[Beignet] [PATCH] backend: add double support to log

rander rander.wang at intel.com
Fri Mar 24 01:36:11 UTC 2017


	get it from flibm and refine it to pass cft

Signed-off-by: rander <rander.wang at intel.com>
---
 backend/src/libocl/tmpl/ocl_math.tmpl.cl    | 119 ++++++++++++++++++++++++++++
 backend/src/libocl/tmpl/ocl_math.tmpl.h     |   4 +
 backend/src/libocl/tmpl/ocl_math_20.tmpl.cl |  95 ++++++++++++++++++++++
 backend/src/libocl/tmpl/ocl_math_20.tmpl.h  |   4 +
 4 files changed, 222 insertions(+)

diff --git a/backend/src/libocl/tmpl/ocl_math.tmpl.cl b/backend/src/libocl/tmpl/ocl_math.tmpl.cl
index 8c5fe8c..a3428c8 100644
--- a/backend/src/libocl/tmpl/ocl_math.tmpl.cl
+++ b/backend/src/libocl/tmpl/ocl_math.tmpl.cl
@@ -3901,6 +3901,31 @@ OVERLOADABLE half rootn(half x, int n) {
 
 
 //-----------------double -----------------------
+INLINE int  __HI(double x){
+    long x64 = as_long(x);
+    int high = convert_int((x64 >> 32) & 0xFFFFFFFF);
+    return high;
+};
+
+INLINE int  __LO(double x){
+    long x64 = as_long(x);
+    int low = convert_int(x64  & 0xFFFFFFFFUL);
+    return low;
+};
+
+INLINE void  __setHigh(double *x, int val){
+    long x64 = as_long(*x);
+    long high = x64  & 0x00000000FFFFFFFF;
+    high |= ((long)val << 32);
+    *x = as_double(high);
+};
+
+INLINE void  __setLow(double *x, int val){
+    long x64 = as_long(*x);
+    long low = x64  & 0xFFFFFFFF00000000;
+    low |= val;
+    *x = as_double(low);
+};
 
 OVERLOADABLE double ceil(double x)
 {
@@ -3984,3 +4009,97 @@ OVERLOADABLE double fract(double x, private double *p)
     return x -ret;
 }
 
