objective-ci 0.0.2 → 0.0.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +15 -0
- data/README.md +1 -3
- data/Rakefile +1 -1
- data/bin/oclint +4 -0
- data/bin/oclint-0.8 +1 -1
- data/bin/oclint-json-compilation-database +1 -1
- data/bin/oclint-xcodebuild +1 -1
- data/externals/oclint/LICENSE +69 -0
- data/externals/oclint/bin/oclint +0 -0
- data/externals/oclint/{oclint-0.8 → bin/oclint-0.8} +0 -0
- data/externals/oclint/{oclint-json-compilation-database → bin/oclint-json-compilation-database} +0 -0
- data/externals/oclint/{oclint-xcodebuild → bin/oclint-xcodebuild} +0 -0
- data/externals/oclint/lib/clang/3.4/asan_blacklist.txt +10 -0
- data/externals/oclint/lib/clang/3.4/include/Intrin.h +784 -0
- data/externals/oclint/lib/clang/3.4/include/__wmmintrin_aes.h +67 -0
- data/externals/oclint/lib/clang/3.4/include/__wmmintrin_pclmul.h +34 -0
- data/externals/oclint/lib/clang/3.4/include/altivec.h +11856 -0
- data/externals/oclint/lib/clang/3.4/include/ammintrin.h +68 -0
- data/externals/oclint/lib/clang/3.4/include/arm_neon.h +6802 -0
- data/externals/oclint/lib/clang/3.4/include/avx2intrin.h +1206 -0
- data/externals/oclint/lib/clang/3.4/include/avxintrin.h +1224 -0
- data/externals/oclint/lib/clang/3.4/include/bmi2intrin.h +94 -0
- data/externals/oclint/lib/clang/3.4/include/bmiintrin.h +115 -0
- data/externals/oclint/lib/clang/3.4/include/cpuid.h +156 -0
- data/externals/oclint/lib/clang/3.4/include/emmintrin.h +1451 -0
- data/externals/oclint/lib/clang/3.4/include/f16cintrin.h +58 -0
- data/externals/oclint/lib/clang/3.4/include/float.h +124 -0
- data/externals/oclint/lib/clang/3.4/include/fma4intrin.h +231 -0
- data/externals/oclint/lib/clang/3.4/include/fmaintrin.h +229 -0
- data/externals/oclint/lib/clang/3.4/include/immintrin.h +118 -0
- data/externals/oclint/lib/clang/3.4/include/iso646.h +43 -0
- data/externals/oclint/lib/clang/3.4/include/limits.h +119 -0
- data/externals/oclint/lib/clang/3.4/include/lzcntintrin.h +55 -0
- data/externals/oclint/lib/clang/3.4/include/mm3dnow.h +162 -0
- data/externals/oclint/lib/clang/3.4/include/mm_malloc.h +75 -0
- data/externals/oclint/lib/clang/3.4/include/mmintrin.h +503 -0
- data/externals/oclint/lib/clang/3.4/include/module.map +156 -0
- data/externals/oclint/lib/clang/3.4/include/nmmintrin.h +35 -0
- data/externals/oclint/lib/clang/3.4/include/pmmintrin.h +117 -0
- data/externals/oclint/lib/clang/3.4/include/popcntintrin.h +45 -0
- data/externals/oclint/lib/clang/3.4/include/prfchwintrin.h +39 -0
- data/externals/oclint/lib/clang/3.4/include/rdseedintrin.h +52 -0
- data/externals/oclint/lib/clang/3.4/include/rtmintrin.h +54 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/asan_interface.h +137 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/common_interface_defs.h +54 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/dfsan_interface.h +87 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/linux_syscall_hooks.h +3070 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/lsan_interface.h +52 -0
- data/externals/oclint/lib/clang/3.4/include/sanitizer/msan_interface.h +162 -0
- data/externals/oclint/lib/clang/3.4/include/shaintrin.h +74 -0
- data/externals/oclint/lib/clang/3.4/include/smmintrin.h +468 -0
- data/externals/oclint/lib/clang/3.4/include/stdalign.h +35 -0
- data/externals/oclint/lib/clang/3.4/include/stdarg.h +50 -0
- data/externals/oclint/lib/clang/3.4/include/stdbool.h +44 -0
- data/externals/oclint/lib/clang/3.4/include/stddef.h +102 -0
- data/externals/oclint/lib/clang/3.4/include/stdint.h +708 -0
- data/externals/oclint/lib/clang/3.4/include/stdnoreturn.h +30 -0
- data/externals/oclint/lib/clang/3.4/include/tbmintrin.h +158 -0
- data/externals/oclint/lib/clang/3.4/include/tgmath.h +1374 -0
- data/externals/oclint/lib/clang/3.4/include/tmmintrin.h +225 -0
- data/externals/oclint/lib/clang/3.4/include/unwind.h +280 -0
- data/externals/oclint/lib/clang/3.4/include/varargs.h +26 -0
- data/externals/oclint/lib/clang/3.4/include/wmmintrin.h +42 -0
- data/externals/oclint/lib/clang/3.4/include/x86intrin.h +79 -0
- data/externals/oclint/lib/clang/3.4/include/xmmintrin.h +1001 -0
- data/externals/oclint/lib/clang/3.4/include/xopintrin.h +804 -0
- data/externals/oclint/lib/clang/3.4/lib/darwin/libclang_rt.asan_osx_dynamic.dylib +0 -0
- data/externals/oclint/lib/clang/3.4/lib/darwin/libclang_rt.i386.a +0 -0
- data/externals/oclint/lib/clang/3.4/lib/darwin/libclang_rt.profile_osx.a +0 -0
- data/externals/oclint/lib/clang/3.4/lib/darwin/libclang_rt.ubsan_osx.a +0 -0
- data/externals/oclint/lib/clang/3.4/lib/darwin/libclang_rt.x86_64.a +0 -0
- data/externals/oclint/lib/oclint/reporters/libHTMLReporter.dylib +0 -0
- data/externals/oclint/lib/oclint/reporters/libJSONReporter.dylib +0 -0
- data/externals/oclint/lib/oclint/reporters/libPMDReporter.dylib +0 -0
- data/externals/oclint/lib/oclint/reporters/libTextReporter.dylib +0 -0
- data/externals/oclint/lib/oclint/reporters/libXMLReporter.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libAvoidBranchingStatementAsLastInLoopRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libBitwiseOperatorInConditionalRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libBrokenNullCheckRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libBrokenOddnessCheckRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libCollapsibleIfStatementsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libConstantConditionalOperatorRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libConstantIfExpressionRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libCyclomaticComplexityRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libDeadCodeRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libDefaultLabelNotLastInSwitchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libDoubleNegativeRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyCatchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyDoWhileStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyElseBlockRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyFinallyStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyForStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyIfStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptySwitchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyTryStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libEmptyWhileStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libForLoopShouldBeWhileLoopRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libGotoStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libInvertedLogicRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libJumbledIncrementerRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libLongClassRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libLongLineRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libLongMethodRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libLongVariableNameRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libMisplacedNullCheckRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libMissingBreakInSwitchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libMultipleUnaryOperatorRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libNPathComplexityRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libNcssMethodCountRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libNestedBlockDepthRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libNonCaseLabelInSwitchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libObjCBoxedExpressionsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libObjCContainerLiteralsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libObjCNSNumberLiteralsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libObjCObjectSubscriptingRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libParameterReassignmentRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libRedundantConditionalOperatorRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libRedundantIfStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libRedundantLocalVariableRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libRedundantNilCheckRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libReturnFromFinallyBlockRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libShortVariableNameRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libSwitchStatementsShouldHaveDefaultRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libThrowExceptionFromFinallyBlockRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libTooFewBranchesInSwitchStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libTooManyFieldsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libTooManyMethodsRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libTooManyParametersRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libUnnecessaryElseStatementRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libUnnecessaryNullCheckForCXXDeallocRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libUnusedLocalVariableRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libUnusedMethodParameterRule.dylib +0 -0
- data/externals/oclint/lib/oclint/rules/libUselessParenthesesRule.dylib +0 -0
- data/lib/objective_ci/ci_tasks.rb +1 -1
- data/lib/objective_ci/version.rb +1 -1
- metadata +200 -84
@@ -0,0 +1,26 @@
|
|
1
|
+
/*===---- varargs.h - Variable argument handling -------------------------------------===
|
2
|
+
*
|
3
|
+
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
4
|
+
* of this software and associated documentation files (the "Software"), to deal
|
5
|
+
* in the Software without restriction, including without limitation the rights
|
6
|
+
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
7
|
+
* copies of the Software, and to permit persons to whom the Software is
|
8
|
+
* furnished to do so, subject to the following conditions:
|
9
|
+
*
|
10
|
+
* The above copyright notice and this permission notice shall be included in
|
11
|
+
* all copies or substantial portions of the Software.
