[libclc] r312402 - Add halfN types and enable fp16 when generating builtin declarations
Author: awatry Date: Fri Sep 1 19:23:16 2017 New Revision: 312402 URL: http://llvm.org/viewvc/llvm-project?rev=312402&view=rev Log: Add halfN types and enable fp16 when generating builtin declarations Uses the same mechanism to enable fp16 as we use for fp64 when processing clc.h Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/include/clc/clctypes.h Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=312402&r1=312401&r2=312402&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Fri Sep 1 19:23:16 2017 @@ -8,6 +8,10 @@ #pragma OPENCL EXTENSION cl_khr_fp64 : enable #endif +#ifdef cl_khr_fp16 +#pragma OPENCL EXTENSION cl_khr_fp16 : enable +#endif + /* Function Attributes */ #include Modified: libclc/trunk/generic/include/clc/clctypes.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clctypes.h?rev=312402&r1=312401&r2=312402&view=diff == --- libclc/trunk/generic/include/clc/clctypes.h (original) +++ libclc/trunk/generic/include/clc/clctypes.h Fri Sep 1 19:23:16 2017 @@ -85,3 +85,11 @@ typedef __attribute__((ext_vector_type(4 typedef __attribute__((ext_vector_type(8))) double double8; typedef __attribute__((ext_vector_type(16))) double double16; #endif + +#ifdef cl_khr_fp16 +typedef __attribute__((ext_vector_type(2))) half half2; +typedef __attribute__((ext_vector_type(3))) half half3; +typedef __attribute__((ext_vector_type(4))) half half4; +typedef __attribute__((ext_vector_type(8))) half half8; +typedef __attribute__((ext_vector_type(16))) half half16; +#endif ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r312403 - relational: Implement shuffle builtin
Author: awatry Date: Fri Sep 1 19:23:26 2017 New Revision: 312403 URL: http://llvm.org/viewvc/llvm-project?rev=312403&view=rev Log: relational: Implement shuffle builtin This was added in CL 1.1 Tested with a Radeon HD 7850 (Pitcairn) using the CL CTS via: test_conformance/relationals/test_relationals shuffle_built_in v2: Add half-precision support to shuffle when available. Move to misc/ and add section 6.12.12 to clc.h Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely Added: libclc/trunk/generic/include/clc/misc/ libclc/trunk/generic/include/clc/misc/shuffle.h libclc/trunk/generic/lib/misc/ libclc/trunk/generic/lib/misc/shuffle.cl Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/lib/SOURCES Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=312403&r1=312402&r2=312403&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Fri Sep 1 19:23:26 2017 @@ -237,6 +237,9 @@ /* 6.11.13 Image Read and Write Functions */ +/* 6.12.12 Miscellaneous Vector Functions */ +#include + #include #include Added: libclc/trunk/generic/include/clc/misc/shuffle.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/misc/shuffle.h?rev=312403&view=auto == --- libclc/trunk/generic/include/clc/misc/shuffle.h (added) +++ libclc/trunk/generic/include/clc/misc/shuffle.h Fri Sep 1 19:23:26 2017 @@ -0,0 +1,47 @@ +//===-- generic/include/clc/misc/shuffle.h --===// +// +// The LLVM Compiler Infrastructure +// +// This file is dual licensed under both the University of Illinois Open Source +// License and the MIT license. See LICENSE.TXT for details. +// +//===--===// + +#define _CLC_SHUFFLE_DECL(TYPE, MASKTYPE, RETTYPE) \ + _CLC_OVERLOAD _CLC_DECL RETTYPE shuffle(TYPE x, MASKTYPE mask); + +//Return type is same base type as the input type, with the same vector size as the mask. +//Elements in the mask must be the same size (number of bits) as the input value. +//E.g. char8 ret = shuffle(char2 x, uchar8 mask); + +#define _CLC_VECTOR_SHUFFLE_MASKSIZE(INBASE, INTYPE, MASKTYPE) \ + _CLC_SHUFFLE_DECL(INTYPE, MASKTYPE##2, INBASE##2) \ + _CLC_SHUFFLE_DECL(INTYPE, MASKTYPE##4, INBASE##4) \ + _CLC_SHUFFLE_DECL(INTYPE, MASKTYPE##8, INBASE##8) \ + _CLC_SHUFFLE_DECL(INTYPE, MASKTYPE##16, INBASE##16) \ + +#define _CLC_VECTOR_SHUFFLE_INSIZE(TYPE, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE_MASKSIZE(TYPE, TYPE##2, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE_MASKSIZE(TYPE, TYPE##4, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE_MASKSIZE(TYPE, TYPE##8, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE_MASKSIZE(TYPE, TYPE##16, MASKTYPE) \ + +_CLC_VECTOR_SHUFFLE_INSIZE(char, uchar) +_CLC_VECTOR_SHUFFLE_INSIZE(short, ushort) +_CLC_VECTOR_SHUFFLE_INSIZE(int, uint) +_CLC_VECTOR_SHUFFLE_INSIZE(long, ulong) +_CLC_VECTOR_SHUFFLE_INSIZE(uchar, uchar) +_CLC_VECTOR_SHUFFLE_INSIZE(ushort, ushort) +_CLC_VECTOR_SHUFFLE_INSIZE(uint, uint) +_CLC_VECTOR_SHUFFLE_INSIZE(ulong, ulong) +_CLC_VECTOR_SHUFFLE_INSIZE(float, uint) +#ifdef cl_khr_fp64 +_CLC_VECTOR_SHUFFLE_INSIZE(double, ulong) +#endif +#ifdef cl_khr_fp16 +_CLC_VECTOR_SHUFFLE_INSIZE(half, ushort) +#endif + +#undef _CLC_SHUFFLE_DECL +#undef _CLC_VECTOR_SHUFFLE_MASKSIZE +#undef _CLC_VECTOR_SHUFFLE_INSIZE Modified: libclc/trunk/generic/lib/SOURCES URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=312403&r1=312402&r2=312403&view=diff == --- libclc/trunk/generic/lib/SOURCES (original) +++ libclc/trunk/generic/lib/SOURCES Fri Sep 1 19:23:26 2017 @@ -123,6 +123,7 @@ math/sqrt.cl math/tan.cl math/tanh.cl math/tgamma.cl +misc/shuffle.cl relational/all.cl relational/any.cl relational/bitselect.cl @@ -139,6 +140,7 @@ relational/isnormal.cl relational/isnotequal.cl relational/isordered.cl relational/isunordered.cl +relational/shuffle2.cl relational/signbit.cl shared/clamp.cl shared/max.cl Added: libclc/trunk/generic/lib/misc/shuffle.cl URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/misc/shuffle.cl?rev=312403&view=auto == --- libclc/trunk/generic/lib/misc/shuffle.cl (added) +++ libclc/trunk/generic/lib/misc/shuffle.cl Fri Sep 1 19:23:26 2017 @@ -0,0 +1,157 @@ +//===-- generic/lib/misc/shuffle.cl --===// +// +// The LLVM Compiler Infrastructure +// +// This file is dual licensed under both the University of Illinois Open Source +// License and the MIT license. See LICENSE.TXT for details. +// +//===--
[libclc] r312404 - relational: Implement shuffle2 builtin
