diff --git a/source/RISCV/specialize.h b/source/RISCV/specialize.h
index 70fb808..4137bd7 100644
--- a/source/RISCV/specialize.h
+++ b/source/RISCV/specialize.h
@@ -118,6 +118,17 @@
  softfloat_propagateNaNF16UI( uint_fast16_t uiA, uint_fast16_t uiB );
 
 /*----------------------------------------------------------------------------
+| The bit pattern for a default generated 16-bit BF16 floating-point NaN.
+*----------------------------------------------------------------------------*/
+#define defaultNaNF16UI 0x7FC0
+
+/*----------------------------------------------------------------------------
+| Converts the common NaN pointed to by 'aPtr' into a 16-bit floating-point
+| NaN, and returns the bit pattern of this value as an unsigned integer.
+*----------------------------------------------------------------------------*/
+#define softfloat_commonNaNToBF16UI( aPtr ) ((uint_fast16_t) defaultNaNBF16UI)
+
+/*----------------------------------------------------------------------------
 | The bit pattern for a default generated 32-bit floating-point NaN.
 *----------------------------------------------------------------------------*/
 #define defaultNaNF32UI 0x7FC00000
diff --git a/source/f32_to_bf16.c b/source/f32_to_bf16.c
new file mode 100644
index 0000000..82d9b16
--- /dev/null
+++ b/source/f32_to_bf16.c
@@ -0,0 +1,91 @@
+
+/*============================================================================
+
+This C source file is part of the SoftFloat IEEE Floating-Point Arithmetic
+Package, Release 3e, by John R. Hauser.
+
+Copyright 2011, 2012, 2013, 2014, 2015 The Regents of the University of
+California.  All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+ 1. Redistributions of source code must retain the above copyright notice,
+    this list of conditions, and the following disclaimer.
+
+ 2. Redistributions in binary form must reproduce the above copyright notice,
+    this list of conditions, and the following disclaimer in the documentation
+    and/or other materials provided with the distribution.
+
+ 3. Neither the name of the University nor the names of its contributors may
+    be used to endorse or promote products derived from this software without
+    specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS", AND ANY
+EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ARE
+DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
+DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+=============================================================================*/
+
+#include <stdbool.h>
+#include <stdint.h>
+#include "platform.h"
+#include "internals.h"
+#include "specialize.h"
+#include "softfloat.h"
+
+bfloat16_t f32_to_bf16( float32_t a )
+{
+    union ui32_f32 uA;
+    uint_fast32_t uiA;
+    bool sign;
+    int_fast16_t exp;
+    uint_fast32_t frac;
+    struct commonNaN commonNaN;
+    uint_fast16_t uiZ, frac16;
+    union ui16_bf16 uZ;
+
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    uA.f = a;
+    uiA = uA.ui;
+    sign = signF32UI( uiA );
+    exp  = expF32UI( uiA );
+    frac = fracF32UI( uiA );
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    // infinity or NaN cases
+    if ( exp == 0xFF ) {
+        if ( frac ) {
+            // NaN case
+            softfloat_f32UIToCommonNaN( uiA, &commonNaN );
+            uiZ = softfloat_commonNaNToBF16UI( &commonNaN );
+        } else {
+            // infinity case
+            uiZ = packToBF16UI( sign, 0xFF, 0 );
+        }
+        goto uiZ;
+    }
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    frac16 = frac>>16 | ((frac & 0xFFFF) != 0);
+    if ( ! (exp | frac16) ) {
+        uiZ = packToBF16UI( sign, 0, 0 );
+        goto uiZ;
+    }
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    return softfloat_roundPackToBF16( sign, exp, frac16 | 0x4000 );
+ uiZ:
+    uZ.ui = uiZ;
+    return uZ.f;
+
+}
+
diff --git a/source/f32_to_f16.c b/source/f32_to_f16.c
index a00b2e8..45005fe 100644
--- a/source/f32_to_f16.c
+++ b/source/f32_to_f16.c
@@ -72,6 +72,9 @@
     }
     /*------------------------------------------------------------------------
     *------------------------------------------------------------------------*/
+    // frac is a 24-bit significand, the bottom 9 bits LSB are extracted and OR-red
+    // into a sticky flag, the top 15 MSBs are extracted, the LSB of this top slice
+    // is OR-red with the sticky 
     frac16 = frac>>9 | ((frac & 0x1FF) != 0);
     if ( ! (exp | frac16) ) {
         uiZ = packToF16UI( sign, 0, 0 );
diff --git a/source/include/internals.h b/source/include/internals.h
index ccd7747..ea14a7d 100644
--- a/source/include/internals.h
+++ b/source/include/internals.h
@@ -102,6 +102,17 @@
 
 /*----------------------------------------------------------------------------
 *----------------------------------------------------------------------------*/
+#define signBF16UI( a ) ((bool) ((uint16_t) (a)>>15))
+#define expBF16UI( a ) ((int_fast8_t) ((a)>>7) & 0xFF)
+#define fracBF16UI( a ) ((a) & 0x07F)
+#define packToBF16UI( sign, exp, sig ) (((uint16_t) (sign)<<15) + ((uint16_t) (exp)<<7) + (sig))
+
+#define isNaNBF16UI( a ) (((~(a) & 0x7FC0) == 0) && ((a) & 0x07F))
+
+bfloat16_t softfloat_roundPackToBF16( bool, int_fast16_t, uint_fast16_t );
+
+/*----------------------------------------------------------------------------
+*----------------------------------------------------------------------------*/
 #define signF32UI( a ) ((bool) ((uint32_t) (a)>>31))
 #define expF32UI( a ) ((int_fast16_t) ((a)>>23) & 0xFF)
 #define fracF32UI( a ) ((a) & 0x007FFFFF)
diff --git a/source/s_roundPackToBF16.c b/source/s_roundPackToBF16.c
new file mode 100644
index 0000000..66e19be
--- /dev/null
+++ b/source/s_roundPackToBF16.c
@@ -0,0 +1,113 @@
+
+/*============================================================================
+
+This C source file is part of the SoftFloat IEEE Floating-Point Arithmetic
+Package, Release 3e, by John R. Hauser.
