As I know, CUDA supports C and C++. But I can't' use C++ in my kernel.
I try a simple example like this
__global__ void simple(){
cout<<"abc";
}
That is error. But if I change to printf("abc"); it is right.
Can you explain for me? Thank you very much!
From CUDA 7.5 nvidia slides:
C++11 Supported features:
auto
lambdas
std::initializer_list
variadic templates
static_asserts
constexpr
rvalue references
range based for loops
C++ Not supported features
thread_local
Standard libraries: std::*
std::cout is defined in the C++ standard library which is not supported by CUDA. Use C printf
From CUDA 6.5, the ‘compute_11′, ‘compute_12′, ‘compute_13′, ‘sm_11′, ‘sm_12′, and ‘sm_13′ architectures are deprecated. So nvcc will compile by default to CC 2.0 enabling printf support.
More info here and here
CUDA doesn't link the libraries & header files that are required to use the cout function. However, you can enable the use of printf()
This answer explains the process which enables this feature:
printing from cuda kernels
quoted here for easier access:
To enable use of plain printf() on devices of Compute Capability >= 2.0, it's important to compile for CC of at least CC 2.0 and disable the default, which includes a build for CC 1.0.
Right-click the .cu file in your project, select Properties, select Configuration Properties | CUDA C/C++ | Device. Click on the Code Generation line, click the triangle, select Edit. In the Code Generation dialog box, uncheck Inherit from parent or project defaults, type compute_20,sm_20 in the top window, click OK.
Related
[summary]
I can not compile vulkan program including "vkGetAccelerationStructureBuildSizesKHR" or "vkCreateAccelerationStructureKHR" with errors "undefined reference to `vkGetAccelerationStructureBuildSizesKHR'".
[environment]
OS : Ubuntu 20.04
command : clang++ with option -lvulkan
vulkan version : 1.2.170
[what tried]
I guess that it is need more library.
I added "-lvulkan_radeon", but it does not make any change.
For reference, If my program does not include any new feature of ray tracing, I can build it normally.
I read that ray tracing features are officially supported from version 1.2.167, so I expect that I can build it without any additional works, but I can't.
Is there any additional library to be linked?
As with all functions provided by extensions and not being part of the core, you have to manually define and get the function pointers in your application before you can call it:
PFN_vkGetAccelerationStructureBuildSizesKHR pfnGetAccelerationStructureBuildSizesKHR;
pfnGetAccelerationStructureBuildSizesKHR = reinterpret_cast<PFN_vkGetAccelerationStructureBuildSizesKHR>(vkGetDeviceProcAddr(device, "vkGetAccelerationStructureBuildSizesKHR"));
Within my bazel.rc file for the C++ Bazel build options, I have specified:
'build --cxxopt="-std=c++1y"'
When I try to change to build --cxxopt="-std=c++14", bazel Throws error of :
''build --cxxopt="-std=c++14"' is not an option'
, with a cursor below 4 in c++14. If I change the 4 to y (c++1y), it compiles with no problem. I tried adding 'std:make_unique' in my code (c++14 addition) , and it clearly does not compile (make _unique is not part of std) which is assume is complaining that c++14 is not set as the standard, hence no make_unique for me.
What is the latest supported version of C++ in bazel? More specifically, how do I enable C++14 / C++17 (and even C++2x) for Bazel build, it those are supported? Thanks!
NOTE: this is not the same as How to set C++ standard version when build with Bazel??. (I am not asking about cpp 11 and I am asking about a very specific toolchain -- GCC/g++) I am not using any of the tools suggested by that thread. I am using GCC / g++ and am restricted in not being able to use the tools suggested by the answer in the previous question, I've RTFMd and googled. Thanks in advance.
Question
Modern Fortran offers a few cross-platform mechanisms to record the compiler version and settings used to build an application. What methods does C++17 have to capture this information? The book by Horton and Van Weert, Beginning C++17, does not appear to address this question.
The Fortran tools are surveyed below.
1. Access to compiler versions and options
The iso_fortran_env in Fortran provides a standard way to access the compiler version and settings used to compile a code. A sample snippet follows.