Author: awatry Date: Fri Sep 1 19:23:28 2017 New Revision: 312404 URL: http://llvm.org/viewvc/llvm-project?rev=312404&view=rev Log: relational: Implement shuffle2 builtin This was added in CL 1.1 Tested with a Radeon HD 7850 (Pitcairn) using the CL CTS via: test_conformance/relationals/test_relationals shuffle_built_in_dual_input v2: Add half support to shuffle2 Move shuffle2 to misc/ Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely Added: libclc/trunk/generic/include/clc/misc/shuffle2.h libclc/trunk/generic/lib/misc/shuffle2.cl Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/lib/SOURCES Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=312404&r1=312403&r2=312404&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Fri Sep 1 19:23:28 2017 @@ -239,6 +239,7 @@ /* 6.12.12 Miscellaneous Vector Functions */ #include +#include #include #include Added: libclc/trunk/generic/include/clc/misc/shuffle2.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/misc/shuffle2.h?rev=312404&view=auto == --- libclc/trunk/generic/include/clc/misc/shuffle2.h (added) +++ libclc/trunk/generic/include/clc/misc/shuffle2.h Fri Sep 1 19:23:28 2017 @@ -0,0 +1,47 @@ +//===-- generic/include/clc/misc/shuffle2.h --===// +// +// The LLVM Compiler Infrastructure +// +// This file is dual licensed under both the University of Illinois Open Source +// License and the MIT license. See LICENSE.TXT for details. +// +//===--===// + +#define _CLC_SHUFFLE2_DECL(TYPE, MASKTYPE, RETTYPE) \ + _CLC_OVERLOAD _CLC_DECL RETTYPE shuffle2(TYPE x, TYPE y, MASKTYPE mask); + +//Return type is same base type as the input type, with the same vector size as the mask. +//Elements in the mask must be the same size (number of bits) as the input value. +//E.g. char8 ret = shuffle2(char2 x, char2 y, uchar8 mask); + +#define _CLC_VECTOR_SHUFFLE2_MASKSIZE(INBASE, INTYPE, MASKTYPE) \ + _CLC_SHUFFLE2_DECL(INTYPE, MASKTYPE##2, INBASE##2) \ + _CLC_SHUFFLE2_DECL(INTYPE, MASKTYPE##4, INBASE##4) \ + _CLC_SHUFFLE2_DECL(INTYPE, MASKTYPE##8, INBASE##8) \ + _CLC_SHUFFLE2_DECL(INTYPE, MASKTYPE##16, INBASE##16) \ + +#define _CLC_VECTOR_SHUFFLE2_INSIZE(TYPE, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE2_MASKSIZE(TYPE, TYPE##2, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE2_MASKSIZE(TYPE, TYPE##4, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE2_MASKSIZE(TYPE, TYPE##8, MASKTYPE) \ + _CLC_VECTOR_SHUFFLE2_MASKSIZE(TYPE, TYPE##16, MASKTYPE) \ + +_CLC_VECTOR_SHUFFLE2_INSIZE(char, uchar) +_CLC_VECTOR_SHUFFLE2_INSIZE(short, ushort) +_CLC_VECTOR_SHUFFLE2_INSIZE(int, uint) +_CLC_VECTOR_SHUFFLE2_INSIZE(long, ulong) +_CLC_VECTOR_SHUFFLE2_INSIZE(uchar, uchar) +_CLC_VECTOR_SHUFFLE2_INSIZE(ushort, ushort) +_CLC_VECTOR_SHUFFLE2_INSIZE(uint, uint) +_CLC_VECTOR_SHUFFLE2_INSIZE(ulong, ulong) +_CLC_VECTOR_SHUFFLE2_INSIZE(float, uint) +#ifdef cl_khr_fp64 +_CLC_VECTOR_SHUFFLE2_INSIZE(double, ulong) +#endif +#ifdef cl_khr_fp16 +_CLC_VECTOR_SHUFFLE2_INSIZE(half, ushort) +#endif + +#undef _CLC_SHUFFLE_DECL +#undef _CLC_VECTOR_SHUFFLE2_MASKSIZE +#undef _CLC_VECTOR_SHUFFLE2_INSIZE Modified: libclc/trunk/generic/lib/SOURCES URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=312404&r1=312403&r2=312404&view=diff == --- libclc/trunk/generic/lib/SOURCES (original) +++ libclc/trunk/generic/lib/SOURCES Fri Sep 1 19:23:28 2017 @@ -124,6 +124,7 @@ math/tan.cl math/tanh.cl math/tgamma.cl misc/shuffle.cl +misc/shuffle2.cl relational/all.cl relational/any.cl relational/bitselect.cl @@ -140,7 +141,6 @@ relational/isnormal.cl relational/isnotequal.cl relational/isordered.cl relational/isunordered.cl -relational/shuffle2.cl relational/signbit.cl shared/clamp.cl shared/max.cl Added: libclc/trunk/generic/lib/misc/shuffle2.cl URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/misc/shuffle2.cl?rev=312404&view=auto == --- libclc/trunk/generic/lib/misc/shuffle2.cl (added) +++ libclc/trunk/generic/lib/misc/shuffle2.cl Fri Sep 1 19:23:28 2017 @@ -0,0 +1,160 @@ +//===-- generic/lib/misc/shuffle2.cl --===// +// +// The LLVM Compiler Infrastructure +// +// This file is dual licensed under both the University of Illinois Open Source +// License and the MIT license. See LICENSE.TXT for details. +// +//===--===// + +#include + +#define _CLC_ELEMENT_
[libclc] r312854 - integer: Add popcount implementation using ctpop intrinsic
Author: awatry Date: Fri Sep 8 19:23:54 2017 New Revision: 312854 URL: http://llvm.org/viewvc/llvm-project?rev=312854&view=rev Log: integer: Add popcount implementation using ctpop intrinsic Also copy/modify the unary_intrin.inc from math/ to make the intrinsic declaration somewhat reusable. Passes CL CTS integer_ops/test_integer_ops popcount tests for CL 1.2 Tested-by on GCN 1.0 (Pitcairn) Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely Added: libclc/trunk/generic/include/clc/integer/popcount.h libclc/trunk/generic/include/clc/integer/unary_intrin.inc Modified: libclc/trunk/generic/include/clc/clc.h Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=312854&r1=312853&r2=312854&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Fri Sep 8 19:23:54 2017 @@ -126,6 +126,7 @@ #include #include #include +#include #include #include #include Added: libclc/trunk/generic/include/clc/integer/popcount.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/integer/popcount.h?rev=312854&view=auto == --- libclc/trunk/generic/include/clc/integer/popcount.h (added) +++ libclc/trunk/generic/include/clc/integer/popcount.h Fri Sep 8 19:23:54 2017 @@ -0,0 +1,6 @@ +#undef popcount +#define popcount __clc_popcount + +#define __CLC_FUNCTION __clc_popcount +#define __CLC_INTRINSIC "llvm.ctpop" +#include Added: libclc/trunk/generic/include/clc/integer/unary_intrin.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/integer/unary_intrin.inc?rev=312854&view=auto == --- libclc/trunk/generic/include/clc/integer/unary_intrin.inc (added) +++ libclc/trunk/generic/include/clc/integer/unary_intrin.inc Fri Sep 8 19:23:54 2017 @@ -0,0 +1,20 @@ +#define __CLC_INTRINSIC_DEF(SCALAR_TYPE, BIT_SIZE) \ +_CLC_OVERLOAD SCALAR_TYPE __CLC_FUNCTION(SCALAR_TYPE x) __asm(__CLC_INTRINSIC ".i" BIT_SIZE); \ +_CLC_OVERLOAD SCALAR_TYPE##2 __CLC_FUNCTION(SCALAR_TYPE##2 x) __asm(__CLC_INTRINSIC ".v2i" BIT_SIZE); \ +_CLC_OVERLOAD SCALAR_TYPE##3 __CLC_FUNCTION(SCALAR_TYPE##3 x) __asm(__CLC_INTRINSIC ".v3i" BIT_SIZE); \ +_CLC_OVERLOAD SCALAR_TYPE##4 __CLC_FUNCTION(SCALAR_TYPE##4 x) __asm(__CLC_INTRINSIC ".v4i" BIT_SIZE); \ +_CLC_OVERLOAD SCALAR_TYPE##8 __CLC_FUNCTION(SCALAR_TYPE##8 x) __asm(__CLC_INTRINSIC ".v8i" BIT_SIZE); \ +_CLC_OVERLOAD SCALAR_TYPE##16 __CLC_FUNCTION(SCALAR_TYPE##16 x) __asm(__CLC_INTRINSIC ".v16i" BIT_SIZE); + +__CLC_INTRINSIC_DEF(char, "8") +__CLC_INTRINSIC_DEF(uchar, "8") +__CLC_INTRINSIC_DEF(short, "16") +__CLC_INTRINSIC_DEF(ushort, "16") +__CLC_INTRINSIC_DEF(int, "32") +__CLC_INTRINSIC_DEF(uint, "32") +__CLC_INTRINSIC_DEF(long, "64") +__CLC_INTRINSIC_DEF(ulong, "64") + +#undef __CLC_FUNCTION +#undef __CLC_INTRINSIC +#undef __CLC_INTRINSIC_DEF ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r313107 - Add native_recip(x) as ((1)/(x))