|
12
|
+
*
|
13
|
+
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
14
|
+
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
15
|
+
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
16
|
+
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
17
|
+
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
18
|
+
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
19
|
+
* THE SOFTWARE.
|
20
|
+
*
|
21
|
+
*===-----------------------------------------------------------------------===
|
22
|
+
*/
|
23
|
+
#ifndef __VARARGS_H
|
24
|
+
#define __VARARGS_H
|
25
|
+
#error "Please use <stdarg.h> instead of <varargs.h>"
|
26
|
+
#endif
|
@@ -0,0 +1,42 @@
|
|
1
|
+
/*===---- wmmintrin.h - AES intrinsics ------------------------------------===
|
2
|
+
*
|
3
|
+
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
4
|
+
* of this software and associated documentation files (the "Software"), to deal
|
5
|
+
* in the Software without restriction, including without limitation the rights
|
6
|
+
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
7
|
+
* copies of the Software, and to permit persons to whom the Software is
|
8
|
+
* furnished to do so, subject to the following conditions:
|
9
|
+
*
|
10
|
+
* The above copyright notice and this permission notice shall be included in
|
11
|
+
* all copies or substantial portions of the Software.
|
12
|
+
*
|
13
|
+
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
14
|
+
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
15
|
+
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
16
|
+
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
17
|
+
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
18
|
+
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
19
|
+
* THE SOFTWARE.
|
20
|
+
*
|
21
|
+
*===-----------------------------------------------------------------------===
|
22
|
+
*/
|
23
|
+
|
24
|
+
#ifndef _WMMINTRIN_H
|
25
|
+
#define _WMMINTRIN_H
|
26
|
+
|
27
|
+
#include <emmintrin.h>
|
28
|
+
|
29
|
+
#if !defined (__AES__) && !defined (__PCLMUL__)
|
30
|
+
# error "AES/PCLMUL instructions not enabled"
|
31
|
+
#else
|
32
|
+
|
33
|
+
#ifdef __AES__
|
34
|
+
#include <__wmmintrin_aes.h>
|
35
|
+
#endif /* __AES__ */
|
36
|
+
|
37
|
+
#ifdef __PCLMUL__
|
38
|
+
#include <__wmmintrin_pclmul.h>
|
39
|
+
#endif /* __PCLMUL__ */
|
40
|
+
|
41
|
+
#endif /* __AES__ || __PCLMUL__ */
|
42
|
+
#endif /* _WMMINTRIN_H */
|
@@ -0,0 +1,79 @@
|
|
1
|
+
/*===---- x86intrin.h - X86 intrinsics -------------------------------------===
|
2
|
+
*
|
3
|
+
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
4
|
+
* of this software and associated documentation files (the "Software"), to deal
|
5
|
+
* in the Software without restriction, including without limitation the rights
|
6
|
+
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
7
|
+
* copies of the Software, and to permit persons to whom the Software is
|
8
|
+
* furnished to do so, subject to the following conditions:
|
9
|
+
*
|
10
|
+
* The above copyright notice and this permission notice shall be included in
|
11
|
+
* all copies or substantial portions of the Software.
|
12
|
+
*
|
13
|
+
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
14
|
+
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
15
|
+
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
16
|
+
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
17
|
+
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
18
|
+
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
19
|
+
* THE SOFTWARE.
|
20
|
+
*
|
21
|
+
*===-----------------------------------------------------------------------===
|
22
|
+
*/
|
23
|
+
|
24
|
+
#ifndef __X86INTRIN_H
|
25
|
+
#define __X86INTRIN_H
|
26
|
+
|
27
|
+
#include <immintrin.h>
|
28
|
+
|
29
|
+
#ifdef __3dNOW__
|
30
|
+
#include <mm3dnow.h>
|
31
|
+
#endif
|
32
|
+
|
33
|
+
#ifdef __BMI__
|
34
|
+
#include <bmiintrin.h>
|
35
|
+
#endif
|
36
|
+
|
37
|
+
#ifdef __BMI2__
|
38
|
+
#include <bmi2intrin.h>
|
39
|
+
#endif
|
40
|
+
|
41
|
+
#ifdef __LZCNT__
|
42
|
+
#include <lzcntintrin.h>
|
43
|
+
#endif
|
44
|
+
|
45
|
+
#ifdef __POPCNT__
|
46
|
+
#include <popcntintrin.h>
|
47
|
+
#endif
|
48
|
+
|
49
|
+
#ifdef __RDSEED__
|
50
|
+
#include <rdseedintrin.h>
|
51
|
+
#endif
|
52
|
+
|
53
|
+
#ifdef __PRFCHW__
|
54
|
+
#include <prfchwintrin.h>
|
55
|
+
#endif
|
56
|
+
|
57
|
+
#ifdef __SSE4A__
|
58
|
+
#include <ammintrin.h>
|
59
|
+
#endif
|
60
|
+
|
61
|
+
#ifdef __FMA4__
|
62
|
+
#include <fma4intrin.h>
|
63
|
+
#endif
|
64
|
+
|
65
|
+
#ifdef __XOP__
|
66
|
+
#include <xopintrin.h>
|
67
|
+
#endif
|
68
|
+
|
69
|
+
#ifdef __TBM__
|
70
|
+
#include <tbmintrin.h>
|
71
|
+
#endif
|
72
|
+
|
73
|
+
#ifdef __F16C__
|
74
|
+
#include <f16cintrin.h>
|
75
|
+
#endif
|
76
|
+
|
77
|
+
// FIXME: LWP
|
78
|
+
|
79
|
+
#endif /* __X86INTRIN_H */
|
@@ -0,0 +1,1001 @@
|
|
1
|
+
/*===---- xmmintrin.h - SSE intrinsics -------------------------------------===
|
2
|
+
*
|
3
|
+
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
4
|
+
* of this software and associated documentation files (the "Software"), to deal
|
5
|
+
* in the Software without restriction, including without limitation the rights
|
6
|
+
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
7
|
+
* copies of the Software, and to permit persons to whom the Software is
|
8
|
+
* furnished to do so, subject to the following conditions:
|
9
|
+
*
|
10
|
+
* The above copyright notice and this permission notice shall be included in
|
11
|
+
* all copies or substantial portions of the Software.
