This is the mail archive of the
glibc-cvs@sourceware.org
mailing list for the glibc project.
GNU C Library master sources branch master updated. glibc-2.21-451-g24a2718
- From: andros at sourceware dot org
- To: glibc-cvs at sourceware dot org
- Date: 9 Jun 2015 11:54:58 -0000
- Subject: GNU C Library master sources branch master updated. glibc-2.21-451-g24a2718
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "GNU C Library master sources".
The branch, master has been updated
via 24a2718f595bc11dc6abb31303ceb8fdcb664f2f (commit)
from 2193311288b97cf11dfabf1be22eac89b4ff7366 (commit)
Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.
- Log -----------------------------------------------------------------
http://sourceware.org/git/gitweb.cgi?p=glibc.git;a=commitdiff;h=24a2718f595bc11dc6abb31303ceb8fdcb664f2f
commit 24a2718f595bc11dc6abb31303ceb8fdcb664f2f
Author: Andrew Senkevich <andrew.senkevich@intel.com>
Date: Tue Jun 9 14:51:52 2015 +0300
Addition of testing infrastructure for vector math functions.
We test vector math functions using scalar tests infrastructure with
help of special wrappers from scalar versions to vector ones. Wrapper
implemented using platform specific vector types and placed in separate
file for compilation with architecture specific options, main part of
test has no such options. With help of system of definitions unfolding
of which is drived from test code we have wrapper called in individual
testing function instead of scalar function. Also system of definitions
includes generated during make check header math/libm-have-vector-test.h
with series of conditional definitions which help to avoid build fails
for functions having no vector versions; runtime architecture check
to prevent runtime fails of test run on inappropriate hardware.
* math/Makefile: Added rules for vector tests.
* math/gen-libm-have-vector-test.sh: Added generation of wrapper
declaration under condition.
* math/test-double-vlen2.h: New file.
* math/test-double-vlen4.h: New file.
* math/test-double-vlen8.h: New file.
* math/test-vec-loop.h: Added initialization macro.
* sysdeps/x86_64/fpu/Makefile: Added variables for vector tests.
* sysdeps/x86_64/fpu/libm-test-ulps: Regenarated.
* sysdeps/x86_64/fpu/math-tests-arch.h: New file.
* sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen2.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen8.c: New file.
diff --git a/ChangeLog b/ChangeLog
index 0877e05..47318a7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -28,6 +28,25 @@
* INSTALL: Regenerated.
* NEWS: Mention addition of libmvec and x86_64 vector cos.
+ * math/Makefile: Added rules for vector tests.
+ * math/gen-libm-have-vector-test.sh: Added generation of wrapper
+ declaration under condition.
+ * math/test-double-vlen2.h: New file.
+ * math/test-double-vlen4.h: New file.
+ * math/test-double-vlen8.h: New file.
+ * math/test-vec-loop.h: Added initialization macro.
+ * sysdeps/x86_64/fpu/Makefile: Added variables for vector tests.
+ * sysdeps/x86_64/fpu/libm-test-ulps: Regenarated.
+ * sysdeps/x86_64/fpu/math-tests-arch.h: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen2.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen4-avx2.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen4.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c: New file.
+ * sysdeps/x86_64/fpu/test-double-vlen8.c: New file.
+
2015-06-09 Marko Myllynen <myllynen@redhat.com>
* locale/C-ctype.c (PREDEFINED_CLASSES): Remove.