Author: awatry Date: Tue Sep 12 18:40:25 2017 New Revision: 313107 URL: http://llvm.org/viewvc/llvm-project?rev=313107&view=rev Log: Add native_recip(x) as ((1)/(x)) Signed-off-by: Aaron Watry Acked-by: Jan Vesely Added: libclc/trunk/generic/include/clc/math/native_recip.h Modified: libclc/trunk/generic/include/clc/clc.h Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=313107&r1=313106&r2=313107&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Tue Sep 12 18:40:25 2017 @@ -106,6 +106,7 @@ #include #include #include +#include #include #include #include Added: libclc/trunk/generic/include/clc/math/native_recip.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/native_recip.h?rev=313107&view=auto == --- libclc/trunk/generic/include/clc/math/native_recip.h (added) +++ libclc/trunk/generic/include/clc/math/native_recip.h Tue Sep 12 18:40:25 2017 @@ -0,0 +1 @@ +#define native_recip(x) ((1) / (x)) ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r281565 - math: Implement lgamma
Author: awatry
Date: Wed Sep 14 19:17:31 2016
New Revision: 281565
URL: http://llvm.org/viewvc/llvm-project?rev=281565&view=rev
Log:
math: Implement lgamma
Just use lgamma_r and ignore the value returned in the second argument
Signed-off-by: Aaron Watry
Reviewed-by: Tom Stellard
Added:
libclc/trunk/generic/include/clc/math/lgamma.h
libclc/trunk/generic/include/clc/math/lgamma.inc
libclc/trunk/generic/lib/math/lgamma.cl
Modified:
libclc/trunk/generic/include/clc/clc.h
libclc/trunk/generic/lib/SOURCES
Modified: libclc/trunk/generic/include/clc/clc.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=281565&r1=281564&r2=281565&view=diff
==
--- libclc/trunk/generic/include/clc/clc.h (original)
+++ libclc/trunk/generic/include/clc/clc.h Wed Sep 14 19:17:31 2016
@@ -69,6 +69,7 @@
#include
#include
#include
+#include
#include
#include
#include
Added: libclc/trunk/generic/include/clc/math/lgamma.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/lgamma.h?rev=281565&view=auto
==
--- libclc/trunk/generic/include/clc/math/lgamma.h (added)
+++ libclc/trunk/generic/include/clc/math/lgamma.h Wed Sep 14 19:17:31 2016
@@ -0,0 +1,2 @@
+#define __CLC_BODY
+#include
Added: libclc/trunk/generic/include/clc/math/lgamma.inc
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/lgamma.inc?rev=281565&view=auto
==
--- libclc/trunk/generic/include/clc/math/lgamma.inc (added)
+++ libclc/trunk/generic/include/clc/math/lgamma.inc Wed Sep 14 19:17:31 2016
@@ -0,0 +1 @@
+_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE lgamma(__CLC_GENTYPE a);
Modified: libclc/trunk/generic/lib/SOURCES
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=281565&r1=281564&r2=281565&view=diff
==
--- libclc/trunk/generic/lib/SOURCES (original)
+++ libclc/trunk/generic/lib/SOURCES Wed Sep 14 19:17:31 2016
@@ -97,6 +97,7 @@ math/hypot.cl
math/ilogb.cl
math/clc_ldexp.cl
math/ldexp.cl
+math/lgamma.cl
math/lgamma_r.cl
math/log.cl
math/log10.cl
Added: libclc/trunk/generic/lib/math/lgamma.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/lgamma.cl?rev=281565&view=auto
==
--- libclc/trunk/generic/lib/math/lgamma.cl (added)
+++ libclc/trunk/generic/lib/math/lgamma.cl Wed Sep 14 19:17:31 2016
@@ -0,0 +1,44 @@
+/*
+ * 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
+#include "../clcmacro.h"
+
+_CLC_OVERLOAD _CLC_DEF float lgamma(float x) {
+int s;
+return lgamma_r(x, &s);
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, lgamma, float)
+
+#ifdef cl_khr_fp64
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+
+_CLC_OVERLOAD _CLC_DEF double lgamma(double x) {
+int s;
+return lgamma_r(x, &s);
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, lgamma, double)
+
+#endif
\ No newline at end of file
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r281564 - math: Implement lgamma_r
Author: awatry Date: Wed Sep 14 19:17:28 2016 New Revision: 281564 URL: http://llvm.org/viewvc/llvm-project?rev=281564&view=rev Log: math: Implement lgamma_r Ported from the amd-builtins branch, which is itself based on the Sun Microsystems implementation. Signed-off-by: Aaron Watry Reviewed-by: Tom Stellard Added: libclc/trunk/generic/include/clc/math/lgamma_r.h libclc/trunk/generic/include/clc/math/lgamma_r.inc libclc/trunk/generic/lib/math/lgamma_r.cl libclc/trunk/generic/lib/math/lgamma_r.inc Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/lib/SOURCES Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=281564&r1=281563&r2=281564&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Wed Sep 14 19:17:28 2016 @@ -69,6 +69,7 @@ #include #include #include +#include #include #include #include Added: libclc/trunk/generic/include/clc/math/lgamma_r.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/lgamma_r.h?rev=281564&view=auto == --- libclc/trunk/generic/include/clc/math/lgamma_r.h (added) +++ libclc/trunk/generic/include/clc/math/lgamma_r.h Wed Sep 14 19:17:28 2016 @@ -0,0 +1,2 @@ +#define __CLC_BODY +#include Added: libclc/trunk/generic/include/clc/math/lgamma_r.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/lgamma_r.inc?rev=281564&view=auto == --- libclc/trunk/generic/include/clc/math/lgamma_r.inc (added) +++ libclc/trunk/generic/include/clc/math/lgamma_r.inc Wed Sep 14 19:17:28 2016 @@ -0,0 +1,3 @@ +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE lgamma_r(__CLC_GENTYPE x, global __CLC_INTN *iptr); +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE lgamma_r(__CLC_GENTYPE x, local __CLC_INTN *iptr); +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE lgamma_r(__CLC_GENTYPE x, private __CLC_INTN *iptr); Modified: libclc/trunk/generic/lib/SOURCES URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=281564&r1=281563&r2=281564&view=diff == --- libclc/trunk/generic/lib/SOURCES (original) +++ libclc/trunk/generic/lib/SOURCES Wed Sep 14 19:17:28 2016 @@ -97,6 +97,7 @@ math/hypot.cl math/ilogb.cl math/clc_ldexp.cl math/ldexp.cl +math/lgamma_r.cl math/log.cl math/log10.cl math/log1p.cl Added: libclc/trunk/generic/lib/math/lgamma_r.cl URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/lgamma_r.cl?rev=281564&view=auto == --- libclc/trunk/generic/lib/math/lgamma_r.cl (added) +++ libclc/trunk/generic/lib/math/lgamma_r.cl Wed Sep 14 19:17:28 2016 @@ -0,0 +1,11 @@ +#include + +#include "../clcmacro.h" +#include "math.h" + +#ifdef cl_khr_fp64 +#pragma OPENCL EXTENSION cl_khr_fp64 : enable +#endif + +#define __CLC_BODY +#include Added: libclc/trunk/generic/lib/math/lgamma_r.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/lgamma_r.inc?rev=281564&view=auto == --- libclc/trunk/generic/lib/math/lgamma_r.inc (added) +++ libclc/trunk/generic/lib/math/lgamma_r.inc Wed Sep 14 19:17:28 2016 @@ -0,0 +1,500 @@ +/* + * Copyright (c) 2014 Advanced Micro Devices, Inc. + * Copyright (c) 2016 Aaron Watry + * + * 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. + */ + +#if __CLC_FPSIZE == 32 +#ifdef __CLC_SCALAR +/* + * + * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. + * + *
[libclc] r281563 - Add ADDR_SPACE parameter to _CLC_V_V_VP_VECTORIZE
Author: awatry
Date: Wed Sep 14 19:17:22 2016
New Revision: 281563
URL: http://llvm.org/viewvc/llvm-project?rev=281563&view=rev
Log:
Add ADDR_SPACE parameter to _CLC_V_V_VP_VECTORIZE
This macro is currently unused, but I plan to use it shortly.
