Hi eliben, TylerNowicki, aaron.ballman, gribozavr,
This patch extends the documentation for "#pragma clang loop" hints to include
the unroll and unroll_count directives.
http://reviews.llvm.org/D4198
Files:
docs/LanguageExtensions.rst
docs/ReleaseNotes.rst
include/clang/Basic/AttrDocs.td
Index: docs/LanguageExtensions.rst
===================================================================
--- docs/LanguageExtensions.rst
+++ docs/LanguageExtensions.rst
@@ -1770,8 +1770,9 @@
The ``#pragma clang loop`` directive is used to specify hints for optimizing the
subsequent for, while, do-while, or c++11 range-based for loop. The directive
-provides options for vectorization and interleaving. Loop hints can be specified
-before any loop and will be ignored if the optimization is not safe to apply.
+provides options for vectorization, interleaving, and unrolling. Loop hints can
+be specified before any loop and will be ignored if the optimization is not safe
+to apply.
A vectorized loop performs multiple iterations of the original loop
in parallel using vector instructions. The instruction set of the target
@@ -1786,9 +1787,16 @@
a cost model that depends on the register pressure and generated code size to
select the interleaving count.
-Vectorization is enabled by ``vectorize(enable)`` and interleaving is enabled
-by ``interleave(enable)``. This is useful when compiling with ``-Os`` to
-manually enable vectorization or interleaving.
+Loop unrolling replicates the body of a loop increasing the loop size
+and reducing the loop count. Loop control overhead can be reduced or
+eliminated, and additional ILP can be exposed. The unroller selects
+an unroll count based on a limit on the growth of code size and
+whether the loop can be unrolled completely.
+
+Vectorization is enabled by ``vectorize(enable)``, interleaving is enabled
+by ``interleave(enable)``, and unrolling is enabled by ``unrolling(enable)``.
+This is useful when compiling with ``-Os`` to manually enable vectorization or
+interleaving.
.. code-block:: c++
@@ -1798,22 +1806,28 @@
...
}
-The vector width is specified by ``vectorize_width(_value_)`` and the interleave
-count is specified by ``interleave_count(_value_)``, where
-_value_ is a positive integer. This is useful for specifying the optimal
-width/count of the set of target architectures supported by your application.
+If ``unroll(enable)`` is specified the unroller will attempt to unroll
+the loop completely if the trip count is known at compile time. If
+the loop count is not known the loop will still be unrolled subject to
+a limit on growth of code size.
+
+The vector width is specified by ``vectorize_width(_value_)``, interleave count
+is specified by ``interleave_count(_value_)``, and unroll count is specified by
+``unroll_count(_value)``, where _value_ is a positive integer. This is useful
+for specifying the optimal width/count of the set of target architectures
+supported by your application.
.. code-block:: c++
-
#pragma clang loop vectorize_width(2)
#pragma clang loop interleave_count(2)
+ #pragma clang loop unroll_count(4)
for(...) {
...
}
Specifying a width/count of 1 disables the optimization, and is equivalent to
-``vectorize(disable)`` or ``interleave(disable)``.
+``vectorize(disable)``, ``interleave(disable)``, or ``unroll(disable)``.
For convenience multiple loop hints can be specified on a single line.
Index: docs/ReleaseNotes.rst
===================================================================
--- docs/ReleaseNotes.rst
+++ docs/ReleaseNotes.rst
@@ -101,9 +101,10 @@
-----------------------
Loop optimization hints can be specified using the new `#pragma clang loop`
-directive just prior to the desired loop. The directive allows vectorization
-and interleaving to be enabled or disabled, and the vector width and interleave
-count to be manually specified. See language extensions for details.
+directive just prior to the desired loop. The directive allows vectorization,
+interleaving, and unrolling to be enabled or disabled. Vector width as well
+as interleave and unrolling count can be manually specified. See language
+extensions for details.
C Language Changes in Clang
---------------------------
Index: include/clang/Basic/AttrDocs.td
===================================================================
--- include/clang/Basic/AttrDocs.td
+++ include/clang/Basic/AttrDocs.td
@@ -1016,9 +1016,10 @@
let Category = DocCatStmt;
let Content = [{
The ``#pragma clang loop'' directive allows loop optimization hints to be
-specified for the subsequent loop. The directive allows vectorization
-and interleaving to be enabled or disabled, and the vector width and interleave
-count to be manually specified. See `language extensions
+specified for the subsequent loop. The directive allows vectorization,
+interleaving, and unrolling to be enabled or disabled. Vector width as well
+as interleave and unrolling count can be manually specified. See
+`language extensions
<http://clang.llvm.org/docs/LanguageExtensions.html#extensions-for-loop-hint-optimizations>'_
for details.
}];
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
