Use AlignedBuffer helper class in rgbProc, use SSE in standard tone curve application.

rtengine/alignedbuffer.h

@@ -21,6 +21,10 @@
 #include <cstdlib>
 #include <utility>
 
+inline size_t padToAlignment(size_t size, size_t align = 16) {
+    return align * ((size + align - 1) / align);
+}
+
 // Aligned buffer that should be faster
 template <class T> class AlignedBuffer
 {