+
+/* @(#)e_log.c 1.3 95/01/18 */
+/*
+ * ====================================================
+ * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
+ *
+ * Developed at SunSoft, a Sun Microsystems, Inc. business.
+ * Permission to use, copy, modify, and distribute this
+ * software is freely granted, provided that this notice
+ * is preserved.
+ * ====================================================
+ */
+OVERLOADABLE double log(double x)
+{
+	double ln2_hi	=  6.93147180369123816490e-01,	/* 3fe62e42 fee00000 */
+	ln2_lo	=  1.90821492927058770002e-10,	/* 3dea39ef 35793c76 */
+	two54	=  1.80143985094819840000e+16,	/* 43500000 00000000 */
+	Lg1 = 6.666666666666735130e-01,  /* 3FE55555 55555593 */
+	Lg2 = 3.999999999940941908e-01,  /* 3FD99999 9997FA04 */
+	Lg3 = 2.857142874366239149e-01,  /* 3FD24924 94229359 */
+	Lg4 = 2.222219843214978396e-01,  /* 3FCC71C5 1D8E78AF */
+	Lg5 = 1.818357216161805012e-01,  /* 3FC74664 96CB03DE */
+	Lg6 = 1.531383769920937332e-01,  /* 3FC39A09 D078C69F */
+	Lg7 = 1.479819860511658591e-01;  /* 3FC2F112 DF3E5244 */
+
+	double zero = 0;
+	double hfsq,f,s,z,R,w,t1,t2,dk;
+	int k,hx,i,j;
+	unsigned lx;
+
+	hx = __HI(x);		/* high word of x */
+	lx = __LO(x);		/* low  word of x */
+
+	k=0;
+	if (hx < 0x00100000)
+	{			/* x < 2**-1022  */
+		if (((hx&0x7fffffff)|lx)==0)
+		return -two54/zero;		/* log(+-0)=-inf */
+		if (hx<0)
+			return (x-x)/zero;	/* log(-#) = NaN */
+		k -= 54; x *= two54; /* subnormal number, scale up x */
+		hx = __HI(x);		/* high word of x */
+	}
+	if (hx >= 0x7ff00000) return x+x;
+	k += (hx>>20)-1023;
+	hx &= 0x000fffff;
+	i = (hx+0x95f64)&0x100000;
+	__setHigh(&x, (hx|(i^0x3ff00000)));	/* normalize x or x/2 */
+	k += (i>>20);
+	f = x-1.0;
+	if((0x000fffff&(2+hx))<3) {	/* |f| < 2**-20 */
+		if(f==zero)
+		{
+			if(k==0) return zero;
+			else
+			{
+				dk=(double)k;
+				return dk*ln2_hi+dk*ln2_lo;
+			}
+		}
+
+		R = f*f*(0.5-0.33333333333333333*f);
+		if(k==0)
+			return f-R;
+		else {dk=(double)k;
+			return dk*ln2_hi-((R-dk*ln2_lo)-f);}
+	}
+	s = f/(2.0+f);
+	dk = (double)k;
+	z = s*s;
+	i = hx-0x6147a;
+	w = z*z;
+	j = 0x6b851-hx;
+	t1= w*(Lg2+w*(Lg4+w*Lg6));
+	t2= z*(Lg1+w*(Lg3+w*(Lg5+w*Lg7)));
+	i |= j;
+	R = t2+t1;
+	if(i>0) {
+	hfsq=0.5*f*f;
+	if(k==0)
+		return f-(hfsq-s*(hfsq+R));
+	else
+		return dk*ln2_hi-((hfsq-(s*(hfsq+R)+dk*ln2_lo))-f);
+	}
+	else
+	{
+		if(k==0)
+			return f-s*(f-R);
+		else
+			return dk*ln2_hi-((s*(f-R)-dk*ln2_lo)-f);
+	}
+
+}
+
diff --git a/backend/src/libocl/tmpl/ocl_math.tmpl.h b/backend/src/libocl/tmpl/ocl_math.tmpl.h
index dc79c08..ce6b075 100644
--- a/backend/src/libocl/tmpl/ocl_math.tmpl.h
+++ b/backend/src/libocl/tmpl/ocl_math.tmpl.h
@@ -239,4 +239,8 @@ OVERLOADABLE double floor(double x);
 OVERLOADABLE double fract(double x, global double *p);
 OVERLOADABLE double fract(double x, local double *p);
 OVERLOADABLE double fract(double x, private double *p);
+OVERLOADABLE double log(double x);
+OVERLOADABLE double log2(double x);
+OVERLOADABLE double log10(double x);
+
 
diff --git a/backend/src/libocl/tmpl/ocl_math_20.tmpl.cl b/backend/src/libocl/tmpl/ocl_math_20.tmpl.cl
index 701d006..0f355b1 100644
--- a/backend/src/libocl/tmpl/ocl_math_20.tmpl.cl
+++ b/backend/src/libocl/tmpl/ocl_math_20.tmpl.cl
@@ -3886,3 +3886,98 @@ OVERLOADABLE double fract(double x, private double *p)
 }
 
