maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/CodeGen/builtins-ppc-p8vector.c

855 lines
28 KiB
C
Raw Normal View History

Revert r240872, "Suppress clang/test/CodeGen/builtins-ppc-p8vector.c for -Asserts for now. Will fix later." This has been fixed since r240912. llvm-svn: 240920
2015-06-29 07:14:35 +08:00
// REQUIRES: powerpc-registered-target
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// RUN: %clang_cc1 -faltivec -target-feature +power8-vector -triple powerpc64-unknown-unknown -emit-llvm %s -o - | FileCheck %s
// RUN: %clang_cc1 -faltivec -target-feature +power8-vector -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-LE
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
// RUN: not %clang_cc1 -faltivec -target-feature +vsx -triple powerpc64-unknown-unknown -emit-llvm %s -o - 2>&1 | FileCheck %s -check-prefix=CHECK-PPC
// Added -target-feature +vsx above to avoid errors about "vector double" and to
// generate the correct errors for functions that are only overloaded with VSX
// (vec_cmpge, vec_cmple). Without this option, there is only one overload so
// it is selected.
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
Clang support for vector quad bit permute and gather instructions through builtins This patch corresponds to review: http://reviews.llvm.org/D10095 This is for just two instructions and related builtins: vbpermq vgbbd llvm-svn: 239506
2015-06-11 14:25:36 +08:00
vector signed char vsc = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5 };
vector unsigned char vuc = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5 };
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
vector int vi = { -1, 2, -3, 4 };
vector unsigned int vui = { 1, 2, 3, 4 };
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
vector bool int vbi = {0, -1, -1, 0};
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
vector bool long long vbll = { 1, 0 };
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
vector signed long long vsll = { 1, 2 };
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
vector unsigned long long vull = { 1, 2 };
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
vector double vda = { 1.e-11, -132.23e10 };
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
int res_i;
Clang support for vector quad bit permute and gather instructions through builtins This patch corresponds to review: http://reviews.llvm.org/D10095 This is for just two instructions and related builtins: vbpermq vgbbd llvm-svn: 239506
2015-06-11 14:25:36 +08:00
vector signed char res_vsc;
vector unsigned char res_vuc;
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
vector int res_vi;
vector unsigned int res_vui;
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
vector bool int res_vbi;
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
vector bool long long res_vbll;
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
vector signed long long res_vsll;
vector unsigned long long res_vull;
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
vector double res_vd;
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK-LABEL: define void @test1
void test1() {
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
/* vec_abs */
res_vsll = vec_abs(vsll);
[PPC] fixes typos in builtins-ppc-p8vector.c The extra ] causes %{{[0-9]]*}} to match only %<single digit> such as %1. llvm-svn: 240912
2015-06-29 02:30:36 +08:00
// CHECK: call <2 x i64> @llvm.ppc.altivec.vmaxsd(<2 x i64> %{{[0-9]*}}, <2 x i64>
// CHECK-LE: call <2 x i64> @llvm.ppc.altivec.vmaxsd(<2 x i64> %{{[0-9]*}}, <2 x i64>
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
// CHECK-PPC: error: call to 'vec_abs' is ambiguous
res_vd = vec_abs(vda);
// CHECK: store <2 x i64> <i64 9223372036854775807, i64 9223372036854775807>, <2 x i64>*
// CHECK: and <2 x i64>
// CHECK-LE: store <2 x i64> <i64 9223372036854775807, i64 9223372036854775807>, <2 x i64>*
// CHECK-LE: and <2 x i64>
// CHECK-PPC: error: call to 'vec_abs' is ambiguous
/* vec_add */
res_vsll = vec_add(vsll, vsll);
// CHECK: add <2 x i64>
// CHECK-LE: add <2 x i64>
// CHECK-PPC: error: call to 'vec_add' is ambiguous
res_vull = vec_add(vull, vull);
// CHECK: add <2 x i64>
// CHECK-LE: add <2 x i64>
// CHECK-PPC: error: call to 'vec_add' is ambiguous
This patch adds support for the vector merge even word and vector merge odd word instructions introduced in POWER8. These are the Clang-related changes for http://reviews.llvm.org/D10704 All builtins are added in altivec.h and guarded with the POWER8_VECTOR macro. Phabricator review: http://reviews.llvm.org/D10736 llvm-svn: 241293