|
12
|
+
*
|
13
|
+
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
14
|
+
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
15
|
+
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
16
|
+
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
17
|
+
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
18
|
+
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
|
19
|
+
* THE SOFTWARE.
|
20
|
+
*
|
21
|
+
*===-----------------------------------------------------------------------===
|
22
|
+
*/
|
23
|
+
|
24
|
+
#ifndef __XMMINTRIN_H
|
25
|
+
#define __XMMINTRIN_H
|
26
|
+
|
27
|
+
#ifndef __SSE__
|
28
|
+
#error "SSE instruction set not enabled"
|
29
|
+
#else
|
30
|
+
|
31
|
+
#include <mmintrin.h>
|
32
|
+
|
33
|
+
typedef int __v4si __attribute__((__vector_size__(16)));
|
34
|
+
typedef float __v4sf __attribute__((__vector_size__(16)));
|
35
|
+
typedef float __m128 __attribute__((__vector_size__(16)));
|
36
|
+
|
37
|
+
// This header should only be included in a hosted environment as it depends on
|
38
|
+
// a standard library to provide allocation routines.
|
39
|
+
#if __STDC_HOSTED__
|
40
|
+
#include <mm_malloc.h>
|
41
|
+
#endif
|
42
|
+
|
43
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
44
|
+
_mm_add_ss(__m128 __a, __m128 __b)
|
45
|
+
{
|
46
|
+
__a[0] += __b[0];
|
47
|
+
return __a;
|
48
|
+
}
|
49
|
+
|
50
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
51
|
+
_mm_add_ps(__m128 __a, __m128 __b)
|
52
|
+
{
|
53
|
+
return __a + __b;
|
54
|
+
}
|
55
|
+
|
56
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
57
|
+
_mm_sub_ss(__m128 __a, __m128 __b)
|
58
|
+
{
|
59
|
+
__a[0] -= __b[0];
|
60
|
+
return __a;
|
61
|
+
}
|
62
|
+
|
63
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
64
|
+
_mm_sub_ps(__m128 __a, __m128 __b)
|
65
|
+
{
|
66
|
+
return __a - __b;
|
67
|
+
}
|
68
|
+
|
69
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
70
|
+
_mm_mul_ss(__m128 __a, __m128 __b)
|
71
|
+
{
|
72
|
+
__a[0] *= __b[0];
|
73
|
+
return __a;
|
74
|
+
}
|
75
|
+
|
76
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
77
|
+
_mm_mul_ps(__m128 __a, __m128 __b)
|
78
|
+
{
|
79
|
+
return __a * __b;
|
80
|
+
}
|
81
|
+
|
82
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
83
|
+
_mm_div_ss(__m128 __a, __m128 __b)
|
84
|
+
{
|
85
|
+
__a[0] /= __b[0];
|
86
|
+
return __a;
|
87
|
+
}
|
88
|
+
|
89
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
90
|
+
_mm_div_ps(__m128 __a, __m128 __b)
|
91
|
+
{
|
92
|
+
return __a / __b;
|
93
|
+
}
|
94
|
+
|
95
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
96
|
+
_mm_sqrt_ss(__m128 __a)
|
97
|
+
{
|
98
|
+
__m128 __c = __builtin_ia32_sqrtss(__a);
|
99
|
+
return (__m128) { __c[0], __a[1], __a[2], __a[3] };
|
100
|
+
}
|
101
|
+
|
102
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
103
|
+
_mm_sqrt_ps(__m128 __a)
|
104
|
+
{
|
105
|
+
return __builtin_ia32_sqrtps(__a);
|
106
|
+
}
|
107
|
+
|
108
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
109
|
+
_mm_rcp_ss(__m128 __a)
|
110
|
+
{
|
111
|
+
__m128 __c = __builtin_ia32_rcpss(__a);
|
112
|
+
return (__m128) { __c[0], __a[1], __a[2], __a[3] };
|
113
|
+
}
|
114
|
+
|
115
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
116
|
+
_mm_rcp_ps(__m128 __a)
|
117
|
+
{
|
118
|
+
return __builtin_ia32_rcpps(__a);
|
119
|
+
}
|
120
|
+
|
121
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
122
|
+
_mm_rsqrt_ss(__m128 __a)
|
123
|
+
{
|
124
|
+
__m128 __c = __builtin_ia32_rsqrtss(__a);
|
125
|
+
return (__m128) { __c[0], __a[1], __a[2], __a[3] };
|
126
|
+
}
|
127
|
+
|
128
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
129
|
+
_mm_rsqrt_ps(__m128 __a)
|
130
|
+
{
|
131
|
+
return __builtin_ia32_rsqrtps(__a);
|
132
|
+
}
|
133
|
+
|
134
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
135
|
+
_mm_min_ss(__m128 __a, __m128 __b)
|
136
|
+
{
|
137
|
+
return __builtin_ia32_minss(__a, __b);
|
138
|
+
}
|
139
|
+
|
140
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
141
|
+
_mm_min_ps(__m128 __a, __m128 __b)
|
142
|
+
{
|
143
|
+
return __builtin_ia32_minps(__a, __b);
|
144
|
+
}
|
145
|
+
|
146
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
147
|
+
_mm_max_ss(__m128 __a, __m128 __b)
|
148
|
+
{
|
149
|
+
return __builtin_ia32_maxss(__a, __b);
|
150
|
+
}
|
151
|
+
|
152
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
153
|
+
_mm_max_ps(__m128 __a, __m128 __b)
|
154
|
+
{
|
155
|
+
return __builtin_ia32_maxps(__a, __b);
|
156
|
+
}
|
157
|
+
|
158
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
159
|
+
_mm_and_ps(__m128 __a, __m128 __b)
|
160
|
+
{
|
161
|
+
return (__m128)((__v4si)__a & (__v4si)__b);
|
162
|
+
}
|
163
|
+
|
164
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
165
|
+
_mm_andnot_ps(__m128 __a, __m128 __b)
|
166
|
+
{
|
167
|
+
return (__m128)(~(__v4si)__a & (__v4si)__b);
|
168
|
+
}
|
169
|
+
|
170
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
171
|
+
_mm_or_ps(__m128 __a, __m128 __b)
|
172
|
+
{
|
173
|
+
return (__m128)((__v4si)__a | (__v4si)__b);
|
174
|
+
}
|
175
|
+
|
176
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
177
|
+
_mm_xor_ps(__m128 __a, __m128 __b)
|
178
|
+
{
|
179
|
+
return (__m128)((__v4si)__a ^ (__v4si)__b);
|
180
|
+
}
|
181
|
+
|
182
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
183
|
+
_mm_cmpeq_ss(__m128 __a, __m128 __b)
|
184
|
+
{
|
185
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 0);
|
186
|
+
}
|
187
|
+
|
188
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
189
|
+
_mm_cmpeq_ps(__m128 __a, __m128 __b)
|
190
|
+
{
|
191
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 0);
|
192
|
+
}
|
193
|
+
|
194
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
195
|
+
_mm_cmplt_ss(__m128 __a, __m128 __b)
|
196
|
+
{
|
197
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 1);
|
198
|
+
}
|
199
|
+
|
200
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
201
|
+
_mm_cmplt_ps(__m128 __a, __m128 __b)
|
202
|
+
{
|
203
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 1);
|
204
|
+
}
|
205
|
+
|
206
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
207
|
+
_mm_cmple_ss(__m128 __a, __m128 __b)
|
208
|
+
{
|
209