diff --git a/math/Makefile b/math/Makefile
index 9a3cf32..8f14f25 100644
--- a/math/Makefile
+++ b/math/Makefile
@@ -115,8 +115,9 @@ tests-static = test-fpucw-static test-fpucw-ieee-static
test-longdouble-yes = test-ldouble test-ildoubl
ifneq (no,$(PERL))
+libm-vec-tests = $(addprefix test-,$(libmvec-tests))
libm-tests = test-float test-double $(test-longdouble-$(long-double-fcts)) \
- test-ifloat test-idouble
+ test-ifloat test-idouble $(libm-vec-tests)
libm-tests.o = $(addsuffix .o,$(libm-tests))
tests += $(libm-tests)
@@ -143,8 +144,22 @@ $(objpfx)test-double.o: $(objpfx)libm-test.stmp
$(objpfx)test-idouble.o: $(objpfx)libm-test.stmp
$(objpfx)test-ldouble.o: $(objpfx)libm-test.stmp
$(objpfx)test-ildoubl.o: $(objpfx)libm-test.stmp
+$(addprefix $(objpfx), $(addsuffix .o, $(libm-vec-tests))): $(objpfx)libm-test.stmp
endif
+libm-test-fast-math-cflags = -fno-builtin -D__FAST_MATH__ -DTEST_FAST_MATH
+libm-test-vec-cflags = $(libm-test-fast-math-cflags) -fno-inline \
+ -ffloat-store -D_OPENMP=201307 -Wno-unknown-pragmas
+
+CFLAGS-test-double-vlen2.c = $(libm-test-vec-cflags)
+CFLAGS-test-double-vlen2-wrappers.c = $(double-vlen2-arch-ext-cflags)
+
+CFLAGS-test-double-vlen4.c = $(libm-test-vec-cflags)
+CFLAGS-test-double-vlen4-wrappers.c = $(double-vlen4-arch-ext-cflags)
+
+CFLAGS-test-double-vlen8.c = $(libm-test-vec-cflags)
+CFLAGS-test-double-vlen8-wrappers.c = $(double-vlen8-arch-ext-cflags)
+
CFLAGS-test-float.c = -fno-inline -ffloat-store -fno-builtin
CFLAGS-test-double.c = -fno-inline -ffloat-store -fno-builtin
CFLAGS-test-ldouble.c = -fno-inline -ffloat-store -fno-builtin
@@ -155,12 +170,12 @@ CFLAGS-test-tgmath.c = -fno-builtin
CFLAGS-test-tgmath2.c = -fno-builtin
CFLAGS-test-tgmath-ret.c = -fno-builtin
CFLAGS-test-powl.c = -fno-builtin
-CPPFLAGS-test-ifloat.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__ \
- -DTEST_FAST_MATH -fno-builtin
-CPPFLAGS-test-idouble.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__ \
- -DTEST_FAST_MATH -fno-builtin
-CPPFLAGS-test-ildoubl.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__ \
- -DTEST_FAST_MATH -fno-builtin
+CPPFLAGS-test-ifloat.c = -U__LIBC_INTERNAL_MATH_INLINES \
+ $(libm-test-fast-math-cflags)
+CPPFLAGS-test-idouble.c = -U__LIBC_INTERNAL_MATH_INLINES \
+ $(libm-test-fast-math-cflags)
+CPPFLAGS-test-ildoubl.c = -U__LIBC_INTERNAL_MATH_INLINES \
+ $(libm-test-fast-math-cflags)
# The -lieee module sets the _LIB_VERSION_ switch to IEEE mode
@@ -236,8 +251,10 @@ $(objpfx)libieee.a: $(objpfx)ieee-math.o
$(patsubst %/,cd % &&,$(objpfx)) \
$(LN_S) $(<F) $(@F)
-$(addprefix $(objpfx),$(filter-out $(tests-static),$(tests))): $(libm)
+$(addprefix $(objpfx),$(filter-out $(tests-static) $(libm-vec-tests),$(tests))): $(libm)
$(addprefix $(objpfx),$(tests-static)): $(objpfx)libm.a
+$(addprefix $(objpfx), $(libm-vec-tests)): $(objpfx)%: $(libm) $(libmvec) \