The previous form did casts of pointers without an address space, which
doesn't work so well for CL 1.x.
Signed-off-by: Aaron Watry
Reviewed-by: Tom Stellard
Modified:
libclc/trunk/generic/lib/clcmacro.h
Modified: libclc/trunk/generic/lib/clcmacro.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/clcmacro.h?rev=281563&r1=281562&r2=281563&view=diff
==
--- libclc/trunk/generic/lib/clcmacro.h (original)
+++ libclc/trunk/generic/lib/clcmacro.h Wed Sep 14 19:17:22 2016
@@ -109,26 +109,41 @@
} \
\
-#define _CLC_V_V_VP_VECTORIZE(DECLSPEC, RET_TYPE, FUNCTION, ARG1_TYPE,
ARG2_TYPE) \
- DECLSPEC RET_TYPE##2 FUNCTION(ARG1_TYPE##2 x, ARG2_TYPE##2 *y) { \
-return (RET_TYPE##2)(FUNCTION(x.x, (ARG2_TYPE*)y), FUNCTION(x.y,
(ARG2_TYPE*)y+1)); \
+#define _CLC_V_V_VP_VECTORIZE(DECLSPEC, RET_TYPE, FUNCTION, ARG1_TYPE,
ADDR_SPACE, ARG2_TYPE) \
+ DECLSPEC RET_TYPE##2 FUNCTION(ARG1_TYPE##2 x, ADDR_SPACE ARG2_TYPE##2 *y) { \
+return (RET_TYPE##2)( \
+FUNCTION(x.x, (ARG2_TYPE*)y), \
+FUNCTION(x.y, (ADDR_SPACE ARG2_TYPE*)((ADDR_SPACE ARG2_TYPE*)y+1)) \
+); \
} \
\
- DECLSPEC RET_TYPE##3 FUNCTION(ARG1_TYPE##3 x, ARG2_TYPE##3 *y) { \
-return (RET_TYPE##3)(FUNCTION(x.x, (ARG2_TYPE*)y), FUNCTION(x.y,
(ARG2_TYPE*)y+1), \
- FUNCTION(x.z, (ARG2_TYPE*)y+2)); \
+ DECLSPEC RET_TYPE##3 FUNCTION(ARG1_TYPE##3 x, ADDR_SPACE ARG2_TYPE##3 *y) { \
+return (RET_TYPE##3)( \
+FUNCTION(x.x, (ARG2_TYPE*)y), \
+FUNCTION(x.y, (ADDR_SPACE ARG2_TYPE*)((ADDR_SPACE ARG2_TYPE*)y+1)), \
+FUNCTION(x.z, (ADDR_SPACE ARG2_TYPE*)((ADDR_SPACE ARG2_TYPE*)y+2)) \
+); \
} \
\
- DECLSPEC RET_TYPE##4 FUNCTION(ARG1_TYPE##4 x, ARG2_TYPE##4 *y) { \
-return (RET_TYPE##4)(FUNCTION(x.lo, (ARG2_TYPE##2*)y), FUNCTION(x.hi,
(ARG2_TYPE##2*)((ARG2_TYPE*)y+2))); \
+ DECLSPEC RET_TYPE##4 FUNCTION(ARG1_TYPE##4 x, ADDR_SPACE ARG2_TYPE##4 *y) { \
+return (RET_TYPE##4)( \
+FUNCTION(x.lo, (ARG2_TYPE##2*)y), \
+FUNCTION(x.hi, (ADDR_SPACE ARG2_TYPE##2*)((ADDR_SPACE ARG2_TYPE*)y+2))
\
+); \
} \
\
- DECLSPEC RET_TYPE##8 FUNCTION(ARG1_TYPE##8 x, ARG2_TYPE##8 *y) { \
-return (RET_TYPE##8)(FUNCTION(x.lo, (ARG2_TYPE##4*)y), FUNCTION(x.hi,
(ARG2_TYPE##4*)((ARG2_TYPE*)y+4))); \
+ DECLSPEC RET_TYPE##8 FUNCTION(ARG1_TYPE##8 x, ADDR_SPACE ARG2_TYPE##8 *y) { \
+return (RET_TYPE##8)( \
+FUNCTION(x.lo, (ARG2_TYPE##4*)y), \
+FUNCTION(x.hi, (ADDR_SPACE ARG2_TYPE##4*)((ADDR_SPACE ARG2_TYPE*)y+4))
\
+); \
} \
\
- DECLSPEC RET_TYPE##16 FUNCTION(ARG1_TYPE##16 x, ARG2_TYPE##16 *y) { \
-return (RET_TYPE##16)(FUNCTION(x.lo, (ARG2_TYPE##8*)y), FUNCTION(x.hi,
(ARG2_TYPE##8*)((ARG2_TYPE*)y+8))); \
+ DECLSPEC RET_TYPE##16 FUNCTION(ARG1_TYPE##16 x, ADDR_SPACE ARG2_TYPE##16 *y)
{ \
+return (RET_TYPE##16)( \
+FUNCTION(x.lo, (ARG2_TYPE##8*)y), \
+FUNCTION(x.hi, (ADDR_SPACE ARG2_TYPE##8*)((ADDR_SPACE ARG2_TYPE*)y+8))
\
+); \
}
#define _CLC_DEFINE_BINARY_BUILTIN(RET_TYPE, FUNCTION, BUILTIN, ARG1_TYPE,
ARG2_TYPE) \
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r281566 - math: Implement tgamma
Author: awatry
Date: Wed Sep 14 19:17:34 2016
New Revision: 281566
URL: http://llvm.org/viewvc/llvm-project?rev=281566&view=rev
Log:
math: Implement tgamma
Signed-off-by: Aaron Watry
Reviewed-by: Tom Stellard
Added:
libclc/trunk/generic/include/clc/math/tgamma.h
libclc/trunk/generic/include/clc/math/tgamma.inc
libclc/trunk/generic/lib/math/tgamma.cl
Modified:
libclc/trunk/generic/include/clc/clc.h
libclc/trunk/generic/lib/SOURCES
Modified: libclc/trunk/generic/include/clc/clc.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=281566&r1=281565&r2=281566&view=diff
==
--- libclc/trunk/generic/include/clc/clc.h (original)
+++ libclc/trunk/generic/include/clc/clc.h Wed Sep 14 19:17:34 2016
@@ -88,6 +88,7 @@
#include
#include
#include
+#include
#include
#include
#include
Added: libclc/trunk/generic/include/clc/math/tgamma.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/tgamma.h?rev=281566&view=auto
==
--- libclc/trunk/generic/include/clc/math/tgamma.h (added)
+++ libclc/trunk/generic/include/clc/math/tgamma.h Wed Sep 14 19:17:34 2016
@@ -0,0 +1,3 @@
+#define __CLC_BODY
+#include
+#undef __CLC_BODY
Added: libclc/trunk/generic/include/clc/math/tgamma.inc
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/tgamma.inc?rev=281566&view=auto
==
--- libclc/trunk/generic/include/clc/math/tgamma.inc (added)
+++ libclc/trunk/generic/include/clc/math/tgamma.inc Wed Sep 14 19:17:34 2016
@@ -0,0 +1 @@
+_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE tgamma(__CLC_GENTYPE a);
Modified: libclc/trunk/generic/lib/SOURCES
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=281566&r1=281565&r2=281566&view=diff
==
--- libclc/trunk/generic/lib/SOURCES (original)
+++ libclc/trunk/generic/lib/SOURCES Wed Sep 14 19:17:34 2016
@@ -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
Added: libclc/trunk/generic/lib/math/tgamma.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/tgamma.cl?rev=281566&view=auto
==
--- libclc/trunk/generic/lib/math/tgamma.cl (added)
+++ libclc/trunk/generic/lib/math/tgamma.cl Wed Sep 14 19:17:34 2016
@@ -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
+
+#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, dou
[libclc] r292334 - math: Add expm1 builtin function
Author: awatry
Date: Tue Jan 17 21:13:37 2017
New Revision: 292334
URL: http://llvm.org/viewvc/llvm-project?rev=292334&view=rev
Log:
math: Add expm1 builtin function
Ported from the amd-builtins branch.