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 2);
|
210
|
+
}
|
211
|
+
|
212
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
213
|
+
_mm_cmple_ps(__m128 __a, __m128 __b)
|
214
|
+
{
|
215
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 2);
|
216
|
+
}
|
217
|
+
|
218
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
219
|
+
_mm_cmpgt_ss(__m128 __a, __m128 __b)
|
220
|
+
{
|
221
|
+
return (__m128)__builtin_shufflevector(__a,
|
222
|
+
__builtin_ia32_cmpss(__b, __a, 1),
|
223
|
+
4, 1, 2, 3);
|
224
|
+
}
|
225
|
+
|
226
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
227
|
+
_mm_cmpgt_ps(__m128 __a, __m128 __b)
|
228
|
+
{
|
229
|
+
return (__m128)__builtin_ia32_cmpps(__b, __a, 1);
|
230
|
+
}
|
231
|
+
|
232
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
233
|
+
_mm_cmpge_ss(__m128 __a, __m128 __b)
|
234
|
+
{
|
235
|
+
return (__m128)__builtin_shufflevector(__a,
|
236
|
+
__builtin_ia32_cmpss(__b, __a, 2),
|
237
|
+
4, 1, 2, 3);
|
238
|
+
}
|
239
|
+
|
240
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
241
|
+
_mm_cmpge_ps(__m128 __a, __m128 __b)
|
242
|
+
{
|
243
|
+
return (__m128)__builtin_ia32_cmpps(__b, __a, 2);
|
244
|
+
}
|
245
|
+
|
246
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
247
|
+
_mm_cmpneq_ss(__m128 __a, __m128 __b)
|
248
|
+
{
|
249
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 4);
|
250
|
+
}
|
251
|
+
|
252
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
253
|
+
_mm_cmpneq_ps(__m128 __a, __m128 __b)
|
254
|
+
{
|
255
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 4);
|
256
|
+
}
|
257
|
+
|
258
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
259
|
+
_mm_cmpnlt_ss(__m128 __a, __m128 __b)
|
260
|
+
{
|
261
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 5);
|
262
|
+
}
|
263
|
+
|
264
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
265
|
+
_mm_cmpnlt_ps(__m128 __a, __m128 __b)
|
266
|
+
{
|
267
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 5);
|
268
|
+
}
|
269
|
+
|
270
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
271
|
+
_mm_cmpnle_ss(__m128 __a, __m128 __b)
|
272
|
+
{
|
273
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 6);
|
274
|
+
}
|
275
|
+
|
276
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
277
|
+
_mm_cmpnle_ps(__m128 __a, __m128 __b)
|
278
|
+
{
|
279
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 6);
|
280
|
+
}
|
281
|
+
|
282
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
283
|
+
_mm_cmpngt_ss(__m128 __a, __m128 __b)
|
284
|
+
{
|
285
|
+
return (__m128)__builtin_shufflevector(__a,
|
286
|
+
__builtin_ia32_cmpss(__b, __a, 5),
|
287
|
+
4, 1, 2, 3);
|
288
|
+
}
|
289
|
+
|
290
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
291
|
+
_mm_cmpngt_ps(__m128 __a, __m128 __b)
|
292
|
+
{
|
293
|
+
return (__m128)__builtin_ia32_cmpps(__b, __a, 5);
|
294
|
+
}
|
295
|
+
|
296
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
297
|
+
_mm_cmpnge_ss(__m128 __a, __m128 __b)
|
298
|
+
{
|
299
|
+
return (__m128)__builtin_shufflevector(__a,
|
300
|
+
__builtin_ia32_cmpss(__b, __a, 6),
|
301
|
+
4, 1, 2, 3);
|
302
|
+
}
|
303
|
+
|
304
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
305
|
+
_mm_cmpnge_ps(__m128 __a, __m128 __b)
|
306
|
+
{
|
307
|
+
return (__m128)__builtin_ia32_cmpps(__b, __a, 6);
|
308
|
+
}
|
309
|
+
|
310
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
311
|
+
_mm_cmpord_ss(__m128 __a, __m128 __b)
|
312
|
+
{
|
313
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 7);
|
314
|
+
}
|
315
|
+
|
316
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
317
|
+
_mm_cmpord_ps(__m128 __a, __m128 __b)
|
318
|
+
{
|
319
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 7);
|
320
|
+
}
|
321
|
+
|
322
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
323
|
+
_mm_cmpunord_ss(__m128 __a, __m128 __b)
|
324
|
+
{
|
325
|
+
return (__m128)__builtin_ia32_cmpss(__a, __b, 3);
|
326
|
+
}
|
327
|
+
|
328
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
329
|
+
_mm_cmpunord_ps(__m128 __a, __m128 __b)
|
330
|
+
{
|
331
|
+
return (__m128)__builtin_ia32_cmpps(__a, __b, 3);
|
332
|
+
}
|
333
|
+
|
334
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
335
|
+
_mm_comieq_ss(__m128 __a, __m128 __b)
|
336
|
+
{
|
337
|
+
return __builtin_ia32_comieq(__a, __b);
|
338
|
+
}
|
339
|
+
|
340
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
341
|
+
_mm_comilt_ss(__m128 __a, __m128 __b)
|
342
|
+
{
|
343
|
+
return __builtin_ia32_comilt(__a, __b);
|
344
|
+
}
|
345
|
+
|
346
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
347
|
+
_mm_comile_ss(__m128 __a, __m128 __b)
|
348
|
+
{
|
349
|
+
return __builtin_ia32_comile(__a, __b);
|
350
|
+
}
|
351
|
+
|
352
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
353
|
+
_mm_comigt_ss(__m128 __a, __m128 __b)
|
354
|
+
{
|
355
|
+
return __builtin_ia32_comigt(__a, __b);
|
356
|
+
}
|
357
|
+
|
358
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
359
|
+
_mm_comige_ss(__m128 __a, __m128 __b)
|
360
|
+
{
|
361
|
+
return __builtin_ia32_comige(__a, __b);
|
362
|
+
}
|
363
|
+
|
364
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
365
|
+
_mm_comineq_ss(__m128 __a, __m128 __b)
|
366
|
+
{
|
367
|
+
return __builtin_ia32_comineq(__a, __b);
|
368
|
+
}
|
369
|
+
|
370
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
371
|
+
_mm_ucomieq_ss(__m128 __a, __m128 __b)
|
372
|
+
{
|
373
|
+
return __builtin_ia32_ucomieq(__a, __b);
|
374
|
+
}
|
375
|
+
|
376
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
377
|
+
_mm_ucomilt_ss(__m128 __a, __m128 __b)
|
378
|
+
{
|
379
|
+
return __builtin_ia32_ucomilt(__a, __b);
|
380
|
+
}
|
381
|
+
|
382
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
383
|
+
_mm_ucomile_ss(__m128 __a, __m128 __b)
|
384
|
+
{
|
385
|
+
return __builtin_ia32_ucomile(__a, __b);
|
386
|
+
}
|
387
|
+
|
388
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
389
|
+
_mm_ucomigt_ss(__m128 __a, __m128 __b)
|
390
|
+
{
|
391
|
+
return __builtin_ia32_ucomigt(__a, __b);
|
392
|
+
}
|
393
|
+
|
394
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
395
|
+
_mm_ucomige_ss(__m128 __a, __m128 __b)
|
396
|
+
{
|
397
|
+
return __builtin_ia32_ucomige(__a, __b);