+
+Copyright 2011, 2012, 2013, 2014, 2015, 2017 The Regents of the University of
+California.  All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+ 1. Redistributions of source code must retain the above copyright notice,
+    this list of conditions, and the following disclaimer.
+
+ 2. Redistributions in binary form must reproduce the above copyright notice,
+    this list of conditions, and the following disclaimer in the documentation
+    and/or other materials provided with the distribution.
+
+ 3. Neither the name of the University nor the names of its contributors may
+    be used to endorse or promote products derived from this software without
+    specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS", AND ANY
+EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ARE
+DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY
+DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+=============================================================================*/
+
+#include <stdbool.h>
+#include <stdint.h>
+#include "platform.h"
+#include "internals.h"
+#include "softfloat.h"
+
+bfloat16_t
+ softfloat_roundPackToBF16( bool sign, int_fast16_t exp, uint_fast16_t sig )
+{
+    uint_fast8_t roundingMode;
+    bool roundNearEven;
+    uint_fast8_t roundIncrement, roundBits;
+    bool isTiny;
+    uint_fast16_t uiZ;
+    union ui16_bf16 uZ;
+
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    roundingMode = softfloat_roundingMode;
+    roundNearEven = (roundingMode == softfloat_round_near_even);
+    roundIncrement = 0x8;
+    if ( ! roundNearEven && (roundingMode != softfloat_round_near_maxMag) ) {
+        roundIncrement =
+            (roundingMode
+                 == (sign ? softfloat_round_min : softfloat_round_max))
+                ? 0xF
+                : 0;
+    }
+    roundBits = sig & 0xF;
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    if ( 0x1D <= (unsigned int) exp ) {
+        if ( exp < 0 ) {
+            /*----------------------------------------------------------------
+            *----------------------------------------------------------------*/
+            isTiny =
+                (softfloat_detectTininess == softfloat_tininess_beforeRounding)
+                    || (exp < -1) || (sig + roundIncrement < 0x8000);
+            sig = softfloat_shiftRightJam32( sig, -exp );
+            exp = 0;
+            roundBits = sig & 0xF;
+            if ( isTiny && roundBits ) {
+                softfloat_raiseFlags( softfloat_flag_underflow );
+            }
+        } else if ( (0x1D < exp) || (0x8000 <= sig + roundIncrement) ) {
+            /*----------------------------------------------------------------
+            *----------------------------------------------------------------*/
+            softfloat_raiseFlags(
+                softfloat_flag_overflow | softfloat_flag_inexact );
+            uiZ = packToF16UI( sign, 0x1F, 0 ) - ! roundIncrement;
+            goto uiZ;
+        }
+    }
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+    sig = (sig + roundIncrement)>>4;
+    if ( roundBits ) {
+        softfloat_exceptionFlags |= softfloat_flag_inexact;
+#ifdef SOFTFLOAT_ROUND_ODD
+        if ( roundingMode == softfloat_round_odd ) {
+            sig |= 1;
+            goto packReturn;
+        }
+#endif
+    }
+    sig &= ~(uint_fast16_t) (! (roundBits ^ 8) & roundNearEven);
+    if ( ! sig ) exp = 0;
+    /*------------------------------------------------------------------------
+    *------------------------------------------------------------------------*/
+ packReturn:
+    uiZ = packToF16UI( sign, exp, sig );
+ uiZ:
+    uZ.ui = uiZ;
+    return uZ.f;
+
+}
+