Signed-off-by: Aaron Watry
Reviewed-by: Matt Arsenault
CC: Tom Stellard
Added:
libclc/trunk/generic/include/clc/math/expm1.h
libclc/trunk/generic/lib/math/expm1.cl
Modified:
libclc/trunk/generic/include/clc/clc.h
libclc/trunk/generic/lib/SOURCES
libclc/trunk/generic/lib/math/tables.cl
libclc/trunk/generic/lib/math/tables.h
Modified: libclc/trunk/generic/include/clc/clc.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=292334&r1=292333&r2=292334&view=diff
==
--- libclc/trunk/generic/include/clc/clc.h (original)
+++ libclc/trunk/generic/include/clc/clc.h Tue Jan 17 21:13:37 2017
@@ -53,6 +53,7 @@
#include
#include
#include
+#include
#include
#include
#include
Added: libclc/trunk/generic/include/clc/math/expm1.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/expm1.h?rev=292334&view=auto
==
--- libclc/trunk/generic/include/clc/math/expm1.h (added)
+++ libclc/trunk/generic/include/clc/math/expm1.h Tue Jan 17 21:13:37 2017
@@ -0,0 +1,9 @@
+#undef exp
+
+#define __CLC_BODY
+#define __CLC_FUNCTION expm1
+
+#include
+
+#undef __CLC_BODY
+#undef __CLC_FUNCTION
Modified: libclc/trunk/generic/lib/SOURCES
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=292334&r1=292333&r2=292334&view=diff
==
--- libclc/trunk/generic/lib/SOURCES (original)
+++ libclc/trunk/generic/lib/SOURCES Tue Jan 17 21:13:37 2017
@@ -83,6 +83,7 @@ math/erf.cl
math/erfc.cl
math/exp.cl
math/exp_helper.cl
+math/expm1.cl
math/exp2.cl
math/exp10.cl
math/fdim.cl
Added: libclc/trunk/generic/lib/math/expm1.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/expm1.cl?rev=292334&view=auto
==
--- libclc/trunk/generic/lib/math/expm1.cl (added)
+++ libclc/trunk/generic/lib/math/expm1.cl Tue Jan 17 21:13:37 2017
@@ -0,0 +1,142 @@
+#include
+
+#include "math.h"
+#include "tables.h"
+#include "../clcmacro.h"
+
+/* Refer to the exp routine for the underlying algorithm */
+
+_CLC_OVERLOAD _CLC_DEF float expm1(float x) {
+const float X_MAX = 0x1.62e42ep+6f; // 128*log2 : 88.722839111673
+const float X_MIN = -0x1.9d1da0p+6f; // -149*log2 : -103.27892990343184
+
+const float R_64_BY_LOG2 = 0x1.715476p+6f; // 64/log2 :
92.332482616893657
+const float R_LOG2_BY_64_LD = 0x1.62p-7f; // log2/64 lead:
0.0108032227
+const float R_LOG2_BY_64_TL = 0x1.c85fdep-16f; // log2/64 tail:
0.272020388
+
+uint xi = as_uint(x);
+int n = (int)(x * R_64_BY_LOG2);
+float fn = (float)n;
+
+int j = n & 0x3f;
+int m = n >> 6;
+
+float r = mad(fn, -R_LOG2_BY_64_TL, mad(fn, -R_LOG2_BY_64_LD, x));
+
+// Truncated Taylor series
+float z2 = mad(r*r, mad(r, mad(r, 0x1.56p-5f, 0x1.56p-3f), 0.5f),
r);
+
+float m2 = as_float((m + EXPBIAS_SP32) << EXPSHIFTBITS_SP32);
+float2 tv = USE_TABLE(exp_tbl_ep, j);
+
+float two_to_jby64_h = tv.s0 * m2;
+float two_to_jby64_t = tv.s1 * m2;
+float two_to_jby64 = two_to_jby64_h + two_to_jby64_t;
+
+z2 = mad(z2, two_to_jby64, two_to_jby64_t) + (two_to_jby64_h - 1.0f);
+ //Make subnormals work
+z2 = x == 0.f ? x : z2;
+z2 = x < X_MIN | m < -24 ? -1.0f : z2;
+z2 = x > X_MAX ? as_float(PINFBITPATT_SP32) : z2;
+z2 = isnan(x) ? x : z2;
+
+return z2;
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, expm1, float)
+
+#ifdef cl_khr_fp64
+
+#include "exp_helper.h"
+
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+
+_CLC_OVERLOAD _CLC_DEF double expm1(double x) {
+const double max_expm1_arg = 709.8;
+const double min_expm1_arg = -37.42994775023704;
+const double log_OnePlus_OneByFour = 0.22314355131420976;
//0x3FCC8FF7C79A9A22 = log(1+1/4)
+const double log_OneMinus_OneByFour = -0.28768207245178096;
//0xBFD269621134DB93 = log(1-1/4)
+const double sixtyfour_by_lnof2 = 92.33248261689366;
//0x40571547652b82fe
+const double lnof2_by_64_head = 0.010830424696223417;
//0x3f862e42fefa
+const double lnof2_by_64_tail = 2.5728046223276688e-14;
//0x3d1cf79abc9e3b39
+
+// First, assume log(1-1/4) < x < log(1+1/4) i.e -0.28768 < x < 0.22314
+double u = as_double(as_ulong(x) & 0xff00UL);
+double v = x - u;
+double y = u * u * 0.5;
+double z = v * (x + u) * 0.5;
+
+double q = fma(x,
+ fma(x,
+ fma(x,
+ fma(x,
+
[libclc] r292335 - math: Add logb builtin
Author: awatry
Date: Tue Jan 17 21:14:10 2017
New Revision: 292335
URL: http://llvm.org/viewvc/llvm-project?rev=292335&view=rev
Log:
math: Add logb builtin
Ported from the amd-builtins branch.