|
398
|
+
}
|
399
|
+
|
400
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
401
|
+
_mm_ucomineq_ss(__m128 __a, __m128 __b)
|
402
|
+
{
|
403
|
+
return __builtin_ia32_ucomineq(__a, __b);
|
404
|
+
}
|
405
|
+
|
406
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
407
|
+
_mm_cvtss_si32(__m128 __a)
|
408
|
+
{
|
409
|
+
return __builtin_ia32_cvtss2si(__a);
|
410
|
+
}
|
411
|
+
|
412
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
413
|
+
_mm_cvt_ss2si(__m128 __a)
|
414
|
+
{
|
415
|
+
return _mm_cvtss_si32(__a);
|
416
|
+
}
|
417
|
+
|
418
|
+
#ifdef __x86_64__
|
419
|
+
|
420
|
+
static __inline__ long long __attribute__((__always_inline__, __nodebug__))
|
421
|
+
_mm_cvtss_si64(__m128 __a)
|
422
|
+
{
|
423
|
+
return __builtin_ia32_cvtss2si64(__a);
|
424
|
+
}
|
425
|
+
|
426
|
+
#endif
|
427
|
+
|
428
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
429
|
+
_mm_cvtps_pi32(__m128 __a)
|
430
|
+
{
|
431
|
+
return (__m64)__builtin_ia32_cvtps2pi(__a);
|
432
|
+
}
|
433
|
+
|
434
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
435
|
+
_mm_cvt_ps2pi(__m128 __a)
|
436
|
+
{
|
437
|
+
return _mm_cvtps_pi32(__a);
|
438
|
+
}
|
439
|
+
|
440
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
441
|
+
_mm_cvttss_si32(__m128 __a)
|
442
|
+
{
|
443
|
+
return __a[0];
|
444
|
+
}
|
445
|
+
|
446
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
447
|
+
_mm_cvtt_ss2si(__m128 __a)
|
448
|
+
{
|
449
|
+
return _mm_cvttss_si32(__a);
|
450
|
+
}
|
451
|
+
|
452
|
+
static __inline__ long long __attribute__((__always_inline__, __nodebug__))
|
453
|
+
_mm_cvttss_si64(__m128 __a)
|
454
|
+
{
|
455
|
+
return __a[0];
|
456
|
+
}
|
457
|
+
|
458
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
459
|
+
_mm_cvttps_pi32(__m128 __a)
|
460
|
+
{
|
461
|
+
return (__m64)__builtin_ia32_cvttps2pi(__a);
|
462
|
+
}
|
463
|
+
|
464
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
465
|
+
_mm_cvtt_ps2pi(__m128 __a)
|
466
|
+
{
|
467
|
+
return _mm_cvttps_pi32(__a);
|
468
|
+
}
|
469
|
+
|
470
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
471
|
+
_mm_cvtsi32_ss(__m128 __a, int __b)
|
472
|
+
{
|
473
|
+
__a[0] = __b;
|
474
|
+
return __a;
|
475
|
+
}
|
476
|
+
|
477
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
478
|
+
_mm_cvt_si2ss(__m128 __a, int __b)
|
479
|
+
{
|
480
|
+
return _mm_cvtsi32_ss(__a, __b);
|
481
|
+
}
|
482
|
+
|
483
|
+
#ifdef __x86_64__
|
484
|
+
|
485
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
486
|
+
_mm_cvtsi64_ss(__m128 __a, long long __b)
|
487
|
+
{
|
488
|
+
__a[0] = __b;
|
489
|
+
return __a;
|
490
|
+
}
|
491
|
+
|
492
|
+
#endif
|
493
|
+
|
494
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
495
|
+
_mm_cvtpi32_ps(__m128 __a, __m64 __b)
|
496
|
+
{
|
497
|
+
return __builtin_ia32_cvtpi2ps(__a, (__v2si)__b);
|
498
|
+
}
|
499
|
+
|
500
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
501
|
+
_mm_cvt_pi2ps(__m128 __a, __m64 __b)
|
502
|
+
{
|
503
|
+
return _mm_cvtpi32_ps(__a, __b);
|
504
|
+
}
|
505
|
+
|
506
|
+
static __inline__ float __attribute__((__always_inline__, __nodebug__))
|
507
|
+
_mm_cvtss_f32(__m128 __a)
|
508
|
+
{
|
509
|
+
return __a[0];
|
510
|
+
}
|
511
|
+
|
512
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
513
|
+
_mm_loadh_pi(__m128 __a, const __m64 *__p)
|
514
|
+
{
|
515
|
+
typedef float __mm_loadh_pi_v2f32 __attribute__((__vector_size__(8)));
|
516
|
+
struct __mm_loadh_pi_struct {
|
517
|
+
__mm_loadh_pi_v2f32 __u;
|
518
|
+
} __attribute__((__packed__, __may_alias__));
|
519
|
+
__mm_loadh_pi_v2f32 __b = ((struct __mm_loadh_pi_struct*)__p)->__u;
|
520
|
+
__m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
|
521
|
+
return __builtin_shufflevector(__a, __bb, 0, 1, 4, 5);
|
522
|
+
}
|
523
|
+
|
524
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
525
|
+
_mm_loadl_pi(__m128 __a, const __m64 *__p)
|
526
|
+
{
|
527
|
+
typedef float __mm_loadl_pi_v2f32 __attribute__((__vector_size__(8)));
|
528
|
+
struct __mm_loadl_pi_struct {
|
529
|
+
__mm_loadl_pi_v2f32 __u;
|
530
|
+
} __attribute__((__packed__, __may_alias__));
|
531
|
+
__mm_loadl_pi_v2f32 __b = ((struct __mm_loadl_pi_struct*)__p)->__u;
|
532
|
+
__m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
|
533
|
+
return __builtin_shufflevector(__a, __bb, 4, 5, 2, 3);
|
534
|
+
}
|
535
|
+
|
536
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
537
|
+
_mm_load_ss(const float *__p)
|
538
|
+
{
|
539
|
+
struct __mm_load_ss_struct {
|
540
|
+
float __u;
|
541
|
+
} __attribute__((__packed__, __may_alias__));
|
542
|
+
float __u = ((struct __mm_load_ss_struct*)__p)->__u;
|
543
|
+
return (__m128){ __u, 0, 0, 0 };
|
544
|
+
}
|
545
|
+
|
546
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
547
|
+
_mm_load1_ps(const float *__p)
|
548
|
+
{
|
549
|
+
struct __mm_load1_ps_struct {
|
550
|
+
float __u;
|
551
|
+
} __attribute__((__packed__, __may_alias__));
|
552
|
+
float __u = ((struct __mm_load1_ps_struct*)__p)->__u;
|
553
|
+
return (__m128){ __u, __u, __u, __u };
|
554
|
+
}
|
555
|
+
|
556
|
+
#define _mm_load_ps1(p) _mm_load1_ps(p)
|
557
|
+
|
558
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
559
|
+
_mm_load_ps(const float *__p)
|
560
|
+
{
|
561
|
+
return *(__m128*)__p;
|
562
|
+
}
|
563
|
+
|
564
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
565
|
+
_mm_loadu_ps(const float *__p)
|
566
|
+
{
|
567
|
+
struct __loadu_ps {
|
568
|
+
__m128 __v;
|
569
|
+
} __attribute__((__packed__, __may_alias__));
|
570
|
+
return ((struct __loadu_ps*)__p)->__v;
|
571
|
+
}
|
572
|
+
|
573
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
574
|
+
_mm_loadr_ps(const float *__p)
|
575
|
+
{
|
576
|
+
__m128 __a = _mm_load_ps(__p);
|
577
|
+
return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
|
578
|
+
}
|
579
|
+
|
580
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
581
|
+
_mm_set_ss(float __w)
|
582
|
+
{
|
583
|
+
return (__m128){ __w, 0, 0, 0 };
|
584
|
+
}
|
585
|
+
|
586
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
587
|
+
_mm_set1_ps(float __w)
|
588
|
+
{
|
589
|
+
return (__m128){ __w, __w, __w, __w };
|
590
|
+
}
|
591
|
+
|
592
|
+
// Microsoft specific.
