.net/soloud/c_api.html
so i guess we need only one file: soloud_c.cpp
let's try to just include it in SoloudStatic.vcxproj file generated by you with Genie. like this:
..
<ClCompile Include="..\..\src\c_api\soloud_c.cpp">
</ClCompile>
..
and recompile our static library. i use this command in powershell:
& 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Preview\MSBuild\Current\Bin\MSBuild.exe' /p:PlatformToolset=v142`;WindowsTargetPlatformVersion=10`;Configuration=Release`;Platform=x64 .\SoloudStatic.vcxproj
but you can compile how you want. just make sure that it's architecture is really x64. you can check it with this command:
dumpbin /headers soloud_static_x64.lib | more
finally just link it with your nim file. add this line to the top:
{.link:"soloud_static_x64.lib".}
and compile nim file with this command:
nim c --cc:vcc --dynlibOverride:libsoloud.dll -r "examples\00-ogg\Example00_ogg.nim"
I have a CUDA project in a .cu file that I would like to compile to a .mex file using mexcuda. Because my code makes use of the 64-bit floating point atomic operation atomicAdd(double *, double), which is only supposed for GPU devices of compute capability 6.0 or higher, I need to specify this as a flag when I am compiling.
In my standard IDE, this works fine, but when compiling with mexcuda, this is not working as I would like. In this post on MathWorks, it was suggested to use the following command (edited from the comment by Joss Knight):
mexcuda('-v', 'mexGPUExample.cu', 'NVCCFLAGS=-gencode=arch=compute_60,code=sm_60')
but when I use this command on my file, the verbose option spits out the following line last:
Building with 'NVIDIA CUDA Compiler'.
nvcc -c --compiler-options=/Zp8,/GR,/W3,/EHs,/nologo,/MD -
gencode=arch=compute_30,code=sm_30 -gencode=arch=compute_50,code=sm_50 -
gencode=arch=compute_60,code=sm_60 -
gencode=arch=compute_70,code=\"sm_70,compute_70\"
(and so on), which signals to me that the specified flag was not passed to the nvcc properly. And indeed, compilation fails with the following error:
C:/path/mexGPUExample.cu(35): error: no instance of overloaded function "atomicAdd" matches
the argument list. Argument types are: (double *, double)
The only other post I could find on this topic was this post on SO, but it is almost three years old and seemed to me more like a workaround - one which I do not understand even after some research, otherwise I would have tried it - rather than a true solution to the problem.
Is there a setting I missed, or can this simply not be done without a workaround?
I was able to work my way around this problem after some messing around with the standard xml-files in the MatLab folder. The following steps allowed me to compile using -mexcuda:
-1) Go to the folder C:\Program Files\MATLAB\-version-\toolbox\distcomp\gpu\extern\src\mex\win64, which contains xml-files for different versions of msvcpp;
-2) Make a backup of the file that corresponds to the version you are using. In my case, I made a copy of the file nvcc_msvcpp2017 and named it nvcc_msvcpp2017_old, to always have the original.
-3) Open nvcc_msvcppYEAR with notepad, and scroll to the following block of lines:
COMPILER="nvcc"
COMPFLAGS="--compiler-options=/Zp8,/GR,/W3,/EHs,/nologo,/MD $ARCHFLAGS"
ARCHFLAGS="-gencode=arch=compute_30,code=sm_30 -gencode=arch=compute_50,code=sm_50 -gencode=arch=compute_60,code=sm_60 -gencode=arch=compute_70,code=\"sm_70,compute_70\" $NVCC_FLAGS"
COMPDEFINES="--compiler-options=/D_CRT_SECURE_NO_DEPRECATE,/D_SCL_SECURE_NO_DEPRECATE,/D_SECURE_SCL=0,$MATLABMEX"
MATLABMEX="/DMATLAB_MEX_FILE"
OPTIMFLAGS="--compiler-options=/O2,/Oy-,/DNDEBUG"
INCLUDE="-I"$MATLABROOT\extern\include" -I"$MATLABROOT\simulink\include""
DEBUGFLAGS="--compiler-options=/Z7"
-4) Remove the architectures that will not allow your code to compile, i.e. all the architecture flags below 60 in my case:
ARCHFLAGS="-gencode=arch=compute_60,code=sm_60 -gencode=arch=compute_70,code=\"sm_70,compute_70\" $NVCC_FLAGS"
-5) I was able to compile using mexcuda after this. You do not need to specify any architecture flags in the mexcuda call.
-6) (optional) I suppose you want to revert this change after you are done with the project that required you to make this change, if you want to ensure maximum portability of the code you will compile after this.
Note: you will need administrator permission to make these changes.
I'm getting an RC2247 error (Symbol name too long) when attempting to display the dialogs in a Win32 app. The error is occurring in this file:
C:\Program Files (x86)\Windows Kits\10\Include\10.0.17134.0\um\winnt.h
This worked fine under VS2015. The error started when I upgraded to VS2017.
