* Added support for MSVC's ``__vectorcall`` calling convention as
``x86_vectorcallcc``.
-* ... next change ...
-
.. NOTE
If you would like to document a larger change, then you can add a
subsection about it right here. You can copy the following boilerplate
It is now just a wrapper, which simplifies using object::Binary with other
users of the underlying file.
+
IR in object files is now supported
-----------------------------------
Change in the representation of lazy loaded funcs
-------------------------------------------------
-Lazy loaded functions are now represented is a way that ``isDeclaration``
+Lazy loaded functions are now represented in a way that ``isDeclaration``
returns the correct answer even before reading the body.
This was done to simplify compatibility with python 3.
+
The leak detector has been removed
----------------------------------
-In practice tools like asan and valgrind were finding way more bugs than
+In practice, tools like asan and valgrind were finding way more bugs than
the old leak detector, so it was removed.
@g = global i32 0, comdat($c)
@c = global i32 0, comdat
-The version without the parentheses is a syntatic sugar for a comdat with
+The version without the parentheses is a syntactic sugar for a comdat with
the same name as the global.
+
Added support for Win64 unwind information
------------------------------------------
users should call ``CaptureStackBackTrace``, or implement equivalent
functionality by consulting the unwind tables present in the binary.
+
Diagnostic infrastructure used by lib/Linker and lib/Bitcode
------------------------------------------------------------
but both mechanisms will be supported in 3.7.
-Changes to the ARM Backend
---------------------------
-
- During this release ...
-
-
Changes to the MIPS Target
--------------------------
microMIPS for the O32 ABI on little endian targets, and code generation for
microMIPS is almost completely passing the test-suite.
+
ABI
^^^
has been fixed when the fastcc calling convention is used with 64-bit FPU's
and -mno-odd-spreg.
+
LLVMLinux
^^^^^^^^^
* Added support for a number of directives used by Linux to the Integrated
Assembler.
+
Miscellaneous
^^^^^^^^^^^^^
is in use and will be removed in LLVM 3.7. These names have never been
supported by the GNU Assembler for these ABI's.
+
Changes to the PowerPC Target
-----------------------------
* LLVM now has a POWER8 instruction scheduling description.
-* Address Sanitizer (ASAN) support is now fully functional.
+* AddressSanitizer (ASan) support is now fully functional.
* Performance of simple atomic accesses has been greatly improved.
* There have been many smaller bug fixes and performance improvements.
+
Changes to the OCaml bindings
-----------------------------
* As usual, many more functions have been exposed to OCaml.
+
Go bindings
-----------
* A set of Go bindings based on `gollvm <https://github.com/go-llvm/llvm>`_
was introduced in this release.
+
External Open Source Projects Using LLVM 3.6
============================================
a lot of other language and tools projects. This section lists some of the
projects that have already been updated to work with LLVM 3.6.
+
Portable Computing Language (pocl)
----------------------------------
the presence of work-group barriers. This enables static parallelization of
the fine-grained static concurrency in the work groups in multiple ways.
+
TTA-based Co-design Environment (TCE)
-------------------------------------
loads them in to the compiler backend as runtime libraries to avoid
per-target recompilation of larger parts of the compiler chain.
+
Likely
------
`Likely <http://www.liblikely.org>`_ is an embeddable just-in-time Lisp for
-image recognition and heterogenous computing. Algorithms are just-in-time
+image recognition and heterogeneous computing. Algorithms are just-in-time
compiled using LLVM's MCJIT infrastructure to execute on single or
multi-threaded CPUs and potentially OpenCL SPIR or CUDA enabled GPUs.
Likely seeks to explore new optimizations for statistical learning
compile-time evaluation of a function (the learning algorithm) with constant
arguments (the training data).
+
LDC - the LLVM-based D compiler
-------------------------------
PowerPC (32/64 bit). Ports to other architectures like ARM, AArch64 and MIPS64
are underway.
+
LLVMSharp & ClangSharp
----------------------