cuda-api-wrappers
Thin C++-flavored wrappers for the CUDA Runtime API
|
Definitions and utility functions relating to run-time compilation (RTC) of CUDA code using the NVRTC library. More...
#include "cuda/api/detail/option_marshalling.hpp"
#include "../api/device_properties.hpp"
#include "../api/device.hpp"
#include "../api/common_ptx_compilation_options.hpp"
#include <unordered_map>
#include <unordered_set>
#include <sstream>
#include <string>
#include <vector>
#include <cstring>
#include <limits>
#include <iterator>
Go to the source code of this file.
Classes | |
struct | cuda::rtc::compilation_options_base_t< Kind > |
Compilation options common to all kinds of JIT-compilable programs. More... | |
class | cuda::rtc::compilation_options_t< Kind > |
Options to be passed to one of the NVIDIA JIT compilers along with a program's source code. More... | |
class | cuda::rtc::compilation_options_t< ptx > |
Options for JIT-compilation of CUDA PTX code. More... | |
struct | cuda::rtc::compilation_options_t< ptx >::caching_mode_spec_t |
Options for fully-specifying a caching mode. More... | |
class | cuda::rtc::compilation_options_t< cuda_cpp > |
Options for JIT-compilation of CUDA C++ code. More... | |
Namespaces | |
cuda | |
Definitions and functionality wrapping CUDA APIs. | |
cuda::rtc | |
Real-time compilation of programs using the NVIDIA libraries. | |
Typedefs | |
using | cuda::rtc::error::number_t = unsigned |
Errors, or problematic findings, by the compiler are identified by a number of this type. | |
Enumerations | |
enum | cpp_dialect_t { cpp03 = 0, cpp11 = 1, cpp14 = 2, cpp17 = 3, last = cpp17 } |
enum | cuda::rtc::error::handling_method_t { raise_error = 0, suppress = 1, warn = 2 } |
Possible ways of handling a potentially problematic finding by the compiler in the program source code. | |
enum | : rtc::optimization_level_t { O0 = 0, no_optimization = O0, O1 = 1, O2 = 2, O3 = 3, maximum_optimization = O3 } |
Commonly-used phrases regarding the optimization level (e.g. More... | |
Functions | |
template<typename CompilationOptions > | |
inline ::std::string | cuda::rtc::render (const CompilationOptions &opts) |
Definitions and utility functions relating to run-time compilation (RTC) of CUDA code using the NVRTC library.