+ $(objpfx)init-arch.o $(objpfx)%-wrappers.o
gmp-objs = $(patsubst %,$(common-objpfx)stdlib/%.o,\
add_n sub_n cmp addmul_1 mul_1 mul_n divmod_1 \
diff --git a/math/gen-libm-have-vector-test.sh b/math/gen-libm-have-vector-test.sh
index eeddf76..8c36078 100644
--- a/math/gen-libm-have-vector-test.sh
+++ b/math/gen-libm-have-vector-test.sh
@@ -25,6 +25,7 @@ print_defs()
echo "#if defined TEST_VECTOR_$1 && TEST_VECTOR_$1"
echo "# define HAVE_VECTOR_$1 1"
echo "# define ${1}_VEC_SUFF WRAPPER_NAME ($1)"
+ echo "WRAPPER_DECL$2 (WRAPPER_NAME ($1))"
echo "#else"
echo "# define HAVE_VECTOR_$1 0"
echo "# define ${1}_VEC_SUFF $1"
@@ -32,12 +33,18 @@ print_defs()
echo
}
-for func in $(grep ALL_RM_TEST libm-test.inc | grep -v define | sed -r "s/.*\(//; s/,.*//"); do
+for func in $(cat libm-test.inc | grep ALL_RM_TEST | grep -v define | grep -v RUN_TEST_LOOP_ff_f | sed -r "s/.*\(//; s/,.*//" ); do
print_defs ${func}
print_defs ${func}f
print_defs ${func}l
done
+for func in $(cat libm-test.inc | grep ALL_RM_TEST | grep RUN_TEST_LOOP_ff_f | sed -r "s/.*\(//; s/,.*//" ); do
+ print_defs ${func} "_ff"
+ print_defs ${func}f "_ff"
+ print_defs ${func}l "_ff"
+done
+
# When all functions will use ALL_RM_TEST instead of using START directly,
# this code can be removed.
for func in $(grep 'START.*;$' libm-test.inc | sed -r "s/.*\(//; s/,.*//"); do
diff --git a/math/test-double-vlen2.h b/math/test-double-vlen2.h
new file mode 100644
index 0000000..37d7060
--- /dev/null
+++ b/math/test-double-vlen2.h
@@ -0,0 +1,71 @@
+/* Definitions for double vector tests with vector length 2.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#define FLOAT double
+#define FUNC(function) function
+#define TEST_MSG "testing double vector math (without inline functions)\n"
+#define MATHCONST(x) x
+#define CHOOSE(Clongdouble,Cdouble,Cfloat,Cinlinelongdouble,Cinlinedouble,Cinlinefloat) Cdouble
+#define PRINTF_EXPR "e"
+#define PRINTF_XEXPR "a"
+#define PRINTF_NEXPR "f"
+#define TEST_DOUBLE 1
+#define TEST_MATHVEC 1
+
+#ifndef __NO_MATH_INLINES
+# define __NO_MATH_INLINES
+#endif
+
+#define EXCEPTION_TESTS_double 0
+#define ROUNDING_TESTS_double(MODE) ((MODE) == FE_TONEAREST)
+
+#define CNCT(x, y) x ## y
+#define CONCAT(a, b) CNCT (a, b)
+
+#define VEC_SUFF _vlen2
+#define WRAPPER_NAME(function) CONCAT (function, VEC_SUFF)
+
+#define FUNC_TEST(function) function ## _VEC_SUFF
+
+#define WRAPPER_DECL(function) extern FLOAT function (FLOAT);
+#define WRAPPER_DECL_ff(function) extern FLOAT function (FLOAT, FLOAT);
+
+// Wrapper from scalar to vector function with vector length 8.
+#define VECTOR_WRAPPER(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE); \
+FLOAT scalar_func (FLOAT x) \
+{ \
+ int i; \
+ VEC_TYPE mx; \
+ INIT_VEC_LOOP (mx, x, 2); \
+ VEC_TYPE mr = vector_func (mx); \
+ TEST_VEC_LOOP (2); \
+}
+
+// Wrapper from scalar 2 argument function to vector one.
