diff --git a/rtengine/helpersse2.h b/rtengine/helpersse2.h
index da1691748..0f1fc5759 100644
--- a/rtengine/helpersse2.h
+++ b/rtengine/helpersse2.h
@@ -26,17 +26,20 @@ typedef __m128i vint2;
 #define LVFU(x) _mm_loadu_ps(&x)
 #define STVF(x,y) _mm_store_ps(&x,y)
 #define STVFU(x,y) _mm_storeu_ps(&x,y)
+#define LVI(x) _mm_load_si128((__m128i*)&x)
 #else // there is a bug in gcc 4.7.x when using openmp and aligned memory and -O3, also need to map the aligned functions to unaligned functions for WIN32 builds
 #define LVF(x) _mm_loadu_ps((float*)&x)
 #define LVFU(x) _mm_loadu_ps(&x)
 #define STVF(x,y) _mm_storeu_ps(&x,y)
 #define STVFU(x,y) _mm_storeu_ps(&x,y)
+#define LVI(x) _mm_loadu_si128((__m128i*)&x)
 #endif
 #else
 #define LVF(x) _mm_load_ps((float*)&x)
 #define LVFU(x) _mm_loadu_ps(&x)
 #define STVF(x,y) _mm_store_ps(&x,y)
 #define STVFU(x,y) _mm_storeu_ps(&x,y)
+#define LVI(x) _mm_load_si128((__m128i*)&x)
 #endif
 
 #if defined(__x86_64__) && defined(__AVX__)
diff --git a/rtgui/histogrampanel.cc b/rtgui/histogrampanel.cc
index e5d3d560e..9c8163c8e 100644
--- a/rtgui/histogrampanel.cc
+++ b/rtgui/histogrampanel.cc
@@ -25,8 +25,7 @@
 #include "rtimage.h"
 #include "../rtengine/improccoordinator.h"
 #include "../rtengine/color.h"
-
-
+#include "../rtengine/opthelper.h"
 using namespace rtengine;
 
 extern Glib::ustring argv0;
@@ -852,7 +851,7 @@ void HistogramArea::update (LUTu &histRed, LUTu &histGreen, LUTu &histBlue, LUTu
     g_idle_add (histupdateUI, haih);
 }
 
-void HistogramArea::renderHistogram ()
+SSEFUNCTION void HistogramArea::renderHistogram ()
 {
 
     if (!is_realized ()) {
@@ -878,7 +877,7 @@ void HistogramArea::renderHistogram ()
 
         // make double copies of LUT, one for faster access, another one to scale down the raw histos
         LUTu rhchanged(256), ghchanged(256), bhchanged(256);
-        unsigned int lhisttemp[256], chisttemp[256], rhtemp[256], ghtemp[256], bhtemp[256];
+        unsigned int lhisttemp[256] ALIGNED16 {0}, chisttemp[256] ALIGNED16 {0}, rhtemp[256] ALIGNED16 {0}, ghtemp[256] ALIGNED16 {0}, bhtemp[256] ALIGNED16 {0};
         const int scale = (rawMode ? 8 : 1);
 
         for(int i = 0; i < 256; i++) {
@@ -937,31 +936,48 @@ void HistogramArea::renderHistogram ()
         if (!fullMode) {
             int area = 0;
 
-            if(!rawMode)
-                for (int i = 0; i < fullhistheight; i++) {
-                    for (int j = 0; j < 256; j++)
-                        if ((needLuma && lhisttemp[j] > i) || (needChroma && chisttemp[j] > i) || (needRed && rhtemp[j] > i) || (needGreen && ghtemp[j] > i) || (needBlue && bhtemp[j] > i)) {
-                            area++;
-                        }
+#ifdef __SSE2__
+            vint onev = _mm_set1_epi32(1);
+            vint iv = (vint)ZEROV;
+#endif
 
-                    if ((double)area / (256 * (i + 1)) < 0.3) {
-                        realhistheight = i;
-                        break;
-                    }
-                }
-            else
-                for (int i = 0; i < fullhistheight; i++) {
-                    for (int j = 0; j < 256; j++)
-                        if ((needRed && rhtemp[j] > i) || (needGreen && ghtemp[j] > i) || (needBlue && bhtemp[j] > i)) {
-                            area++;
-                        }
+            for (int i = 0; i < fullhistheight; i++) {
+#ifdef __SSE2__
+                vint areatempv = (vint)ZEROV;
+
+                for (int j = 0; j < 256; j += 4) {
+                    vmask mask1v = _mm_cmpgt_epi32(LVI(lhisttemp[j]), iv);
+                    vmask mask2v = _mm_cmpgt_epi32(LVI(rhtemp[j]), iv);
+                    vmask mask3v = _mm_cmpgt_epi32(LVI(ghtemp[j]), iv);
+                    vmask mask4v = _mm_cmpgt_epi32(LVI(bhtemp[j]), iv);
+                    mask1v = _mm_or_si128(mask1v, mask2v);
+                    mask3v = _mm_or_si128(mask3v, mask4v);
+                    mask2v = _mm_cmpgt_epi32(LVI(chisttemp[j]), iv);
+                    mask1v = _mm_or_si128(mask1v, mask3v);
+                    mask1v = _mm_or_si128(mask1v, mask2v);
+                    areatempv = _mm_add_epi32(areatempv, _mm_and_si128(mask1v, onev));
 
-                    if ((double)area / (256 * (i + 1)) < 0.3) {
-                        realhistheight = i;
-                        break;
-                    }
                 }
 
+                areatempv = _mm_add_epi32(areatempv, (vint)_mm_movehl_ps((vfloat)areatempv, (vfloat)areatempv));
+                areatempv = _mm_add_epi32(areatempv, _mm_shuffle_epi32(areatempv, 1));
+                area += _mm_cvtsi128_si32(areatempv);
+                iv = _mm_add_epi32(iv, onev);
+
+#else
+
+                for (int j = 0; j < 256; j++)
+                    if (lhisttemp[j] > i || rhtemp[j] > i || ghtemp[j] > i || bhtemp[j] > i || chisttemp[j] > i) {
+                        area++;
+                    }
+
+#endif
+
+                if ((double)area / (256 * (i + 1)) < 0.3) {
+                    realhistheight = i;
+                    break;
+                }
+            }
         }
 
         if (realhistheight < winh - 2) {