 
+/* @(#)e_log.c 1.3 95/01/18 */
+/*
+ * ====================================================
+ * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
+ *
+ * Developed at SunSoft, a Sun Microsystems, Inc. business
+ * Permission to use, copy, modify, and distribute this
+ * software is freely granted, provided that this notice
+ * is preserved.
+ * ====================================================
+ */
+OVERLOADABLE double log(double x)
+{
+	double ln2_hi	=  6.93147180369123816490e-01,	/* 3fe62e42 fee00000 */
+	ln2_lo	=  1.90821492927058770002e-10,	/* 3dea39ef 35793c76 */
+	two54	=  1.80143985094819840000e+16,	/* 43500000 00000000 */
+	Lg1 = 6.666666666666735130e-01,  /* 3FE55555 55555593 */
+	Lg2 = 3.999999999940941908e-01,  /* 3FD99999 9997FA04 */
+	Lg3 = 2.857142874366239149e-01,  /* 3FD24924 94229359 */
+	Lg4 = 2.222219843214978396e-01,  /* 3FCC71C5 1D8E78AF */
+	Lg5 = 1.818357216161805012e-01,  /* 3FC74664 96CB03DE */
+	Lg6 = 1.531383769920937332e-01,  /* 3FC39A09 D078C69F */
+	Lg7 = 1.479819860511658591e-01;  /* 3FC2F112 DF3E5244 */
+
+	double zero = 0;
+	double hfsq,f,s,z,R,w,t1,t2,dk;
+	int k,hx,i,j;
+	unsigned lx;
+
+	hx = __HI(x);		/* high word of x */
+	lx = __LO(x);		/* low	word of x */
+
+	k=0;
+	if (hx < 0x00100000)
+	{			/* x < 2**-1022  */
+		if (((hx&0x7fffffff)|lx)==0)
+		return -two54/zero; 	/* log(+-0)=-inf */
+		if (hx<0)
+			return (x-x)/zero;	/* log(-#) = NaN */
+		k -= 54; x *= two54; /* subnormal number, scale up x */
+		hx = __HI(x);		/* high word of x */
+	}
+	if (hx >= 0x7ff00000) return x+x;
+	k += (hx>>20)-1023;
+	hx &= 0x000fffff;
+	i = (hx+0x95f64)&0x100000;
+	__setHigh(&x, (hx|(i^0x3ff00000))); /* normalize x or x/2 */
+	k += (i>>20);
+	f = x-1.0;
+	if((0x000fffff&(2+hx))<3) { /* |f| < 2**-20 */
+		if(f==zero)
+		{
+			if(k==0) return zero;
+			else
+			{
+				dk=(double)k;
+				return dk*ln2_hi+dk*ln2_lo;
+			}
+		}
+
+		R = f*f*(0.5-0.33333333333333333*f);
+		if(k==0)
+			return f-R;
+		else {dk=(double)k;
+			return dk*ln2_hi-((R-dk*ln2_lo)-f);}
+	}
+	s = f/(2.0+f);
+	dk = (double)k;
+	z = s*s;
+	i = hx-0x6147a;
+	w = z*z;
+	j = 0x6b851-hx;
+	t1= w*(Lg2+w*(Lg4+w*Lg6));
+	t2= z*(Lg1+w*(Lg3+w*(Lg5+w*Lg7)));
+	i |= j;
+	R = t2+t1;
+	if(i>0) {
+	hfsq=0.5*f*f;
+	if(k==0)
+		return f-(hfsq-s*(hfsq+R));
+	else
+		return dk*ln2_hi-((hfsq-(s*(hfsq+R)+dk*ln2_lo))-f);
+	}
+	else
+	{
+		if(k==0)
+			return f-s*(f-R);
+		else
+			return dk*ln2_hi-((s*(f-R)-dk*ln2_lo)-f);
+	}
+
+}
+
+
+
diff --git a/backend/src/libocl/tmpl/ocl_math_20.tmpl.h b/backend/src/libocl/tmpl/ocl_math_20.tmpl.h
index 52a6b23..f65d3e0 100644
--- a/backend/src/libocl/tmpl/ocl_math_20.tmpl.h
+++ b/backend/src/libocl/tmpl/ocl_math_20.tmpl.h
@@ -216,4 +216,8 @@ OVERLOADABLE double floor(double x);
 OVERLOADABLE double fract(double x, global double *p);
 OVERLOADABLE double fract(double x, local double *p);
 OVERLOADABLE double fract(double x, private double *p);
+OVERLOADABLE double log(double x);
+OVERLOADABLE double log2(double x);
+OVERLOADABLE double log10(double x);
+
 
-- 
2.7.4



More information about the Beignet mailing list