+#define VECTOR_WRAPPER_ff(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE, VEC_TYPE); \
+FLOAT scalar_func (FLOAT x, FLOAT y) \
+{ \
+ int i; \
+ VEC_TYPE mx; \
+ INIT_VEC_LOOP (mx, x, 2); \
+ INIT_VEC_LOOP (my, y, 2); \
+ VEC_TYPE mr = vector_func (mx, my); \
+ TEST_VEC_LOOP (2); \
+}
diff --git a/math/test-double-vlen4.h b/math/test-double-vlen4.h
new file mode 100644
index 0000000..4fca9d8
--- /dev/null
+++ b/math/test-double-vlen4.h
@@ -0,0 +1,71 @@
+/* Definitions for double vector tests with vector length 4.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#define FLOAT double
+#define FUNC(function) function
+#define TEST_MSG "testing double vector math (without inline functions)\n"
+#define MATHCONST(x) x
+#define CHOOSE(Clongdouble,Cdouble,Cfloat,Cinlinelongdouble,Cinlinedouble,Cinlinefloat) Cdouble
+#define PRINTF_EXPR "e"
+#define PRINTF_XEXPR "a"
+#define PRINTF_NEXPR "f"
+#define TEST_DOUBLE 1
+#define TEST_MATHVEC 1
+
+#ifndef __NO_MATH_INLINES
+# define __NO_MATH_INLINES
+#endif
+
+#define EXCEPTION_TESTS_double 0
+#define ROUNDING_TESTS_double(MODE) ((MODE) == FE_TONEAREST)
+
+#define CNCT(x, y) x ## y
+#define CONCAT(a, b) CNCT (a, b)
+
+#define VEC_SUFF _vlen4
+#define WRAPPER_NAME(function) CONCAT (function, VEC_SUFF)
+
+#define FUNC_TEST(function) function ## _VEC_SUFF
+
+#define WRAPPER_DECL(function) extern FLOAT function (FLOAT);
+#define WRAPPER_DECL_ff(function) extern FLOAT function (FLOAT, FLOAT);
+
+// Wrapper from scalar to vector function with vector length 4.
+#define VECTOR_WRAPPER(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE); \
+FLOAT scalar_func (FLOAT x) \
+{ \
+ int i; \
+ VEC_TYPE mx; \
+ INIT_VEC_LOOP (mx, x, 4); \
+ VEC_TYPE mr = vector_func (mx); \
+ TEST_VEC_LOOP (4); \
+}
+
+// Wrapper from scalar 2 argument function to vector one.
+#define VECTOR_WRAPPER_ff(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE, VEC_TYPE); \
+FLOAT scalar_func (FLOAT x, FLOAT y) \
+{ \
+ int i; \
+ VEC_TYPE mx, my; \
+ INIT_VEC_LOOP (mx, x, 4); \
+ INIT_VEC_LOOP (my, y, 4); \
+ VEC_TYPE mr = vector_func (mx, my); \
+ TEST_VEC_LOOP (4); \
+}
diff --git a/math/test-double-vlen8.h b/math/test-double-vlen8.h
new file mode 100644
index 0000000..6780e81
--- /dev/null
+++ b/math/test-double-vlen8.h
@@ -0,0 +1,71 @@
+/* Definitions for double vector tests with vector length 8.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#define FLOAT double
+#define FUNC(function) function
+#define TEST_MSG "testing double vector math (without inline functions)\n"
+#define MATHCONST(x) x
+#define CHOOSE(Clongdouble,Cdouble,Cfloat,Cinlinelongdouble,Cinlinedouble,Cinlinefloat) Cdouble
+#define PRINTF_EXPR "e"
+#define PRINTF_XEXPR "a"
+#define PRINTF_NEXPR "f"
+#define TEST_DOUBLE 1
+#define TEST_MATHVEC 1
+
+#ifndef __NO_MATH_INLINES
+# define __NO_MATH_INLINES
+#endif
+
+#define EXCEPTION_TESTS_double 0
+#define ROUNDING_TESTS_double(MODE) ((MODE) == FE_TONEAREST)
+
+#define CNCT(x, y) x ## y
+#define CONCAT(a, b) CNCT (a, b)
+
+#define VEC_SUFF _vlen8
+#define WRAPPER_NAME(function) CONCAT (function, VEC_SUFF)
+
+#define FUNC_TEST(function) function ## _VEC_SUFF
+
+#define WRAPPER_DECL(function) extern FLOAT function (FLOAT);
+#define WRAPPER_DECL_ff(function) extern FLOAT function (FLOAT, FLOAT);
+
+// Wrapper from scalar to vector function with vector length 8.