2015-07-03 03:29:05 +08:00
/* vec_mergee */
res_vbi = vec_mergee(vbi, vbi);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
res_vi = vec_mergee(vi, vi);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
res_vui = vec_mergee(vui, vui);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
// CHECK-PPC: warning: implicit declaration of function 'vec_mergee'
/* vec_mergeo */
res_vbi = vec_mergeo(vbi, vbi);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
res_vi = vec_mergeo(vi, vi);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
res_vui = vec_mergeo(vui, vui);
// CHECK: @llvm.ppc.altivec.vperm
// CHECK-LE: @llvm.ppc.altivec.vperm
// CHECK-PPC: warning: implicit declaration of function 'vec_mergeo'
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
/* vec_cmpeq */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vbll = vec_cmpeq(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd
// CHECK-LE: @llvm.ppc.altivec.vcmpequd
// CHECK-PPC: error: call to 'vec_cmpeq' is ambiguous
res_vbll = vec_cmpeq(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd
// CHECK-LE: @llvm.ppc.altivec.vcmpequd
// CHECK-PPC: error: call to 'vec_cmpeq' is ambiguous
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
/* vec_cmpge */
res_vbll = vec_cmpge(vsll, vsll);
// CHECK: @llvm.ppc.altivec.vcmpgtsd
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd
// CHECK-PPC: error: call to 'vec_cmpge' is ambiguous
res_vbll = vec_cmpge(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud
// CHECK-PPC: error: call to 'vec_cmpge' is ambiguous
/* vec_cmple */
res_vbll = vec_cmple(vsll, vsll);
// CHECK: @llvm.ppc.altivec.vcmpgtsd
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd
// CHECK-PPC: error: call to 'vec_cmple' is ambiguous
res_vbll = vec_cmple(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud
// CHECK-PPC: error: call to 'vec_cmple' is ambiguous
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
/* vec_cmpgt */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vbll = vec_cmpgt(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd
// CHECK-PPC: error: call to 'vec_cmpgt' is ambiguous
res_vbll = vec_cmpgt(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud
// CHECK-PPC: error: call to 'vec_cmpgt' is ambiguous
Add missing builtins to altivec.h for ABI compliance (vol. 1) This patch corresponds to review: http://reviews.llvm.org/D10637 This is the first round of additions of missing builtins listed in the ABI document. More to come (this builds onto what seurer already addes). This patch adds: vector signed long long vec_abs(vector signed long long) vector double vec_abs(vector double) vector signed long long vec_add(vector signed long long, vector signed long long) vector unsigned long long vec_add(vector unsigned long long, vector unsigned long long) vector double vec_add(vector double, vector double) vector double vec_and(vector bool long long, vector double) vector double vec_and(vector double, vector bool long long) vector double vec_and(vector double, vector double) vector signed long long vec_and(vector signed long long, vector signed long long) vector double vec_andc(vector bool long long, vector double) vector double vec_andc(vector double, vector bool long long) vector double vec_andc(vector double, vector double) vector signed long long vec_andc(vector signed long long, vector signed long long) vector double vec_ceil(vector double) vector bool long long vec_cmpeq(vector double, vector double) vector bool long long vec_cmpge(vector double, vector double) vector bool long long vec_cmpge(vector signed long long, vector signed long long) vector bool long long vec_cmpge(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmpgt(vector double, vector double) vector bool long long vec_cmple(vector double, vector double) vector bool long long vec_cmple(vector signed long long, vector signed long long) vector bool long long vec_cmple(vector unsigned long long, vector unsigned long long) vector bool long long vec_cmplt(vector double, vector double) vector bool long long vec_cmplt(vector signed long long, vector signed long long) vector bool long long vec_cmplt(vector unsigned long long, vector unsigned long long) llvm-svn: 240821
2015-06-27 03:27:20 +08:00
/* vec_cmplt */
res_vbll = vec_cmplt(vsll, vsll);
// CHECK: call <2 x i64> @llvm.ppc.altivec.vcmpgtsd(<2 x i64> %{{[0-9]*}}, <2 x i64> %{{[0-9]*}})
// CHECK-LE: call <2 x i64> @llvm.ppc.altivec.vcmpgtsd(<2 x i64> %{{[0-9]*}}, <2 x i64> %{{[0-9]*}})
// CHECK-PPC: error: call to 'vec_cmplt' is ambiguous
res_vbll = vec_cmplt(vull, vull);
