A minor performance improvement -- make swizzle() faster.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5431

diff --git a/coregrind/m_mallocfree.c b/coregrind/m_mallocfree.c
index 98d03a9a0618dfb380254c49822de3274bc2b705..faf3e56e6f63216cfd55de251ea0904911448fea 100644 (file)
--- a/coregrind/m_mallocfree.c
+++ b/coregrind/m_mallocfree.c
@@ -704,7 +704,12 @@ void swizzle ( Arena* a, UInt lno )
    if (p_best == NULL) return;
 
    pn = pp = p_best;
-   for (i = 0; i < 20; i++) {
+
+   // This loop bound was 20 for a long time, but experiments showed that
+   // reducing it to 10 gave the same result in all the tests, and 5 got the
+   // same result in 85--100% of cases.  And it's called often enough to be
+   // noticeable in programs that allocated a lot.
+   for (i = 0; i < 5; i++) {
       pn = get_next_b(pn);
       pp = get_prev_b(pp);
       if (pn < p_best) p_best = pn;