+#define VECTOR_WRAPPER(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE); \
+FLOAT scalar_func (FLOAT x) \
+{ \
+ int i; \
+ VEC_TYPE mx; \
+ INIT_VEC_LOOP (mx, x, 8); \
+ VEC_TYPE mr = vector_func (mx); \
+ TEST_VEC_LOOP (8); \
+}
+
+// Wrapper from scalar 2 argument function to vector one.
+#define VECTOR_WRAPPER_ff(scalar_func, vector_func) \
+extern VEC_TYPE vector_func (VEC_TYPE, VEC_TYPE); \
+FLOAT scalar_func (FLOAT x, FLOAT y) \
+{ \
+ int i; \
+ VEC_TYPE mx, my; \
+ INIT_VEC_LOOP (mx, x, 8); \
+ INIT_VEC_LOOP (my, y, 8); \
+ VEC_TYPE mr = vector_func (mx, my); \
+ TEST_VEC_LOOP (8); \
+}
diff --git a/math/test-vec-loop.h b/math/test-vec-loop.h
new file mode 100644
index 0000000..1a76c3e
--- /dev/null
+++ b/math/test-vec-loop.h
@@ -0,0 +1,42 @@
+/* Loop macro used in vector math functions tests.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+/* This macro is used in VECTOR_WRAPPER macros for vector tests. */
+#define TEST_VEC_LOOP(len) \
+ do \
+ { \
+ for (i = 1; i < len; i++) \
+ { \
+ if ((FLOAT) mr[0] != (FLOAT) mr[i]) \
+ { \
+ return ((FLOAT) mr[0] + 0.1); \
+ } \
+ } \
+ return ((FLOAT) mr[0]); \
+ } \
+ while (0)
+
+#define INIT_VEC_LOOP(vec, val, len) \
+ do \
+ { \
+ for (i = 0; i < len; i++) \
+ { \
+ vec[i] = val; \
+ } \
+ } \
+ while (0)
diff --git a/sysdeps/x86_64/fpu/Makefile b/sysdeps/x86_64/fpu/Makefile
index 2f16323..2d5fcf8 100644
--- a/sysdeps/x86_64/fpu/Makefile
+++ b/sysdeps/x86_64/fpu/Makefile
@@ -3,3 +3,23 @@ libmvec-support += svml_d_cos2_core svml_d_cos4_core_avx \
svml_d_cos4_core svml_d_cos8_core \
svml_d_cos_data init-arch
endif
+
+# Variables for libmvec tests.