Code sample
program check_compiler
use, intrinsic :: iso_fortran_env, only : compiler_options, compiler_version
implicit none
write ( *, 100 ) "compiler version = ", compiler_version ()
write ( *, 100 ) "compiler options = ", trim ( compiler_options () )
100 format ( A, A, / )
stop "normal termination . . ."
end program check_compiler
Sample output
$ gfortran -o check_compiler check_compiler.f08
$ ./check_compiler
compiler version = GCC version 8.0.0 20170604 (experimental)
compiler options = -fPIC -mmacosx-version-min=10.12.7 -mtune=core2
STOP normal termination . . .
2. Probing and interacting with host OS
Fortran commands like execute_command_line, get_command, and get_environment_variable offer another route to record information at compile time.
What methods does C++17 have to capture this information?
None. The C++ standard does not even recognize the concept of "compiler" or "options"; there is merely the "implementation".
Furthermore, it would not really make sense, as different C++ files linked into the same program can be compiled with different options. And I'm not just talking about DLL/SOs; you can in theory statically link files that were compiled with different options or even different compiler versions.
Different compilers have ways to specify what version they are through macros. But each one has its own way to report this.
Searching the C++20 standard draft, which is available in GitHub, I find no results for closely-localted "compiler" and "version", nor have I found something like this looking at the text of the standard.
C++20 is at this time still very close to C++17, and certainly such a mechanism has not been removed, so I think it's pretty safe to say that there's no such thing in C++20.
Each compiler injects their own preproxessor tokens indicating itmwas compiled by them, and what version. These tokens are cross platform on compilers that compile on and to kore than one platdorm, such as icc, gcx and clang.
There are now standard defined ways to detect the existence of some srd header files. Boost has extensive headers that decode compiler capabilities based of a myriad of techniques.
__cplusplus in theory is defined to the standard version, but compilers lie.
The language standard specifies macros __cplusplus that encode the version of the standard that the compiler claims to support. It expands to 201703L on a C++17 compiler, 201710L on a C++14 compiler, and so on. It might also define _STDC and _STDC_VERSION. Beyond that, everything is a vendor-specific extension that you should look up in your compiler's manual.
Some but not all compilers, including GCC and Clang, predefine a macro named __VERSION__ that expands to a string describing the compiler version. You can check for this with #ifdef. Beyond that, many compilers contain macros that expand to version numbers, which you can stringify and concatenate. However, be aware that some compilers treat these as compatibility tests, and will claim to be a different compiler if you ask. In addition to its own version numbers, Clang defines __GNUC__, __GNUC_VERSION__ and __GNUC_PATCHLEVEL__ to indicate its compatibility with GCC, and the Windows version will also define _MSC_VER, _MSC_FULL_VER and so on in its Microsoft-compatiblity mode.
You could therefore create a complicated set of nested #elif blocks to recognize various compilers' version macros, but it could never be complete or forward-compatible.
I have a C++ application using GTKmm version 3.
From one version of the library to another, some method definition change so I would like to handle them in order to allow my sources to compile with a previous and actual version of GTKmm.
When I have written a kernel module, I did the same thing using conditional operators like the following:
#if LINUX_VERSION_CODE >= KERNEL_VERSION(3,6,0)
// Something is done when kernel version is 3.6 or higher
#else
// Something else is done for older versions
#endif
So the question is: is this possible in C++ and GTKMM ?
Yes - gtkmm has the defines GTKMM_MAJOR_VERSION, GTKMM_MINOR_VERSION, and GTKMM_MICRO_VERSION.
How can I use C++11 when programming the Arduino?
I would be fine using either the Arduino IDE or another environment. I am most interested in the core language improvements, not things that require standard library changes.
As of version 1.6.6, the Arduino IDE enables C++11 by default.
For older versions, read on:
It is very easy to change the flags for any element of the toolchain, including the assembler, compiler, linker or archiver.
Tested on the Arduino IDE version 1.5.7 (released on July 2014),
Locate the platform.txt file,
AVR architecture => {install path}\hardware\arduino\avr\platform.txt
SAM architecture => {install path}\hardware\arduino\sam\platform.txt
Inside that file, you can change any flag, for instance,
compiler.c.flags for changing the default compilation flags for C++ files.
compiler.cpp.flags for changing the default compilation flags for C++ files.