Signed-off-by: Aaron Watry
Reviewed-by: Matt Arsenault
CC: Tom Stellard
Added:
libclc/trunk/generic/include/clc/math/logb.h
libclc/trunk/generic/include/clc/math/logb.inc
libclc/trunk/generic/lib/math/logb.cl
Modified:
libclc/trunk/generic/include/clc/clc.h
libclc/trunk/generic/lib/SOURCES
Modified: libclc/trunk/generic/include/clc/clc.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=292335&r1=292334&r2=292335&view=diff
==
--- libclc/trunk/generic/include/clc/clc.h (original)
+++ libclc/trunk/generic/include/clc/clc.h Tue Jan 17 21:14:10 2017
@@ -76,6 +76,7 @@
#include
#include
#include
+#include
#include
#include
#include
Added: libclc/trunk/generic/include/clc/math/logb.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/logb.h?rev=292335&view=auto
==
--- libclc/trunk/generic/include/clc/math/logb.h (added)
+++ libclc/trunk/generic/include/clc/math/logb.h Tue Jan 17 21:14:10 2017
@@ -0,0 +1,2 @@
+#define __CLC_BODY
+#include
Added: libclc/trunk/generic/include/clc/math/logb.inc
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/logb.inc?rev=292335&view=auto
==
--- libclc/trunk/generic/include/clc/math/logb.inc (added)
+++ libclc/trunk/generic/include/clc/math/logb.inc Tue Jan 17 21:14:10 2017
@@ -0,0 +1 @@
+_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE logb(__CLC_GENTYPE a);
Modified: libclc/trunk/generic/lib/SOURCES
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=292335&r1=292334&r2=292335&view=diff
==
--- libclc/trunk/generic/lib/SOURCES (original)
+++ libclc/trunk/generic/lib/SOURCES Tue Jan 17 21:14:10 2017
@@ -104,6 +104,7 @@ math/log.cl
math/log10.cl
math/log1p.cl
math/log2.cl
+math/logb.cl
math/mad.cl
math/modf.cl
math/native_log.cl
Added: libclc/trunk/generic/lib/math/logb.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/logb.cl?rev=292335&view=auto
==
--- libclc/trunk/generic/lib/math/logb.cl (added)
+++ libclc/trunk/generic/lib/math/logb.cl Tue Jan 17 21:14:10 2017
@@ -0,0 +1,31 @@
+#include
+#include "math.h"
+#include "../clcmacro.h"
+
+_CLC_OVERLOAD _CLC_DEF float logb(float x) {
+int ax = as_int(x) & EXSIGNBIT_SP32;
+float s = -118 - clz(ax);
+float r = (ax >> EXPSHIFTBITS_SP32) - EXPBIAS_SP32;
+r = ax >= PINFBITPATT_SP32 ? as_float(ax) : r;
+r = ax < 0x0080 ? s : r;
+r = ax == 0 ? as_float(NINFBITPATT_SP32) : r;
+return r;
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, logb, float);
+
+#ifdef cl_khr_fp64
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+
+_CLC_OVERLOAD _CLC_DEF double logb(double x) {
+long ax = as_long(x) & EXSIGNBIT_DP64;
+double s = -1011L - clz(ax);
+double r = (int) (ax >> EXPSHIFTBITS_DP64) - EXPBIAS_DP64;
+r = ax >= PINFBITPATT_DP64 ? as_double(ax) : r;
+r = ax < 0x0010L ? s : r;
+r = ax == 0L ? as_double(NINFBITPATT_DP64) : r;
+return r;
+}
+
+_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, logb, double)
+#endif
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r295920 - math: Add native_tan as wrapper to tan
Author: awatry Date: Wed Feb 22 19:46:57 2017 New Revision: 295920 URL: http://llvm.org/viewvc/llvm-project?rev=295920&view=rev Log: math: Add native_tan as wrapper to tan Trivially define native_tan as a redirect to tan. If there are any targets with a native implementation, we can deal with it later. Signed-off-by: Aaron Watry Reviewed-by: Matt Arsenault Added: libclc/trunk/generic/include/clc/math/native_tan.h Modified: libclc/trunk/generic/include/clc/clc.h Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=295920&r1=295919&r2=295920&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Wed Feb 22 19:46:57 2017 @@ -103,6 +103,7 @@ #include #include #include +#include #include /* 6.11.2.1 Floating-point macros */ Added: libclc/trunk/generic/include/clc/math/native_tan.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/native_tan.h?rev=295920&view=auto == --- libclc/trunk/generic/include/clc/math/native_tan.h (added) +++ libclc/trunk/generic/include/clc/math/native_tan.h Wed Feb 22 19:46:57 2017 @@ -0,0 +1,10 @@ +//===-- generic/include/clc/math/native_tan.h -===// + +// +// The LLVM Compiler Infrastructure +// +// This file is dual licensed under both the University of Illinois Open Source +// License and the MIT license. See LICENSE.TXT for details. +// +//===--===// +#define native_tan tan ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r247661 - integer: Update integer limits to comply with spec
Author: awatry Date: Mon Sep 14 22:56:21 2015 New Revision: 247661 URL: http://llvm.org/viewvc/llvm-project?rev=247661&view=rev Log: integer: Update integer limits to comply with spec The values for the char/short/integer/long minimums were declared with their actual values, not the definitions from the CL spec (v1.1). As a result, (-2147483648) was actually being treated as a long by the compiler, not an int, which caused issues when trying to add/subtract that value from a vector. Update the definitions to use the values declared by the spec, and also add explicit casts for the char/short/int minimums so that the compiler actually treats them as shorts/chars. Without those casts, they actually end up stored as integers, and the compiler may end up storing the INT_MIN as a long. The compiler can sign extend the values if it needs to convert the char->short, short->int, or int->long v2: Add explicit cast for INT_MIN and fix some type-o's and wrapping in the commit message. Reported-by: Moritz Pflanzer CC: Moritz Pflanzer Reviewed-by: Tom Stellard Signed-off-by: Aaron Watry Modified: libclc/trunk/generic/include/clc/integer/definitions.h Modified: libclc/trunk/generic/include/clc/integer/definitions.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/integer/definitions.h?rev=247661&r1=247660&r2=247661&view=diff == --- libclc/trunk/generic/include/clc/integer/definitions.h (original) +++ libclc/trunk/generic/include/clc/integer/definitions.h Mon Sep 14 22:56:21 2015 @@ -1,14 +1,14 @@ #define CHAR_BIT 8 #define INT_MAX 2147483647 -#define INT_MIN -2147483648 +#define INT_MIN ((int)(-2147483647 - 1)) #define LONG_MAX 0x7fffL -#define LONG_MIN -0x8000L +#define LONG_MIN (-0x7fffL - 1) +#define CHAR_MAX SCHAR_MAX +#define CHAR_MIN SCHAR_MIN #define SCHAR_MAX 127 -#define SCHAR_MIN -128 -#define CHAR_MAX 127 -#define CHAR_MIN -128 +#define SCHAR_MIN ((char)(-127 - 1)) #define SHRT_MAX 32767 -#define SHRT_MIN -32768 +#define SHRT_MIN ((short)(-32767 - 1)) #define UCHAR_MAX 255 #define USHRT_MAX 65535 #define UINT_MAX 0x ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r249445 - integer: remove explicit casts from _MIN definitions