// CHECK: call <2 x i64> @llvm.ppc.altivec.vcmpgtud(<2 x i64> %{{[0-9]*}}, <2 x i64> %{{[0-9]*}})
// CHECK-LE: call <2 x i64> @llvm.ppc.altivec.vcmpgtud(<2 x i64> %{{[0-9]*}}, <2 x i64> %{{[0-9]*}})
// CHECK-PPC: error: call to 'vec_cmplt' is ambiguous
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
/* ----------------------- predicates --------------------------- */
/* vec_all_eq */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_eq(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_eq(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
res_i = vec_all_eq(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
res_i = vec_all_eq(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_eq(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
res_i = vec_all_eq(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
res_i = vec_all_eq(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_eq' is ambiguous
/* vec_all_ne */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ne(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ne(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
res_i = vec_all_ne(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
res_i = vec_all_ne(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ne(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
res_i = vec_all_ne(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
res_i = vec_all_ne(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_all_ne' is ambiguous
/* vec_any_eq */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_eq(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_eq(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
res_i = vec_any_eq(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
res_i = vec_any_eq(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_eq(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
res_i = vec_any_eq(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
res_i = vec_any_eq(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_eq' is ambiguous
/* vec_any_ne */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ne(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ne(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
res_i = vec_any_ne(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
res_i = vec_any_ne(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ne(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
res_i = vec_any_ne(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
res_i = vec_any_ne(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpequd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpequd.p
// CHECK-PPC: error: call to 'vec_any_ne' is ambiguous
/* vec_all_ge */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ge(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ge(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
res_i = vec_all_ge(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
res_i = vec_all_ge(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_ge(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
res_i = vec_all_ge(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
res_i = vec_all_ge(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_ge' is ambiguous
/* vec_all_gt */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_gt(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_gt(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
res_i = vec_all_gt(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
res_i = vec_all_gt(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_gt(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
res_i = vec_all_gt(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
res_i = vec_all_gt(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_gt' is ambiguous
/* vec_all_le */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_le(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_le(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
res_i = vec_all_le(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
res_i = vec_all_le(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_le(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
res_i = vec_all_le(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
res_i = vec_all_le(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_le' is ambiguous
/* vec_all_lt */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_lt(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_lt(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
res_i = vec_all_lt(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