|
593
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
594
|
+
_mm_set_ps1(float __w)
|
595
|
+
{
|
596
|
+
return _mm_set1_ps(__w);
|
597
|
+
}
|
598
|
+
|
599
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
600
|
+
_mm_set_ps(float __z, float __y, float __x, float __w)
|
601
|
+
{
|
602
|
+
return (__m128){ __w, __x, __y, __z };
|
603
|
+
}
|
604
|
+
|
605
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
606
|
+
_mm_setr_ps(float __z, float __y, float __x, float __w)
|
607
|
+
{
|
608
|
+
return (__m128){ __z, __y, __x, __w };
|
609
|
+
}
|
610
|
+
|
611
|
+
static __inline__ __m128 __attribute__((__always_inline__))
|
612
|
+
_mm_setzero_ps(void)
|
613
|
+
{
|
614
|
+
return (__m128){ 0, 0, 0, 0 };
|
615
|
+
}
|
616
|
+
|
617
|
+
static __inline__ void __attribute__((__always_inline__))
|
618
|
+
_mm_storeh_pi(__m64 *__p, __m128 __a)
|
619
|
+
{
|
620
|
+
__builtin_ia32_storehps((__v2si *)__p, __a);
|
621
|
+
}
|
622
|
+
|
623
|
+
static __inline__ void __attribute__((__always_inline__))
|
624
|
+
_mm_storel_pi(__m64 *__p, __m128 __a)
|
625
|
+
{
|
626
|
+
__builtin_ia32_storelps((__v2si *)__p, __a);
|
627
|
+
}
|
628
|
+
|
629
|
+
static __inline__ void __attribute__((__always_inline__))
|
630
|
+
_mm_store_ss(float *__p, __m128 __a)
|
631
|
+
{
|
632
|
+
struct __mm_store_ss_struct {
|
633
|
+
float __u;
|
634
|
+
} __attribute__((__packed__, __may_alias__));
|
635
|
+
((struct __mm_store_ss_struct*)__p)->__u = __a[0];
|
636
|
+
}
|
637
|
+
|
638
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
639
|
+
_mm_storeu_ps(float *__p, __m128 __a)
|
640
|
+
{
|
641
|
+
__builtin_ia32_storeups(__p, __a);
|
642
|
+
}
|
643
|
+
|
644
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
645
|
+
_mm_store1_ps(float *__p, __m128 __a)
|
646
|
+
{
|
647
|
+
__a = __builtin_shufflevector(__a, __a, 0, 0, 0, 0);
|
648
|
+
_mm_storeu_ps(__p, __a);
|
649
|
+
}
|
650
|
+
|
651
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
652
|
+
_mm_store_ps1(float *__p, __m128 __a)
|
653
|
+
{
|
654
|
+
return _mm_store1_ps(__p, __a);
|
655
|
+
}
|
656
|
+
|
657
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
658
|
+
_mm_store_ps(float *__p, __m128 __a)
|
659
|
+
{
|
660
|
+
*(__m128 *)__p = __a;
|
661
|
+
}
|
662
|
+
|
663
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
664
|
+
_mm_storer_ps(float *__p, __m128 __a)
|
665
|
+
{
|
666
|
+
__a = __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
|
667
|
+
_mm_store_ps(__p, __a);
|
668
|
+
}
|
669
|
+
|
670
|
+
#define _MM_HINT_T0 3
|
671
|
+
#define _MM_HINT_T1 2
|
672
|
+
#define _MM_HINT_T2 1
|
673
|
+
#define _MM_HINT_NTA 0
|
674
|
+
|
675
|
+
/* FIXME: We have to #define this because "sel" must be a constant integer, and
|
676
|
+
Sema doesn't do any form of constant propagation yet. */
|
677
|
+
|
678
|
+
#define _mm_prefetch(a, sel) (__builtin_prefetch((void *)(a), 0, (sel)))
|
679
|
+
|
680
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
681
|
+
_mm_stream_pi(__m64 *__p, __m64 __a)
|
682
|
+
{
|
683
|
+
__builtin_ia32_movntq(__p, __a);
|
684
|
+
}
|
685
|
+
|
686
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
687
|
+
_mm_stream_ps(float *__p, __m128 __a)
|
688
|
+
{
|
689
|
+
__builtin_ia32_movntps(__p, __a);
|
690
|
+
}
|
691
|
+
|
692
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
693
|
+
_mm_sfence(void)
|
694
|
+
{
|
695
|
+
__builtin_ia32_sfence();
|
696
|
+
}
|
697
|
+
|
698
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
699
|
+
_mm_extract_pi16(__m64 __a, int __n)
|
700
|
+
{
|
701
|
+
__v4hi __b = (__v4hi)__a;
|
702
|
+
return (unsigned short)__b[__n & 3];
|
703
|
+
}
|
704
|
+
|
705
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
706
|
+
_mm_insert_pi16(__m64 __a, int __d, int __n)
|
707
|
+
{
|
708
|
+
__v4hi __b = (__v4hi)__a;
|
709
|
+
__b[__n & 3] = __d;
|
710
|
+
return (__m64)__b;
|
711
|
+
}
|
712
|
+
|
713
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
714
|
+
_mm_max_pi16(__m64 __a, __m64 __b)
|
715
|
+
{
|
716
|
+
return (__m64)__builtin_ia32_pmaxsw((__v4hi)__a, (__v4hi)__b);
|
717
|
+
}
|
718
|
+
|
719
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
720
|
+
_mm_max_pu8(__m64 __a, __m64 __b)
|
721
|
+
{
|
722
|
+
return (__m64)__builtin_ia32_pmaxub((__v8qi)__a, (__v8qi)__b);
|
723
|
+
}
|
724
|
+
|
725
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
726
|
+
_mm_min_pi16(__m64 __a, __m64 __b)
|
727
|
+
{
|
728
|
+
return (__m64)__builtin_ia32_pminsw((__v4hi)__a, (__v4hi)__b);
|
729
|
+
}
|
730
|
+
|
731
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
732
|
+
_mm_min_pu8(__m64 __a, __m64 __b)
