[Libclc-dev] [PATCH 2/2] amdgcn/fmax: fcanonicalize operands

[Jan Vesely via Libclc-dev]

v_max instruction needs canonicalized operands.
Passes CTS on carrizo
Fixes maxmag CTS on carrizo

Signed-off-by: Jan Vesely <jan.vesely at rutgers.edu>
---
 amdgcn/lib/SOURCES      |  1 +
 amdgcn/lib/math/fmax.cl | 31 +++++++++++++++++++++++++++++++
 2 files changed, 32 insertions(+)
 create mode 100644 amdgcn/lib/math/fmax.cl

diff --git a/amdgcn/lib/SOURCES b/amdgcn/lib/SOURCES
index 95cc276..1f25f36 100644
--- a/amdgcn/lib/SOURCES
+++ b/amdgcn/lib/SOURCES
@@ -1,4 +1,5 @@
 cl_khr_int64_extended_atomics/minmax_helpers.ll
+math/fmax.cl
 math/fmin.cl
 math/ldexp.cl
 mem_fence/fence.cl
diff --git a/amdgcn/lib/math/fmax.cl b/amdgcn/lib/math/fmax.cl
new file mode 100644
index 0000000..c1e6cb9
--- /dev/null
+++ b/amdgcn/lib/math/fmax.cl
@@ -0,0 +1,31 @@
+#include <clc/clc.h>
+
+#include "../../../generic/lib/clcmacro.h"
+
+_CLC_DEF _CLC_OVERLOAD float fmax(float x, float y)
+{
+   /* fcanonicalize removes sNaNs and flushes denormals if not enabled.
+    * Otherwise fmax instruction flushes the values for comparison,
+    * but outputs original denormal */
+   x = __builtin_canonicalizef(x);
+   y = __builtin_canonicalizef(y);
+   return __builtin_fmaxf(x, y);
+}
+_CLC_BINARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, fmax, float, float)
+
+#ifdef cl_khr_fp64
+
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+
+_CLC_DEF _CLC_OVERLOAD double fmax(double x, double y)
+{
+   x = __builtin_canonicalize(x);
+   y = __builtin_canonicalize(y);
+   return __builtin_fmax(x, y);
+}
+_CLC_BINARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, fmax, double, double)
+
+#endif
+
+#define __CLC_BODY <../../../generic/lib/math/fmax.inc>
+#include <clc/math/gentype.inc>
-- 
2.14.3