I have seen the following posts, they don't appear to be relevant because they pertain to prsht.h:
RC2247 : Cannot open Rc file : Resource explorer cannot load resource ; Load failed
http://social.msdn.microsoft.com/Forums/en-US/vcprerelease/thread/4a648d6a-ea81-44d3-89c2-57fa5caa6fd6
The error disappears if I comment out the entire resource.rc file. The error occurs if the RC contains the single line:
#include <winnt.h>
When I click on "Edit code"", nothing happens.
I am grateful for any suggestions.
We must not include windows.h or winnt.h to .rc file - this headers - for are for c/c++ compilers and not designed for RC (resource compiler which process .rc file). as result if we include such files we may get errors.
We need include #include <winres.h> to .rc files. this file specially designed for the RC compiler. Internally it included:
#include <winuser.rh>
#include <commctrl.rh>
#include <dde.rh>
#include <winnt.rh>
#include <dlgs.h>
#include <winver.h>
and define some macros.
All standard windows definitions, which we need/use in rc file - exist in winres.h (and it subincludes). From another side in it no extra symbols/definitions wich exist in windows.h - it's not needed for rc and some time can cause errors.
So simply #include <winres.h> at the beginning of resource files and all will be OK.
I have run into this complaint rc2247 when trying to edit the rc file in a project compiled in vs2010. This had been upgraded from vs2050. I found that if I chose to open the .sin file via 'version selector' instead of going straight to vs2010, then it still opens in vs2010 but there is no problem in editing the rc. I don't pretend to understand !
CalendarMan
After overcoming some troubles with installation I tried to use erlang debugger on simple module:
I included -compile([debug_info]). option in source file and compiled with:
1> c(test_module).
This did not work as expected: After running
2> debugger:start().
the monitor window appeared, then I clicked
Module->Interpret...->test_module.erl
and got error
"Error when interpreting: test_module.erl: No debug_info in BEAM file".
Deleting -compile([debug_info]). line and changing
1> c(test_module).
to
1> c(test_module, [debug_info]).
solved the problem.
What is the difference between these two ways of setting compilation option, why one works and the other does not?
According to the docs, the two ways should be equivalent:
Note that all the options except the include path ({i,Dir}) can also
be given in the file with a -compile([Option,...]). attribute.
(From Erlang -- compile doc)
Check this question for more info.
I have been trying to get GTK 3.0 to work, and have followed all the steps here
How to configure gtk on Visual studio 2010
And changing to 3.0 where needed to get GTK to work, and it seems to have loaded everything it needs in order to compile, but it gives me the error
Error 3 error LNK1104: cannot open file 'gtk-3.lib'
Whenever I try to run the program.
I am using visual studios 2012, but this was the only place i found anything about getting GTK to run on any visual studios.
Here is the code I am using:
#include <gtk-3.0\gtk\gtk.h>
int main(int argc, char* argv[])
{
gtk_init(&argc, &argv);
GtkWidget* window = gtk_window_new(GTK_WINDOW_TOPLEVEL);
//gtk_widget_get_preferred_size(window, 300, 200);
g_signal_connect(G_OBJECT(window), "destroy", G_CALLBACK(gtk_main_quit), NULL);
gtk_window_set_title(GTK_WINDOW(window), "GTK+ with VS2010");
gtk_widget_show(window);
gtk_main();
return 0;
}
I commented out the gtk_widget_get_prefered_size call because it is irrelevant to the problem
any suggestions? I've looked in several places but none came up with clear answers.
The library gtk-3.lib does not exist. In fact, the library reference is not required to build your GTK 3 application. The pkg-config helper doesn't seem to generate the correct linker flags needed to link your application.
Just add in your Additional Options area all the existing libraries found in your GTK package (\gtk3\lib). The lib files for my bundle (gtk+-bundle_3.6.4-20130921) were as follows:
atk-1.0.lib cairo.lib fontconfig.lib gailutil.lib gdk-win32-3.0.lib gdk_pixbuf-2.0.lib gio-2.0.lib glib-2.0.lib gmodule-2.0.lib gobject-2.0.lib gthread-2.0.lib gtk-win32-3.0.lib pango-1.0.lib pangocairo-1.0.lib pangoft2-1.0.lib pangowin32-1.0.lib
(or you can go to your library path via a command prompt and enter dir *.lib /B)
Don't forget to include the /ENTRY:mainCRTStartup flag mention in the initial answer you started with.
you may have to edit your project settings or use a pragma comment to link with your gtk library:
#pragma comment(lib, "gtk-3")//if the libray is on your project's path
#define PATH "C:\\example\\"
#pragma comment(lib, PATH"gtk-3")//if the library is on PATH