* Tried to do something with cucul_powf. It currently sucks.

cucul/math.c

@@ -70,26 +70,36 @@ unsigned int cucul_sqrt(unsigned int a)
     return 2 * cucul_sqrt(a / 4);
 }
 
-
 /**
  * \brief powf substitute (x^y)
  * \param x The value to be raised
- * \param y The power to raise x of.
+ * \param y The power to raise x to (only works with integers)
  * \return \p x raised to the power of \p y
  */
-
 float cucul_powf(float x, float y)
 {
-    int i=((int)y);
-    float r=x;
+#ifdef __i386__
+    /* FIXME: this can be optimised by directly calling fyl2x for x and y */
+    register double logx;
+    register long double v, e;
 
-    if(((int)y)==1 || ((int)x)==1)
-        return x;
+    asm volatile("fldln2; fxch; fyl2x"
+                 : "=t" (logx) : "0" (x) : "st(1)");
 
-    i--;
-    while(i--)
-    {
-        r*=x;
-    }
-    return r;
+    asm volatile("fldl2e\n\t"
+                 "fmul %%st(1)\n\t"
+                 "fst %%st(1)\n\t"
+                 "frndint\n\t"
+                 "fxch\n\t"
+                 "fsub %%st(1)\n\t"
+                 "f2xm1\n\t"
+                 : "=t" (v), "=u" (e) : "0" (y * logx));
+    v += 1.0;
+    asm volatile("fscale"
+                 : "=t" (v) : "0" (v), "u" (e));
+
+    return v;
+#else
+    return x; /* HAHAHA VIEUX PORC */
+#endif
 }