+ifeq ($(subdir),math)
+ifeq ($(build-mathvec),yes)
+libmvec-tests += double-vlen2 double-vlen4 double-vlen4-avx2
+
+ifeq (yes,$(config-cflags-avx512))
+libmvec-tests += double-vlen8
+endif
+
+double-vlen2-arch-ext-cflags = -msse4
+double-vlen4-arch-ext-cflags = -mavx
+double-vlen4-arch-ext2-cflags = -mavx2
+double-vlen8-arch-ext-cflags = -mavx512f
+
+CFLAGS-test-double-vlen4-avx2.c = $(libm-test-vec-cflags)
+CFLAGS-test-double-vlen4-avx2-wrappers.c = $(double-vlen4-arch-ext2-cflags)
+
+endif
+endif
diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps
index e60cc1b..0d1f6e8 100644
--- a/sysdeps/x86_64/fpu/libm-test-ulps
+++ b/sysdeps/x86_64/fpu/libm-test-ulps
@@ -965,6 +965,18 @@ idouble: 1
ildouble: 2
ldouble: 2
+Function: "cos_vlen2":
+double: 1
+
+Function: "cos_vlen4":
+double: 1
+
+Function: "cos_vlen4_avx2":
+double: 1
+
+Function: "cos_vlen8":
+double: 1
+
Function: "cosh":
double: 1
float: 1
diff --git a/sysdeps/x86_64/fpu/math-tests-arch.h b/sysdeps/x86_64/fpu/math-tests-arch.h
new file mode 100644
index 0000000..56714f5
--- /dev/null
+++ b/sysdeps/x86_64/fpu/math-tests-arch.h
@@ -0,0 +1,65 @@
+/* Runtime architecture check for math tests. x86_64 version.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#if defined REQUIRE_AVX2
+# include <init-arch.h>
+
+ /* Set to 1 if AVX2 supported. */
+ static int avx2_usable;
+
+# define INIT_ARCH_EXT \
+ do \
+ { \
+ __init_cpu_features (); \
+ avx2_usable = __cpu_features.feature[index_AVX2_Usable] \
+ & bit_AVX2_Usable; \
+ } \
+ while (0)
+
+# define CHECK_ARCH_EXT \
+ do \
+ { \
+ if (!avx2_usable) return; \
+ } \
+ while (0)
+
+#elif defined REQUIRE_AVX512F
+# include <init-arch.h>
+
+ /* Set to 1 if supported. */
+ static int avx512f_usable;
+
+# define INIT_ARCH_EXT \
+ do \
+ { \
+ __init_cpu_features (); \
+ avx512f_usable = __cpu_features.feature[index_AVX512F_Usable] \
+ & bit_AVX512F_Usable; \
+ } \
+ while (0)
+
+# define CHECK_ARCH_EXT \
+ do \
+ { \
+ if (!avx512f_usable) return; \
+ } \
+ while (0)
+
+#else
+# include <sysdeps/generic/math-tests-arch.h>
+#endif
diff --git a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
new file mode 100644
index 0000000..bcfccb9
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
@@ -0,0 +1,25 @@
+/* Wrapper part of tests for SSE ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen2.h"
+#include "test-vec-loop.h"
+#include <immintrin.h>
+
+#define VEC_TYPE __m128d
+
+VECTOR_WRAPPER (WRAPPER_NAME (cos), _ZGVbN2v_cos)
diff --git a/sysdeps/x86_64/fpu/test-double-vlen2.c b/sysdeps/x86_64/fpu/test-double-vlen2.c
new file mode 100644
index 0000000..03e2046
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen2.c
@@ -0,0 +1,23 @@
+/* Tests for SSE ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen2.h"
+
+#define TEST_VECTOR_cos 1
+
+#include "libm-test.c"
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
new file mode 100644
index 0000000..69e3fb1
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
@@ -0,0 +1,28 @@
+/* Wrapper part of tests for AVX2 ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen4.h"
+#include "test-vec-loop.h"
+#include <immintrin.h>
+
+#undef VEC_SUFF
+#define VEC_SUFF _vlen4_avx2
+
+#define VEC_TYPE __m256d
+
+VECTOR_WRAPPER (WRAPPER_NAME (cos), _ZGVdN4v_cos)
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-avx2.c b/sysdeps/x86_64/fpu/test-double-vlen4-avx2.c
new file mode 100644
index 0000000..fcc4b64
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen4-avx2.c