Author: awatry Date: Tue Oct 6 14:12:12 2015 New Revision: 249445 URL: http://llvm.org/viewvc/llvm-project?rev=249445&view=rev Log: integer: remove explicit casts from _MIN definitions The spec says (section 6.12.3, CL version 1.2): The macro names given in the following list must use the values specified. The values shall all be constant expressions suitable for use in #if preprocessing directives. This commit addresses the second part of that statement. Reviewed-by: Jan Vesely Reviewed-by: Tom Stellard CC: Moritz Pflanzer CC: Serge Martin Modified: libclc/trunk/generic/include/clc/integer/definitions.h Modified: libclc/trunk/generic/include/clc/integer/definitions.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/integer/definitions.h?rev=249445&r1=249444&r2=249445&view=diff == --- libclc/trunk/generic/include/clc/integer/definitions.h (original) +++ libclc/trunk/generic/include/clc/integer/definitions.h Tue Oct 6 14:12:12 2015 @@ -1,14 +1,14 @@ #define CHAR_BIT 8 #define INT_MAX 2147483647 -#define INT_MIN ((int)(-2147483647 - 1)) +#define INT_MIN (-2147483647 - 1) #define LONG_MAX 0x7fffL #define LONG_MIN (-0x7fffL - 1) #define CHAR_MAX SCHAR_MAX #define CHAR_MIN SCHAR_MIN #define SCHAR_MAX 127 -#define SCHAR_MIN ((char)(-127 - 1)) +#define SCHAR_MIN (-127 - 1) #define SHRT_MAX 32767 -#define SHRT_MIN ((short)(-32767 - 1)) +#define SHRT_MIN (-32767 - 1) #define UCHAR_MAX 255 #define USHRT_MAX 65535 #define UINT_MAX 0x ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r260114 - math: Add frexp ported from amd-builtins
Author: awatry Date: Mon Feb 8 11:07:21 2016 New Revision: 260114 URL: http://llvm.org/viewvc/llvm-project?rev=260114&view=rev Log: math: Add frexp ported from amd-builtins The float implementation is almost a direct port from the amd-builtins, but instead of just having a scalar and float4 implementation, it has a scalar and arbitrary width vector implementation. The double scalar is also a direct port from AMD's builtin release. The double vector implementation copies the logic in the float vector implementation using the values from the double scalar version. Both have been tested in piglit using tests sent to that project's mailing list. Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely Added: libclc/trunk/generic/include/clc/math/frexp.h libclc/trunk/generic/include/clc/math/frexp.inc libclc/trunk/generic/lib/math/frexp.cl libclc/trunk/generic/lib/math/frexp.inc Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/include/clc/math/gentype.inc libclc/trunk/generic/lib/SOURCES Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=260114&r1=260113&r2=260114&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Mon Feb 8 11:07:21 2016 @@ -58,6 +58,7 @@ #include #include #include +#include #include #include #include Added: libclc/trunk/generic/include/clc/math/frexp.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/frexp.h?rev=260114&view=auto == --- libclc/trunk/generic/include/clc/math/frexp.h (added) +++ libclc/trunk/generic/include/clc/math/frexp.h Mon Feb 8 11:07:21 2016 @@ -0,0 +1,2 @@ +#define __CLC_BODY +#include Added: libclc/trunk/generic/include/clc/math/frexp.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/frexp.inc?rev=260114&view=auto == --- libclc/trunk/generic/include/clc/math/frexp.inc (added) +++ libclc/trunk/generic/include/clc/math/frexp.inc Mon Feb 8 11:07:21 2016 @@ -0,0 +1,3 @@ +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE frexp(__CLC_GENTYPE x, global __CLC_INTN *iptr); +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE frexp(__CLC_GENTYPE x, local __CLC_INTN *iptr); +_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE frexp(__CLC_GENTYPE x, private __CLC_INTN *iptr); Modified: libclc/trunk/generic/include/clc/math/gentype.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/gentype.inc?rev=260114&r1=260113&r2=260114&view=diff == --- libclc/trunk/generic/include/clc/math/gentype.inc (original) +++ libclc/trunk/generic/include/clc/math/gentype.inc Mon Feb 8 11:07:21 2016 @@ -2,38 +2,50 @@ #define __CLC_FPSIZE 32 #define __CLC_GENTYPE float +#define __CLC_INTN int #define __CLC_SCALAR #include __CLC_BODY #undef __CLC_GENTYPE +#undef __CLC_INTN #undef __CLC_SCALAR #define __CLC_GENTYPE float2 #define __CLC_INTN int2 +#define __CLC_VECSIZE 2 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE float3 #define __CLC_INTN int3 +#define __CLC_VECSIZE 3 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE float4 #define __CLC_INTN int4 +#define __CLC_VECSIZE 4 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE float8 #define __CLC_INTN int8 +#define __CLC_VECSIZE 8 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE float16 #define __CLC_INTN int16 +#define __CLC_VECSIZE 16 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN @@ -47,37 +59,49 @@ #define __CLC_SCALAR #define __CLC_GENTYPE double +#define __CLC_INTN int #include __CLC_BODY #undef __CLC_GENTYPE +#undef __CLC_INTN #undef __CLC_SCALAR #define __CLC_GENTYPE double2 #define __CLC_INTN int2 +#define __CLC_VECSIZE 2 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE double3 #define __CLC_INTN int3 +#define __CLC_VECSIZE 3 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE double4 #define __CLC_INTN int4 +#define __CLC_VECSIZE 4 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE double8 #define __CLC_INTN int8 +#define __CLC_VECSIZE 8 #include __CLC_BODY +#undef __CLC_VECSIZE #undef __CLC_GENTYPE #undef __CLC_INTN #define __CLC_GENTYPE double16 #define __CLC_INTN int16 +#define __CLC_VECSIZE 16
[libclc] r261639 - math: Add ilogb ported from amd-builtins
Author: awatry Date: Tue Feb 23 08:43:09 2016 New Revision: 261639 URL: http://llvm.org/viewvc/llvm-project?rev=261639&view=rev Log: math: Add ilogb ported from amd-builtins The scalar float/double function bodies are a direct copy/paste with usage of the CLC wrappers to vectorize them. This commit also adds in the FP_ILOGB0 and FP_ILOGBNAN macros which are equal to the results of ilogb(0.0f) and ilogb(float nan) respectively. v2: Add FP_ILOGB0 and FP_ILOGBNAN definitions Signed-off-by: Aaron Watry Reviewed-by: Jan Vesely v1 Reviewed-by: Tom Stellard Added: libclc/trunk/generic/include/clc/math/ilogb.h libclc/trunk/generic/include/clc/math/ilogb.inc libclc/trunk/generic/lib/math/ilogb.cl Modified: libclc/trunk/generic/include/clc/clc.h libclc/trunk/generic/include/clc/float/definitions.h libclc/trunk/generic/lib/SOURCES Modified: libclc/trunk/generic/include/clc/clc.