diff options
| author | Devang Patel <dpatel@apple.com> | 2008-06-05 21:44:00 +0000 |
|---|---|---|
| committer | Devang Patel <dpatel@apple.com> | 2008-06-05 21:44:00 +0000 |
| commit | 83a15d7209752bdcc9e37f93b69a4a1b9c21bb16 (patch) | |
| tree | 8c7fee2bce11b28baa8ab4ef96ca0f89cb062fe7 /docs | |
| parent | 6e60ea42e8296265bcb03dc5064e40a927733ae5 (diff) | |
Describe loop index split.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52021 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'docs')
| -rw-r--r-- | docs/ReleaseNotes.html | 16 |
1 files changed, 15 insertions, 1 deletions
diff --git a/docs/ReleaseNotes.html b/docs/ReleaseNotes.html index a752e53bca..3ee762a01a 100644 --- a/docs/ReleaseNotes.html +++ b/docs/ReleaseNotes.html @@ -219,7 +219,21 @@ LLVM 2.3 optimizers support a few major enhancements:</p> <ul> -<li>Loop index set splitting on by default: describe me.</li> +<li>Loop index set splitting on by default.<p> +This transformation hoists conditions from loop bodies and reduces loop's +iteration space to improve performance. For example, <p> +<pre> +for (i = LB; i < UB; ++i) + if (i <= NV) + LOOP_BODY +</pre> +is transformed into +<pre> +NUB = min(NV+1, UB) +for (i = LB; i < NUB; ++i) + LOOP_BODY +</pre> +</li> <li>LLVM includes a new <tt>memcpy</tt> optimization pass which optimizes out dead <tt>memcpy</tt> calls, unneeded copies of aggregates, and handles the return slot optimization. The LLVM optimizer now notices long sequences of consequtive |