@@ -0,0 +1,28 @@
+/* Tests for AVX2 ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen4.h"
+
+#undef VEC_SUFF
+#define VEC_SUFF _vlen4_avx2
+
+#define TEST_VECTOR_cos 1
+
+#define REQUIRE_AVX2
+
+#include "libm-test.c"
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
new file mode 100644
index 0000000..cd9c0bb
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
@@ -0,0 +1,25 @@
+/* Wrapper part of tests for AVX ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen4.h"
+#include "test-vec-loop.h"
+#include <immintrin.h>
+
+#define VEC_TYPE __m256d
+
+VECTOR_WRAPPER (WRAPPER_NAME (cos), _ZGVcN4v_cos)
diff --git a/sysdeps/x86_64/fpu/test-double-vlen4.c b/sysdeps/x86_64/fpu/test-double-vlen4.c
new file mode 100644
index 0000000..dd77b70
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen4.c
@@ -0,0 +1,23 @@
+/* Tests for AVX ISA versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen4.h"
+
+#define TEST_VECTOR_cos 1
+
+#include "libm-test.c"
diff --git a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
new file mode 100644
index 0000000..9381360
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
@@ -0,0 +1,25 @@
+/* Wrapper part of tests for AVX-512 versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen8.h"
+#include "test-vec-loop.h"
+#include <immintrin.h>
+
+#define VEC_TYPE __m512d
+
+VECTOR_WRAPPER (WRAPPER_NAME (cos), _ZGVeN8v_cos)
diff --git a/sysdeps/x86_64/fpu/test-double-vlen8.c b/sysdeps/x86_64/fpu/test-double-vlen8.c
new file mode 100644
index 0000000..e4da2bd
--- /dev/null
+++ b/sysdeps/x86_64/fpu/test-double-vlen8.c
@@ -0,0 +1,25 @@
+/* Tests for AVX-512 versions of vector math functions.
+ Copyright (C) 2014-2015 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include "test-double-vlen8.h"
+
+#define TEST_VECTOR_cos 1
+
+#define REQUIRE_AVX512F
+
+#include "libm-test.c"
-----------------------------------------------------------------------
Summary of changes:
ChangeLog | 19 +++++
math/Makefile | 33 +++++++--
math/gen-libm-have-vector-test.sh | 9 ++-
math/test-double-vlen2.h | 71 ++++++++++++++++++++
math/test-double-vlen4.h | 71 ++++++++++++++++++++
math/test-double-vlen8.h | 71 ++++++++++++++++++++
math/test-vec-loop.h | 42 ++++++++++++
sysdeps/x86_64/fpu/Makefile | 20 ++++++
sysdeps/x86_64/fpu/libm-test-ulps | 12 +++
sysdeps/x86_64/fpu/math-tests-arch.h | 65 ++++++++++++++++++
sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c | 25 +++++++
sysdeps/x86_64/fpu/test-double-vlen2.c | 23 ++++++
.../x86_64/fpu/test-double-vlen4-avx2-wrappers.c | 28 ++++++++
sysdeps/x86_64/fpu/test-double-vlen4-avx2.c | 28 ++++++++
sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c | 25 +++++++
sysdeps/x86_64/fpu/test-double-vlen4.c | 23 ++++++
sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c | 25 +++++++
sysdeps/x86_64/fpu/test-double-vlen8.c | 25 +++++++
18 files changed, 606 insertions(+), 9 deletions(-)
create mode 100644 math/test-double-vlen2.h
create mode 100644 math/test-double-vlen4.h
create mode 100644 math/test-double-vlen8.h
create mode 100644 math/test-vec-loop.h
create mode 100644 sysdeps/x86_64/fpu/math-tests-arch.h
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen2.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen4-avx2.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen4.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c
create mode 100644 sysdeps/x86_64/fpu/test-double-vlen8.c
hooks/post-receive
--
GNU C Library master sources