Do 64-bit loops in QBitArray::count(bool)

On 64-bit platforms, with unaligned loads, this is defintely an
improvement since we can run fewer instructions. On 32-bit platforms
with unaligned loads, we'll do the exact same number of loads. On
platforms without unaligned loads, it's no worse.

Change-Id: Idd5dd5213975d77bbc3adf486adbf6f8ef071341
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>

1 files changed, 6 insertions, 1 deletions

diff --git a/src/corelib/tools/qbitarray.cpp b/src/corelib/tools/qbitarray.cpp
index 169f0ce2c8..da2f48c071 100644
--- a/src/corelib/tools/qbitarray.cpp
+++ b/src/corelib/tools/qbitarray.cpp
@@ -194,7 +194,12 @@ int QBitArray::count(bool on) const
     // it's the QByteArray implicit NUL, so it will not change the bit count
     const quint8 *const end = reinterpret_cast<const quint8 *>(d.end());
 
-    while (bits + 3 <= end) {
+    while (bits + 7 <= end) {
+        quint64 v = qUnalignedLoad<quint64>(bits);
+        bits += 8;
+        numBits += int(qPopulationCount(v));
+    }
+    if (bits + 3 <= end) {
         quint32 v = qUnalignedLoad<quint32>(bits);
         bits += 4;
         numBits += int(qPopulationCount(v));