Namespaces
Name |
Description |
Top‐level LLVM library namespace. |
Macros
Name |
Description |
PTRADD represents pointer arithmetic semantics, for targets that opt in using shouldPreservePtrArith(). ptr = PTRADD ptr, offset |
|
The internal structure of FunctionData. This makes sure that changes to the fields of FunctionData either get automatically captured on the llvm side, or force a manual corresponding update. |
|
DEBUG_WITH_TYPE macro ‐ This macro should be used by passes to emit debug information. If the '‐debug' option is specified on the commandline, and if this is a debug build, then the code specified as the option to the macro will be executed. Otherwise it will not be. Example: |
|
Macro for generating in‐class operand accessor declarations. It should only be called in the public section of the interface. |
|
Macro for generating out‐of‐class operand accessor definitions |
|
TableGen will produce 2 enums, RTLIB::Libcall and RTLIB::LibcallImpl. RTLIB::Libcall describes abstract functionality the compiler may choose to access, RTLIB::LibcallImpl describes a particular ABI implementation, which includes a name and type signature. |
|
The floating‐point selectors get<OP>(EVT) return the <OP> libcall for the given type, or UNKNOWN_LIBCALL. Generated from the RuntimeLibcallFamily table in RuntimeLibcalls.td. |
|
These methods just forward to Create, and are useful when you statically know what type of instruction you're going to create. These helpers just save some typing. |
|
@} |
|
Denotes a value read from an index section that does not correspond to any of the supported standards. |
|
These methods just forward to Create, and are useful when you statically know what type of instruction you're going to create. These helpers just save some typing. |
|
Vocabulary layout constants |
|
output. The default level is 1, and is in increasing level of verbosity. |
|
Helpers to get DEBUG_TYPE as a StringRef, even when DEBUG_TYPE is not defined (in which case it expands to "DEBUG_TYPE") |
|
Helper to provide the default level (=1) or type (=DEBUG_TYPE). This is used when a single argument is passed to LDBG() (or LDBG_OS()), if it is an integer we return DEBUG_TYPE and if it is a string we return 1. When DEBUG_TYPE is not defined, we return the current file name instead. |
|
When a single argument is provided. This can be either a level or the debug type. If a level is provided, we default the debug type to DEBUG_TYPE, if a string is provided, we default the level to 1. |
|
LDBG_OS() is a macro that behaves like LDBG() but instead of directly using it to stream the output, it takes a callback function that will be called with a raw_ostream. This is useful when you need to pass a |
|
This macro is the core of the LDBG_OS() macros. It is used to print the debug output with the given stream, level, type, file, and line number. |
|
This macro expands to the stream to use for output, we use a macro to allow unit‐testing to override. |
|
These macros are helpers to implement LDBG() with an increasing amount of optional arguments made explicit. |
|
This macro is the core of the LDBG() implementation. It is used to print the debug output with the given stream, level, type, file, and line number. |
|
This macro is a helper when LDBG() is called with 2 arguments. In this case we want to force the first argument to be the type for consistency in the codebase. We trick this by casting the first argument to a (const char *) which won't compile with an int. |
|
LLVM_DECLARE_ENUM_AS_BITMASK can be used to declare an enum type as a bit set, so that bitwise operation on such enum does not require static_cast. |
|
Helper macro to declare registry class. |
|
Variants of the registration macros that take an explicit export annotation (EXPORT_ABI) in place of the hard‐coded |
|
Helper macro to define registry class. |
|
LLVM_ENABLE_BITMASK_ENUMS_IN_NAMESPACE() pulls the operator overloads used by LLVM_MARK_AS_BITMASK_ENUM into the current namespace. |
|
LLVM_MARK_AS_BITMASK_ENUM lets you opt in an individual enum type so you can perform bitwise operations on it without putting static_cast everywhere. |
|
macro LLVM_PLUGIN_API_VERSION Identifies the API version understood by this plugin. |
|
Utility for declaring that a std::vector of a particular type should be considered a YAML document list. |
|
Utility for declaring that a std::vector of a particular type should be considered a YAML flow sequence. |
|
Utility for declaring that a std::vector of a particular type should be considered a YAML sequence. |
|
Utility for declaring that std::map<std::string, _type> should be considered a YAML map. |
|
YAML I/O does conversion based on types. But often native data types are just a typedef of built in intergral types (e.g. int). But the C++ type matching system sees through the typedef and all the typedefed types look like a built in type. This will cause the generic YAML I/O conversion to be used. To provide better control over the YAML conversion, you can use this macro instead of typedef. It will create a class with one field and automatic conversion operators to and from the base type. Based on BOOST_STRONG_TYPEDEF |
|
Allow ObjectStore implementations to create internal handles. |
|
Declarations for LLVM‐IR types (simple, array, function and structure) are generated below. Their names are defined and used in OpenMPKinds.def. Here we provide the declarations, the initializeTypes function will provide the values. |
|
@} |
|
TensorSpec encapsulates the specification of a tensor: its dimensions, or "shape" (row‐major), its type (see TensorSpec::getDataType specializations for supported types), its name and port (see "TensorFlow: Large‐Scale Machine Learning on Heterogeneous Distributed Systems", section 4.2, para 2: https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45166.pdf) |