|
733
|
+
{
|
734
|
+
return (__m64)__builtin_ia32_pminub((__v8qi)__a, (__v8qi)__b);
|
735
|
+
}
|
736
|
+
|
737
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
738
|
+
_mm_movemask_pi8(__m64 __a)
|
739
|
+
{
|
740
|
+
return __builtin_ia32_pmovmskb((__v8qi)__a);
|
741
|
+
}
|
742
|
+
|
743
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
744
|
+
_mm_mulhi_pu16(__m64 __a, __m64 __b)
|
745
|
+
{
|
746
|
+
return (__m64)__builtin_ia32_pmulhuw((__v4hi)__a, (__v4hi)__b);
|
747
|
+
}
|
748
|
+
|
749
|
+
#define _mm_shuffle_pi16(a, n) __extension__ ({ \
|
750
|
+
__m64 __a = (a); \
|
751
|
+
(__m64)__builtin_ia32_pshufw((__v4hi)__a, (n)); })
|
752
|
+
|
753
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
754
|
+
_mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
|
755
|
+
{
|
756
|
+
__builtin_ia32_maskmovq((__v8qi)__d, (__v8qi)__n, __p);
|
757
|
+
}
|
758
|
+
|
759
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
760
|
+
_mm_avg_pu8(__m64 __a, __m64 __b)
|
761
|
+
{
|
762
|
+
return (__m64)__builtin_ia32_pavgb((__v8qi)__a, (__v8qi)__b);
|
763
|
+
}
|
764
|
+
|
765
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
766
|
+
_mm_avg_pu16(__m64 __a, __m64 __b)
|
767
|
+
{
|
768
|
+
return (__m64)__builtin_ia32_pavgw((__v4hi)__a, (__v4hi)__b);
|
769
|
+
}
|
770
|
+
|
771
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
772
|
+
_mm_sad_pu8(__m64 __a, __m64 __b)
|
773
|
+
{
|
774
|
+
return (__m64)__builtin_ia32_psadbw((__v8qi)__a, (__v8qi)__b);
|
775
|
+
}
|
776
|
+
|
777
|
+
static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
|
778
|
+
_mm_getcsr(void)
|
779
|
+
{
|
780
|
+
return __builtin_ia32_stmxcsr();
|
781
|
+
}
|
782
|
+
|
783
|
+
static __inline__ void __attribute__((__always_inline__, __nodebug__))
|
784
|
+
_mm_setcsr(unsigned int __i)
|
785
|
+
{
|
786
|
+
__builtin_ia32_ldmxcsr(__i);
|
787
|
+
}
|
788
|
+
|
789
|
+
#define _mm_shuffle_ps(a, b, mask) __extension__ ({ \
|
790
|
+
__m128 __a = (a); \
|
791
|
+
__m128 __b = (b); \
|
792
|
+
(__m128)__builtin_shufflevector((__v4sf)__a, (__v4sf)__b, \
|
793
|
+
(mask) & 0x3, ((mask) & 0xc) >> 2, \
|
794
|
+
(((mask) & 0x30) >> 4) + 4, \
|
795
|
+
(((mask) & 0xc0) >> 6) + 4); })
|
796
|
+
|
797
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
798
|
+
_mm_unpackhi_ps(__m128 __a, __m128 __b)
|
799
|
+
{
|
800
|
+
return __builtin_shufflevector(__a, __b, 2, 6, 3, 7);
|
801
|
+
}
|
802
|
+
|
803
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
804
|
+
_mm_unpacklo_ps(__m128 __a, __m128 __b)
|
805
|
+
{
|
806
|
+
return __builtin_shufflevector(__a, __b, 0, 4, 1, 5);
|
807
|
+
}
|
808
|
+
|
809
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
810
|
+
_mm_move_ss(__m128 __a, __m128 __b)
|
811
|
+
{
|
812
|
+
return __builtin_shufflevector(__a, __b, 4, 1, 2, 3);
|
813
|
+
}
|
814
|
+
|
815
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
816
|
+
_mm_movehl_ps(__m128 __a, __m128 __b)
|
817
|
+
{
|
818
|
+
return __builtin_shufflevector(__a, __b, 6, 7, 2, 3);
|
819
|
+
}
|
820
|
+
|
821
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
822
|
+
_mm_movelh_ps(__m128 __a, __m128 __b)
|
823
|
+
{
|
824
|
+
return __builtin_shufflevector(__a, __b, 0, 1, 4, 5);
|
825
|
+
}
|
826
|
+
|
827
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
828
|
+
_mm_cvtpi16_ps(__m64 __a)
|
829
|
+
{
|
830
|
+
__m64 __b, __c;
|
831
|
+
__m128 __r;
|
832
|
+
|
833
|
+
__b = _mm_setzero_si64();
|
834
|
+
__b = _mm_cmpgt_pi16(__b, __a);
|
835
|
+
__c = _mm_unpackhi_pi16(__a, __b);
|
836
|
+
__r = _mm_setzero_ps();
|
837
|
+
__r = _mm_cvtpi32_ps(__r, __c);
|
838
|
+
__r = _mm_movelh_ps(__r, __r);
|
839
|
+
__c = _mm_unpacklo_pi16(__a, __b);
|
840
|
+
__r = _mm_cvtpi32_ps(__r, __c);
|
841
|
+
|
842
|
+
return __r;
|
843
|
+
}
|
844
|
+
|
845
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
846
|
+
_mm_cvtpu16_ps(__m64 __a)
|
847
|
+
{
|
848
|
+
__m64 __b, __c;
|
849
|
+
__m128 __r;
|
850
|
+
|
851
|
+
__b = _mm_setzero_si64();
|
852
|
+
__c = _mm_unpackhi_pi16(__a, __b);
|
853
|
+
__r = _mm_setzero_ps();
|
854
|
+
__r = _mm_cvtpi32_ps(__r, __c);
|
855
|
+
__r = _mm_movelh_ps(__r, __r);
|
856
|
+
__c = _mm_unpacklo_pi16(__a, __b);
|
857
|
+
__r = _mm_cvtpi32_ps(__r, __c);
|
858
|
+
|
859
|
+
return __r;
|
860
|
+
}
|
861
|
+
|
862
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
863
|
+
_mm_cvtpi8_ps(__m64 __a)
|
864
|
+
{
|
865
|
+
__m64 __b;
|
866
|
+
|
867
|
+
__b = _mm_setzero_si64();
|
868
|
+
__b = _mm_cmpgt_pi8(__b, __a);
|
869
|
+
__b = _mm_unpacklo_pi8(__a, __b);
|
870
|
+
|
871
|
+
return _mm_cvtpi16_ps(__b);
|
872
|
+