res_i = vec_all_lt(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_all_lt(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
res_i = vec_all_lt(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
res_i = vec_all_lt(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_all_lt' is ambiguous
/* vec_any_ge */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ge(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ge(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
res_i = vec_any_ge(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
res_i = vec_any_ge(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_ge(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
res_i = vec_any_ge(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
res_i = vec_any_ge(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_ge' is ambiguous
/* vec_any_gt */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_gt(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_gt(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
res_i = vec_any_gt(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
res_i = vec_any_gt(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_gt(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
res_i = vec_any_gt(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
res_i = vec_any_gt(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_gt' is ambiguous
/* vec_any_le */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_le(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_le(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
res_i = vec_any_le(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
res_i = vec_any_le(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_le(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
res_i = vec_any_le(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
res_i = vec_any_le(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_le' is ambiguous
/* vec_any_lt */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_lt(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_lt(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtsd.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
res_i = vec_any_lt(vull, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
res_i = vec_any_lt(vull, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_i = vec_any_lt(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
res_i = vec_any_lt(vbll, vull);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
res_i = vec_any_lt(vbll, vbll);
// CHECK: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-LE: @llvm.ppc.altivec.vcmpgtud.p
// CHECK-PPC: error: call to 'vec_any_lt' is ambiguous
/* vec_max */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_max(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
// CHECK-PPC: error: call to 'vec_max' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_max(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
// CHECK-PPC: error: call to 'vec_max' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_max(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
// CHECK-PPC: error: call to 'vec_max' is ambiguous
res_vull = vec_max(vull, vull);
// CHECK: @llvm.ppc.altivec.vmaxud
// CHECK-LE: @llvm.ppc.altivec.vmaxud
// CHECK-PPC: error: call to 'vec_max' is ambiguous
res_vull = vec_max(vbll, vull);
// CHECK: @llvm.ppc.altivec.vmaxud
// CHECK-LE: @llvm.ppc.altivec.vmaxud
// CHECK-PPC: error: call to 'vec_max' is ambiguous
res_vull = vec_max(vull, vbll);
// CHECK: @llvm.ppc.altivec.vmaxud
// CHECK-LE: @llvm.ppc.altivec.vmaxud
// CHECK-PPC: error: call to 'vec_max' is ambiguous
/* vec_min */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_min(vsll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
// CHECK-PPC: error: call to 'vec_min' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_min(vbll, vsll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
// CHECK-PPC: error: call to 'vec_min' is ambiguous
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_min(vsll, vbll);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
// CHECK-PPC: error: call to 'vec_min' is ambiguous
res_vull = vec_min(vull, vull);
// CHECK: @llvm.ppc.altivec.vminud
// CHECK-LE: @llvm.ppc.altivec.vminud
// CHECK-PPC: error: call to 'vec_min' is ambiguous
res_vull = vec_min(vbll, vull);
// CHECK: @llvm.ppc.altivec.vminud
// CHECK-LE: @llvm.ppc.altivec.vminud
// CHECK-PPC: error: call to 'vec_min' is ambiguous
res_vull = vec_min(vull, vbll);
// CHECK: @llvm.ppc.altivec.vminud
// CHECK-LE: @llvm.ppc.altivec.vminud
// CHECK-PPC: error: call to 'vec_min' is ambiguous
