diff --git a/src/c/numbers/number_compare.d b/src/c/numbers/number_compare.d
index a1e17b810..c8e7e5e5e 100644
--- a/src/c/numbers/number_compare.d
+++ b/src/c/numbers/number_compare.d
@@ -79,6 +79,10 @@ ecl_number_compare(cl_object x, cl_object y)
     case t_doublefloat:
 #ifdef ECL_LONG_FLOAT
     case t_longfloat:
+#endif
+#ifdef ECL_IEEE_FP
+      if (si_float_infinity_p(y))
+        return(ecl_number_compare(ecl_make_fixnum(0), y));
 #endif
       y = cl_rational(y);
       goto BEGIN;
@@ -101,6 +105,10 @@ ecl_number_compare(cl_object x, cl_object y)
     case t_doublefloat:
 #ifdef ECL_LONG_FLOAT
     case t_longfloat:
+#endif
+#ifdef ECL_IEEE_FP
+      if (si_float_infinity_p(y))
+        return(ecl_number_compare(ecl_make_fixnum(0), y));
 #endif
       y = cl_rational(y);
       goto BEGIN;
@@ -118,6 +126,10 @@ ecl_number_compare(cl_object x, cl_object y)
       return -double_fix_compare(ecl_fixnum(y), dx);
     case t_bignum:
     case t_ratio:
+#ifdef ECL_IEEE_FP
+      if (si_float_infinity_p(x))
+        return(ecl_number_compare(x, ecl_make_fixnum(0)));
+#endif
       x = cl_rational(x);
       goto BEGIN;
     case t_singlefloat: