--- a/src/core/model/int64x64.cc	Mon Sep 29 22:07:14 2014 +0100
+++ a/src/core/model/int64x64.cc	Thu Oct 02 11:01:24 2014 -0700
@@ -35,7 +35,7 @@ 
   int64_t hi = absVal.GetHigh ();
 
   // Save stream format flags
-  const std::streamsize precision = os.precision ();
+  const std::size_t precision = static_cast<std::size_t> (os.precision ());
   std::ios_base::fmtflags ff = os.flags ();
   const bool floatfield = os.flags () & std::ios_base::floatfield;
   os << std::setw (1) << std::noshowpos;
@@ -47,7 +47,7 @@ 
 
 
   int64x64_t low(0, absVal.GetLow ());
-  int places = 0;    // Number of decimal places printed so far
+  std::size_t places = 0;    // Number of decimal places printed so far
   bool more = true;  // Should we print more digits?
 
 #define HEXHILOW(hi, lo) \
--- a/src/fd-net-device/model/fd-net-device.cc	Mon Sep 29 22:07:14 2014 +0100
+++ a/src/fd-net-device/model/fd-net-device.cc	Thu Oct 02 11:01:24 2014 -0700
@@ -288,7 +288,7 @@ 
 ///  instead of memcpy to add the PI header.
 ///  It might be faster in this case to use memmove and avoid the extra mallocs.
 static void
-AddPIHeader (uint8_t *&buf, ssize_t &len)
+AddPIHeader (uint8_t *&buf, size_t &len)
 {
   // Synthesize PI header for our friend the kernel
   uint8_t *buf2 = (uint8_t*)malloc (len + 4);
@@ -526,7 +526,7 @@ 
 
   NS_ASSERT_MSG (packet->GetSize () <= m_mtu, "FdNetDevice::SendFrom(): Packet too big " << packet->GetSize ());
 
-  ssize_t len =  (ssize_t) packet->GetSize ();
+  size_t len =  (size_t)(packet->GetSize ());
   uint8_t *buffer = (uint8_t*)malloc (len);
   packet->CopyData (buffer, len);
 
@@ -539,7 +539,7 @@ 
   ssize_t written = write (m_fd, buffer, len);
   free (buffer);
 
-  if (written == -1 || written != len)
+  if (written < 0 || (size_t)(written) != len)
     {
       m_macTxDropTrace (packet);
       return false;
--- a/src/lte/model/lte-ue-phy.cc	Mon Sep 29 22:07:14 2014 +0100
+++ a/src/lte/model/lte-ue-phy.cc	Thu Oct 02 11:01:24 2014 -0700
@@ -728,7 +728,6 @@ 
 {
   NS_LOG_FUNCTION (this);
 
-
   // apply transmission mode gain
   NS_ASSERT (m_transmissionMode < m_txModeGain.size ());
   SpectrumValue newSinr = sinr;
@@ -742,12 +741,12 @@ 
     {
       cqi = m_amc->CreateCqiFeedbacks (newSinr, m_dlBandwidth);
 
-      int nLayer = TransmissionModesLayers::TxMode2LayerNum (m_transmissionMode);
-      int nbSubChannels = cqi.size ();
+      uint8_t nLayer = TransmissionModesLayers::TxMode2LayerNum (m_transmissionMode);
+      std::size_t nbSubChannels = cqi.size ();
       double cqiSum = 0.0;
-      int activeSubChannels = 0;
+      uint32_t activeSubChannels = 0;
       // average the CQIs of the different RBs
-      for (int i = 0; i < nbSubChannels; i++)
+      for (uint32_t i = 0; i < nbSubChannels; i++)
         {
           if (cqi.at (i) != -1)
             {
@@ -761,7 +760,7 @@ 
       dlcqi.m_cqiType = CqiListElement_s::P10; // Peridic CQI using PUCCH wideband
       NS_ASSERT_MSG (nLayer > 0, " nLayer negative");
       NS_ASSERT_MSG (nLayer < 3, " nLayer limit is 2s");
-      for (int i = 0; i < nLayer; i++)
+      for (uint8_t i = 0; i < nLayer; i++)
         {
           if (activeSubChannels > 0)
             {
@@ -780,14 +779,14 @@ 
   else if (Simulator::Now () > m_a30CqiLast + m_a30CqiPeriocity)
     {
       cqi = m_amc->CreateCqiFeedbacks (newSinr, GetRbgSize ());
-      int nLayer = TransmissionModesLayers::TxMode2LayerNum (m_transmissionMode);
-      int nbSubChannels = cqi.size ();
-      int rbgSize = GetRbgSize ();
+      uint8_t nLayer = TransmissionModesLayers::TxMode2LayerNum (m_transmissionMode);
+      std::size_t nbSubChannels = cqi.size ();
+      uint8_t rbgSize = GetRbgSize ();
       double cqiSum = 0.0;
-      int cqiNum = 0;
+      uint8_t cqiNum = 0;
       SbMeasResult_s rbgMeas;
-      //NS_LOG_DEBUG (this << " Create A30 CQI feedback, RBG " << rbgSize << " cqiNum " << nbSubChannels << " band "  << (uint16_t)m_dlBandwidth);
-      for (int i = 0; i < nbSubChannels; i++)
+      //NS_LOG_DEBUG (this << " Create A30 CQI feedback, RBG " << (uint16_t) rbgSize << " cqiNum " << nbSubChannels << " band "  << (uint16_t)m_dlBandwidth);
+      for (uint32_t i = 0; i < nbSubChannels; i++)
         {
           if (cqi.at (i) != -1)
             {
@@ -801,7 +800,7 @@ 
               //NS_LOG_DEBUG (this << " RBG CQI "  << (uint16_t) cqiSum / rbgSize);
               HigherLayerSelected_s hlCqi;
               hlCqi.m_sbPmi = 0; // not yet used
-              for (int i = 0; i < nLayer; i++)
+              for (uint8_t i = 0; i < nLayer; i++)
                 {
                   hlCqi.m_sbCqi.push_back ((uint16_t) cqiSum / rbgSize);
                 }
--- a/src/stats/model/basic-data-calculators.h	Mon Sep 29 22:07:14 2014 +0100
+++ a/src/stats/model/basic-data-calculators.h	Thu Oct 02 11:01:24 2014 -0700
@@ -160,11 +160,11 @@ 
         }
       else
         {
-          if (i < m_min)
+          if (i - m_min < 0)
             {
               m_min = i;
             }
-          if (i > m_max)
+          if (i - m_max > 0)
             {
               m_max = i;
             }
--- a/wscript	Mon Sep 29 22:07:14 2014 +0100
+++ a/wscript	Thu Oct 02 11:01:24 2014 -0700
@@ -330,7 +330,7 @@ 
             if conf.check_compilation_flag('-march=native'):
                 env.append_value('CXXFLAGS', '-march=native') 
             env.append_value('CXXFLAGS', '-fstrict-overflow')
-            if conf.env['CC_VERSION'] == cc_version_warn_strict_overflow:
+            if conf.env['CC_VERSION'] >= cc_version_warn_strict_overflow:
                 env.append_value('CXXFLAGS', '-Wstrict-overflow=5')
 
         if sys.platform == 'win32':