[Libclc-dev] [PATCH 4/4] math: Implement tgamma

[Aaron Watry via Libclc-dev]

Signed-off-by: Aaron Watry <awatry at gmail.com>
---
 generic/include/clc/clc.h           |  1 +
 generic/include/clc/math/tgamma.h   |  3 ++
 generic/include/clc/math/tgamma.inc |  1 +
 generic/lib/SOURCES                 |  1 +
 generic/lib/math/tgamma.cl          | 71 +++++++++++++++++++++++++++++++++++++
 5 files changed, 77 insertions(+)
 create mode 100644 generic/include/clc/math/tgamma.h
 create mode 100644 generic/include/clc/math/tgamma.inc
 create mode 100644 generic/lib/math/tgamma.cl

diff --git a/generic/include/clc/clc.h b/generic/include/clc/clc.h
index fecbd59..5bdb09f 100644
--- a/generic/include/clc/clc.h
+++ b/generic/include/clc/clc.h
@@ -88,6 +88,7 @@
 #include <clc/math/sqrt.h>
 #include <clc/math/tan.h>
 #include <clc/math/tanh.h>
+#include <clc/math/tgamma.h>
 #include <clc/math/trunc.h>
 #include <clc/math/native_cos.h>
 #include <clc/math/native_divide.h>
diff --git a/generic/include/clc/math/tgamma.h b/generic/include/clc/math/tgamma.h
new file mode 100644
index 0000000..a041d32
--- /dev/null
+++ b/generic/include/clc/math/tgamma.h
@@ -0,0 +1,3 @@
+#define __CLC_BODY <clc/math/tgamma.inc>
+#include <clc/math/gentype.inc>
+#undef __CLC_BODY
diff --git a/generic/include/clc/math/tgamma.inc b/generic/include/clc/math/tgamma.inc
new file mode 100644
index 0000000..9f67fe4
--- /dev/null
+++ b/generic/include/clc/math/tgamma.inc
@@ -0,0 +1 @@
+_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE tgamma(__CLC_GENTYPE a);
diff --git a/generic/lib/SOURCES b/generic/lib/SOURCES
index bcd0991..423a50b 100644
--- a/generic/lib/SOURCES
+++ b/generic/lib/SOURCES
@@ -119,6 +119,7 @@ math/clc_sqrt.cl
 math/sqrt.cl
 math/tan.cl
 math/tanh.cl
+math/tgamma.cl
 relational/all.cl
 relational/any.cl
 relational/bitselect.cl
diff --git a/generic/lib/math/tgamma.cl b/generic/lib/math/tgamma.cl
new file mode 100644
index 0000000..29c069f
--- /dev/null
+++ b/generic/lib/math/tgamma.cl
@@ -0,0 +1,71 @@
+/*
+ * Copyright (c) 2016 Aaron Watry
+ * Copyright (c) 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <clc/clc.h>
+
+#include "math.h"
+#include "../clcmacro.h"
+
+_CLC_OVERLOAD _CLC_DEF float tgamma(float x) {
+    const float pi = 3.1415926535897932384626433832795f;
+    float ax = fabs(x);
+    float lg = lgamma(ax);
+    float g = exp(lg);
+
+    if (x < 0.0f) {
+        float z = sinpi(x);
+        g = g * ax * z;
+        g = pi / g;
+        g = g == 0 ? as_float(PINFBITPATT_SP32) : g;
+        g = z == 0 ? as_float(QNANBITPATT_SP32) : g;
+    }
+
+    return g;
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, tgamma, float);
+
+#ifdef cl_khr_fp64
+
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+
+_CLC_OVERLOAD _CLC_DEF double tgamma(double x) {
+    const double pi = 3.1415926535897932384626433832795;
+    double ax = fabs(x);
+    double lg = lgamma(ax);
+    double g = exp(lg);
+
+    if (x < 0.0) {
+        double z = sinpi(x);
+        g = g * ax * z;
+        g = pi / g;
+        g = g == 0 ? as_double(PINFBITPATT_DP64) : g;
+        g = z == 0 ? as_double(QNANBITPATT_DP64) : g;
+    }
+
+    return g;
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, tgamma, double);
+
+#endif
-- 
2.7.4