You can also change any of the "recipes" or compile patters, at the corresponding section of the configuration file, under the title "AVR/SAM compile patterns".
After making the changes, you must restart the Arduino IDE, at least on version 1.5.7.
For instance,
To enable support for C++11 (C++0x), tested on Arduino IDE versions 1.5.7 and 1.5.8, you will simply add the flag "-std=gnu++11" at the end of the line starting with compiler.cpp.flags=".
It is expected that C++11 is enabled by default in the near future on the Arduino IDE. However, as of version 1.5.8 (Oct 2014) it is still not the case.
Arduino IDE 1.6.6 and newer have C++11 enabled by default (they have the compiler flag "-std=gnu++11" set in the platform.txt file).
Firstly, only GCC 4.7 and above (and therefore AVR-GCC 4.7 and above) support C++11. So, check the versions installed with:
gcc --version
avr-gcc --version
If AVR-GCC is 4.7 or higher, then you may be able to use C++11.
The Arduino IDE does not support custom compiler flags. This has been requested, but has not yet been implemented.
So, you are left with having to use other environments or to compile your program directly from the command line.
In case, of compiling directly from the command line using AVR-GCC, you simply need to add an extra compiler flag for enabling C++11 support.
-std=c++11
For specific development environments, most would support editing of the compiler flags from the build options within the IDE. The above mentioned flag needs to be added to the list of flags for each environment.
C++0x was the name of working draft of the C++11 standard. C++0x support is available GCC 4.3 onwards. However, this is strictly experimental support so you cannot reliably expect C++11 features to be present. Here is the complete list of features available with the corresponding version of GCC. The availability of features in AVR-GCC will be the same as what's available in the corresponding GCC version.
The compiler flag for C++0x is:
-std=c++0x
Please, note, that there is no easy way to specify additional flags from Arduino IDE or use other IDE (Eclipse, Code::Blocks, etc.) or command line.
As a hack, you can use a small proxy program (should be cross-platform):
//============================================================================
// Name : gcc-proxy.cpp
// Copyright : Use as you want
// Description : Based on http://stackoverflow.com/questions/5846934/how-to-pass-a-vector-to-execvp
//============================================================================
#include <unistd.h>
#include <iostream>
#include <vector>
#include <fstream>
using namespace std;
int main(int argc, char *argv[]) {
vector<string> arguments;
vector<const char*> aptrs;
// Additional options, one per line
ifstream cfg((string(argv[0]) + ".ini").c_str());
if (cfg.bad())
cerr << "Could not open ini file (you're using proxy for some reason, er?)" << endl;
string arg;
while (cfg) {
getline(cfg, arg);
if(arg == "\r" || arg == "\n")
continue;
arguments.push_back(arg);
}
for (const string& arg : arguments)
aptrs.push_back(arg.c_str());
for (int i = 1; i < argc; ++i)
aptrs.push_back(argv[i]);
// Add null pointer at the end, execvp expects NULL as last element
aptrs.push_back(nullptr);
// Pass the vector's internal array to execvp
const char **command = &aptrs[0];
return execvp(command[0], command);
}
Compile the program.
Rename the original avr-g++.exe to avr-g++.orig.exe (or any other name).
Create avr-g++.ini file where the first line is FULL path to the original program (e.g. D:\Arduino\hardware\tools\avr\bin\avr-g++.orig.exe) and add additional parameters, one per line, as desired.
You're done!
Example avr-g++.ini:
D:\Arduino\hardware\tools\avr\bin\avr-g++.orig.exe
-std=c++0x
I use Ino and this worked:
ino build -cppflags="-std=c++0x"
This generated a hex file at least 15k in size (that's with optimizations turned on), compared to about 5k for the standard build, which is a consideration for a poor little ATmega328. It might be okay for one of the microcontrollers with a lot more program space.
If you need more control and a better IDE, I recommend using Sloeber Plugin for Eclipse or the Sloeber IDE itself.
Creating more complicated code is much easier using this IDE. It also allows to add flags to the compiler (C, C++ and linker). So to customize the compile, just right click on project and select Properties. In the Properties window, select Arduino → Compiler Option. This way you can add options to your build.