}
|
873
|
+
|
874
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
875
|
+
_mm_cvtpu8_ps(__m64 __a)
|
876
|
+
{
|
877
|
+
__m64 __b;
|
878
|
+
|
879
|
+
__b = _mm_setzero_si64();
|
880
|
+
__b = _mm_unpacklo_pi8(__a, __b);
|
881
|
+
|
882
|
+
return _mm_cvtpi16_ps(__b);
|
883
|
+
}
|
884
|
+
|
885
|
+
static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
|
886
|
+
_mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
|
887
|
+
{
|
888
|
+
__m128 __c;
|
889
|
+
|
890
|
+
__c = _mm_setzero_ps();
|
891
|
+
__c = _mm_cvtpi32_ps(__c, __b);
|
892
|
+
__c = _mm_movelh_ps(__c, __c);
|
893
|
+
|
894
|
+
return _mm_cvtpi32_ps(__c, __a);
|
895
|
+
}
|
896
|
+
|
897
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
898
|
+
_mm_cvtps_pi16(__m128 __a)
|
899
|
+
{
|
900
|
+
__m64 __b, __c;
|
901
|
+
|
902
|
+
__b = _mm_cvtps_pi32(__a);
|
903
|
+
__a = _mm_movehl_ps(__a, __a);
|
904
|
+
__c = _mm_cvtps_pi32(__a);
|
905
|
+
|
906
|
+
return _mm_packs_pi16(__b, __c);
|
907
|
+
}
|
908
|
+
|
909
|
+
static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
|
910
|
+
_mm_cvtps_pi8(__m128 __a)
|
911
|
+
{
|
912
|
+
__m64 __b, __c;
|
913
|
+
|
914
|
+
__b = _mm_cvtps_pi16(__a);
|
915
|
+
__c = _mm_setzero_si64();
|
916
|
+
|
917
|
+
return _mm_packs_pi16(__b, __c);
|
918
|
+
}
|
919
|
+
|
920
|
+
static __inline__ int __attribute__((__always_inline__, __nodebug__))
|
921
|
+
_mm_movemask_ps(__m128 __a)
|
922
|
+
{
|
923
|
+
return __builtin_ia32_movmskps(__a);
|
924
|
+
}
|
925
|
+
|
926
|
+
#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))
|
927
|
+
|
928
|
+
#define _MM_EXCEPT_INVALID (0x0001)
|
929
|
+
#define _MM_EXCEPT_DENORM (0x0002)
|
930
|
+
#define _MM_EXCEPT_DIV_ZERO (0x0004)
|
931
|
+
#define _MM_EXCEPT_OVERFLOW (0x0008)
|
932
|
+
#define _MM_EXCEPT_UNDERFLOW (0x0010)
|
933
|
+
#define _MM_EXCEPT_INEXACT (0x0020)
|
934
|
+
#define _MM_EXCEPT_MASK (0x003f)
|
935
|
+
|
936
|
+
#define _MM_MASK_INVALID (0x0080)
|
937
|
+
#define _MM_MASK_DENORM (0x0100)
|
938
|
+
#define _MM_MASK_DIV_ZERO (0x0200)
|
939
|
+
#define _MM_MASK_OVERFLOW (0x0400)
|
940
|
+
#define _MM_MASK_UNDERFLOW (0x0800)
|
941
|
+
#define _MM_MASK_INEXACT (0x1000)
|
942
|
+
#define _MM_MASK_MASK (0x1f80)
|
943
|
+
|
944
|
+
#define _MM_ROUND_NEAREST (0x0000)
|
945
|
+
#define _MM_ROUND_DOWN (0x2000)
|
946
|
+
#define _MM_ROUND_UP (0x4000)
|
947
|
+
#define _MM_ROUND_TOWARD_ZERO (0x6000)
|
948
|
+
#define _MM_ROUND_MASK (0x6000)
|
949
|
+
|
950
|
+
#define _MM_FLUSH_ZERO_MASK (0x8000)
|
951
|
+
#define _MM_FLUSH_ZERO_ON (0x8000)
|
952
|
+
#define _MM_FLUSH_ZERO_OFF (0x0000)
|
953
|
+
|
954
|
+
#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
|
955
|
+
#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
|
956
|
+
#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
|
957
|
+
#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)
|
958
|
+
|
959
|
+
#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
|
960
|
+
#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
|
961
|
+
#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
|
962
|
+
#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))
|
963
|
+
|
964
|
+
#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
|
965
|
+
do { \
|
966
|
+
__m128 tmp3, tmp2, tmp1, tmp0; \
|
967
|
+
tmp0 = _mm_unpacklo_ps((row0), (row1)); \
|
968
|
+
tmp2 = _mm_unpacklo_ps((row2), (row3)); \
|
969
|
+
tmp1 = _mm_unpackhi_ps((row0), (row1)); \
|
970
|
+
tmp3 = _mm_unpackhi_ps((row2), (row3)); \
|
971
|
+
(row0) = _mm_movelh_ps(tmp0, tmp2); \
|
972
|
+
(row1) = _mm_movehl_ps(tmp2, tmp0); \
|
973
|
+
(row2) = _mm_movelh_ps(tmp1, tmp3); \
|
974
|
+
(row3) = _mm_movehl_ps(tmp3, tmp1); \
|
975
|
+
} while (0)
|
976
|
+
|
977
|
+
/* Aliases for compatibility. */
|
978
|
+
#define _m_pextrw _mm_extract_pi16
|
979
|
+
#define _m_pinsrw _mm_insert_pi16
|
980
|
+
#define _m_pmaxsw _mm_max_pi16
|
981
|
+
#define _m_pmaxub _mm_max_pu8
|
982
|
+
#define _m_pminsw _mm_min_pi16
|
983
|
+
#define _m_pminub _mm_min_pu8
|
984
|
+
#define _m_pmovmskb _mm_movemask_pi8
|
985
|
+
#define _m_pmulhuw _mm_mulhi_pu16
|
986
|
+
#define _m_pshufw _mm_shuffle_pi16
|
987
|
+
#define _m_maskmovq _mm_maskmove_si64
|
988
|
+
#define _m_pavgb _mm_avg_pu8
|
989
|
+
#define _m_pavgw _mm_avg_pu16
|
990
|
+
#define _m_psadbw _mm_sad_pu8
|
991
|
+
#define _m_ _mm_
|
992
|
+
#define _m_ _mm_
|
993
|
+
|
994
|
+
/* Ugly hack for backwards-compatibility (compatible with gcc) */
|
995
|
+
#ifdef __SSE2__
|
996
|
+
#include <emmintrin.h>
|
997
|
+
#endif
|
998
|
+
|
999
|
+
#endif /* __SSE__ */
|
1000
|
+
|
1001
|
+
#endif /* __XMMINTRIN_H */
|