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=261639&r1=261638&r2=261639&view=diff == --- libclc/trunk/generic/include/clc/clc.h (original) +++ libclc/trunk/generic/include/clc/clc.h Tue Feb 23 08:43:09 2016 @@ -62,6 +62,7 @@ #include #include #include +#include #include #include #include Modified: libclc/trunk/generic/include/clc/float/definitions.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/float/definitions.h?rev=261639&r1=261638&r2=261639&view=diff == --- libclc/trunk/generic/include/clc/float/definitions.h (original) +++ libclc/trunk/generic/include/clc/float/definitions.h Tue Feb 23 08:43:09 2016 @@ -14,6 +14,9 @@ #define FLT_MIN 0x1.0p-126f #define FLT_EPSILON 0x1.0p-23f +#define FP_ILOGB0 (-2147483647 - 1) +#define FP_ILOGBNAN (-2147483647 - 1) + #define M_E_F 0x1.5bf0a8p+1f #define M_LOG2E_F 0x1.715476p+0f #define M_LOG10E_F 0x1.bcb7b2p-2f Added: libclc/trunk/generic/include/clc/math/ilogb.h URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/ilogb.h?rev=261639&view=auto == --- libclc/trunk/generic/include/clc/math/ilogb.h (added) +++ libclc/trunk/generic/include/clc/math/ilogb.h Tue Feb 23 08:43:09 2016 @@ -0,0 +1,5 @@ +#define __CLC_BODY + +#include + +#undef __CLC_BODY Added: libclc/trunk/generic/include/clc/math/ilogb.inc URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/ilogb.inc?rev=261639&view=auto == --- libclc/trunk/generic/include/clc/math/ilogb.inc (added) +++ libclc/trunk/generic/include/clc/math/ilogb.inc Tue Feb 23 08:43:09 2016 @@ -0,0 +1 @@ +_CLC_OVERLOAD _CLC_DECL __CLC_INTN ilogb(__CLC_GENTYPE x); Modified: libclc/trunk/generic/lib/SOURCES URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=261639&r1=261638&r2=261639&view=diff == --- libclc/trunk/generic/lib/SOURCES (original) +++ libclc/trunk/generic/lib/SOURCES Tue Feb 23 08:43:09 2016 @@ -90,6 +90,7 @@ math/frexp.cl math/half_rsqrt.cl math/half_sqrt.cl math/hypot.cl +math/ilogb.cl math/clc_ldexp.cl math/ldexp.cl math/log.cl Added: libclc/trunk/generic/lib/math/ilogb.cl URL: http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/ilogb.cl?rev=261639&view=auto == --- libclc/trunk/generic/lib/math/ilogb.cl (added) +++ libclc/trunk/generic/lib/math/ilogb.cl Tue Feb 23 08:43:09 2016 @@ -0,0 +1,57 @@ +/* + * Copyright (c) 2015 Advanced Micro Devices, Inc. + * Copyright (c) 2016 Aaron Watry + * + * 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 SOFTW
[libclc] r261714 - math: Fix ilogb(double) return type
Author: awatry
Date: Tue Feb 23 18:52:15 2016
New Revision: 261714
URL: http://llvm.org/viewvc/llvm-project?rev=261714&view=rev
Log:
math: Fix ilogb(double) return type
Signed-off-by: Aaron Watry
Reviewed-by: Jan Vesely
Modified:
libclc/trunk/generic/lib/math/ilogb.cl
Modified: libclc/trunk/generic/lib/math/ilogb.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/ilogb.cl?rev=261714&r1=261713&r2=261714&view=diff
==
--- libclc/trunk/generic/lib/math/ilogb.cl (original)
+++ libclc/trunk/generic/lib/math/ilogb.cl Tue Feb 23 18:52:15 2016
@@ -41,7 +41,7 @@ _CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_
#ifdef cl_khr_fp64
#pragma OPENCL EXTENSION cl_khr_fp64 : enable
-_CLC_OVERLOAD _CLC_DEF ilogb(double x) {
+_CLC_OVERLOAD _CLC_DEF int ilogb(double x) {
ulong ux = as_ulong(x);
ulong ax = ux & ~SIGNBIT_DP64;
int r = (int) (ax >> EXPSHIFTBITS_DP64) - EXPBIAS_DP64;
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[libclc] r268708 - math: Add fdim implementation
Author: awatry
Date: Thu May 5 22:34:45 2016
New Revision: 268708
URL: http://llvm.org/viewvc/llvm-project?rev=268708&view=rev
Log:
math: Add fdim implementation
Based on the amd-builtin, but explicitly vectorized for all sizes (not just
float4), and includes a vectorized double implementation.
Passes piglit (float) tests on pitcairn.
Signed-off-by: Aaron Watry
Reviewed-by: Jan Vesely
Added:
libclc/trunk/generic/include/clc/math/fdim.h
libclc/trunk/generic/include/clc/math/fdim.inc
libclc/trunk/generic/lib/math/fdim.cl
libclc/trunk/generic/lib/math/fdim.inc
Modified:
libclc/trunk/generic/include/clc/clc.h
libclc/trunk/generic/lib/SOURCES
Modified: libclc/trunk/generic/include/clc/clc.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/clc.h?rev=268708&r1=268707&r2=268708&view=diff
==
--- libclc/trunk/generic/include/clc/clc.h (original)
+++ libclc/trunk/generic/include/clc/clc.h Thu May 5 22:34:45 2016
@@ -52,6 +52,7 @@
#include
#include
#include
+#include
#include
#include
#include
Added: libclc/trunk/generic/include/clc/math/fdim.h
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/fdim.h?rev=268708&view=auto
==
--- libclc/trunk/generic/include/clc/math/fdim.h (added)
+++ libclc/trunk/generic/include/clc/math/fdim.h Thu May 5 22:34:45 2016
@@ -0,0 +1,2 @@
+#define __CLC_BODY
+#include
Added: libclc/trunk/generic/include/clc/math/fdim.inc
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/include/clc/math/fdim.inc?rev=268708&view=auto
==
--- libclc/trunk/generic/include/clc/math/fdim.inc (added)
+++ libclc/trunk/generic/include/clc/math/fdim.inc Thu May 5 22:34:45 2016
@@ -0,0 +1 @@
+_CLC_OVERLOAD _CLC_DECL __CLC_GENTYPE fdim(__CLC_GENTYPE a, __CLC_GENTYPE b);
Modified: libclc/trunk/generic/lib/SOURCES
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/SOURCES?rev=268708&r1=268707&r2=268708&view=diff
==
--- libclc/trunk/generic/lib/SOURCES (original)
+++ libclc/trunk/generic/lib/SOURCES Thu May 5 22:34:45 2016
@@ -82,6 +82,7 @@ math/exp.cl
math/exp_helper.cl
math/exp2.cl
math/exp10.cl
+math/fdim.cl
math/fmax.cl
math/fmin.cl
math/fmod.cl
Added: libclc/trunk/generic/lib/math/fdim.cl
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/fdim.cl?rev=268708&view=auto
==
--- libclc/trunk/generic/lib/math/fdim.cl (added)
+++ libclc/trunk/generic/lib/math/fdim.cl Thu May 5 22:34:45 2016
@@ -0,0 +1,10 @@
+#include
+
+#include "math.h"
+
+#ifdef cl_khr_fp64
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+#endif
+
+#define __CLC_BODY
+#include
Added: libclc/trunk/generic/lib/math/fdim.inc
URL:
http://llvm.org/viewvc/llvm-project/libclc/trunk/generic/lib/math/fdim.inc?rev=268708&view=auto
==
--- libclc/trunk/generic/lib/math/fdim.inc (added)
+++ libclc/trunk/generic/lib/math/fdim.inc Thu May 5 22:34:45 2016
@@ -0,0 +1,71 @@
+/*
+ * Copyright (c) 2014 Advanced Micro Devices, Inc.
+ * Copyright (c) 2016 Aaron Watry
+ *
+ * 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.
+ */
+#if __CLC_FPSIZE == 32
+#ifdef __CLC_SCALAR
+_CLC_OVERLOAD _CLC_DEF __CLC_GENTYPE fdim(__CLC_GENTYPE x, __CLC_GENTYPE y) {
+if (__builtin_isnan(x) || __builtin_isnan(y))
+return as_float(QNANBITPATT_SP32);
+return __builtin_fmax(x - y, 0);
+}
+#define __CLC_FDIM_VEC(width) \
+_CLC_OVERLOAD _CLC_DEF float##width fdim(float##width x, float##width y) { \
+/* Determine if x or y is NaN. */ \
+/*