/* vec_mule */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_mule(vi, vi);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vmulesw
// CHECK-LE: @llvm.ppc.altivec.vmulosw
// CHECK-PPC: error: call to 'vec_mule' is ambiguous
res_vull = vec_mule(vui , vui);
// CHECK: @llvm.ppc.altivec.vmuleuw
// CHECK-LE: @llvm.ppc.altivec.vmulouw
// CHECK-PPC: error: call to 'vec_mule' is ambiguous
/* vec_mulo */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_mulo(vi, vi);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vmulosw
// CHECK-LE: @llvm.ppc.altivec.vmulesw
// CHECK-PPC: error: call to 'vec_mulo' is ambiguous
res_vull = vec_mulo(vui, vui);
// CHECK: @llvm.ppc.altivec.vmulouw
// CHECK-LE: @llvm.ppc.altivec.vmuleuw
// CHECK-PPC: error: call to 'vec_mulo' is ambiguous
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
/* vec_packs */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vi = vec_packs(vsll, vsll);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: @llvm.ppc.altivec.vpksdss
// CHECK-LE: @llvm.ppc.altivec.vpksdss
// CHECK-PPC: error: call to 'vec_packs' is ambiguous
res_vui = vec_packs(vull, vull);
// CHECK: @llvm.ppc.altivec.vpkudus
// CHECK-LE: @llvm.ppc.altivec.vpkudus
// CHECK-PPC: error: call to 'vec_packs' is ambiguous
/* vec_packsu */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vui = vec_packsu(vsll, vsll);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: @llvm.ppc.altivec.vpksdus
// CHECK-LE: @llvm.ppc.altivec.vpksdus
// CHECK-PPC: error: call to 'vec_packsu' is ambiguous
res_vui = vec_packsu(vull, vull);
// CHECK: @llvm.ppc.altivec.vpkudus
// CHECK-LE: @llvm.ppc.altivec.vpkudus
// CHECK-PPC: error: call to 'vec_packsu' is ambiguous
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
/* vec_rl */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_rl(vsll, vull);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: @llvm.ppc.altivec.vrld
// CHECK-LE: @llvm.ppc.altivec.vrld
// CHECK-PPC: error: call to 'vec_rl' is ambiguous
res_vull = vec_rl(vull, vull);
// CHECK: @llvm.ppc.altivec.vrld
// CHECK-LE: @llvm.ppc.altivec.vrld
// CHECK-PPC: error: call to 'vec_rl' is ambiguous
/* vec_sl */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_sl(vsll, vull);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: shl <2 x i64>
// CHECK-LE: shl <2 x i64>
// CHECK-PPC: error: call to 'vec_sl' is ambiguous
res_vull = vec_sl(vull, vull);
// CHECK: shl <2 x i64>
// CHECK-LE: shl <2 x i64>
// CHECK-PPC: error: call to 'vec_sl' is ambiguous
/* vec_sr */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_sr(vsll, vull);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: ashr <2 x i64>
// CHECK-LE: ashr <2 x i64>
// CHECK-PPC: error: call to 'vec_sr' is ambiguous
res_vull = vec_sr(vull, vull);
// CHECK: lshr <2 x i64>
// CHECK-LE: lshr <2 x i64>
// CHECK-PPC: error: call to 'vec_sr' is ambiguous
/* vec_sra */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_sra(vsll, vull);
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
// CHECK: ashr <2 x i64>
// CHECK-LE: ashr <2 x i64>
// CHECK-PPC: error: call to 'vec_sra' is ambiguous
res_vull = vec_sra(vull, vull);
// CHECK: ashr <2 x i64>
// CHECK-LE: ashr <2 x i64>
// CHECK-PPC: error: call to 'vec_sra' is ambiguous
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
/* vec_unpackh */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_unpackh(vi);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vupkhsw
// CHECK-LE: llvm.ppc.altivec.vupklsw
// CHECK-PPC: error: call to 'vec_unpackh' is ambiguous
res_vbll = vec_unpackh(vbi);
// CHECK: llvm.ppc.altivec.vupkhsw
// CHECK-LE: llvm.ppc.altivec.vupklsw
// CHECK-PPC: error: call to 'vec_unpackh' is ambiguous
/* vec_unpackl */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_unpackl(vi);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vupklsw
// CHECK-LE: llvm.ppc.altivec.vupkhsw
// CHECK-PPC: error: call to 'vec_unpackl' is ambiguous
res_vbll = vec_unpackl(vbi);
// CHECK: llvm.ppc.altivec.vupklsw
// CHECK-LE: llvm.ppc.altivec.vupkhsw
// CHECK-PPC: error: call to 'vec_unpackl' is ambiguous
/* vec_vpksdss */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vi = vec_vpksdss(vsll, vsll);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vpksdss
// CHECK-LE: llvm.ppc.altivec.vpksdss
// CHECK-PPC: warning: implicit declaration of function 'vec_vpksdss'
/* vec_vpksdus */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vui = vec_vpksdus(vsll, vsll);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vpksdus
// CHECK-LE: llvm.ppc.altivec.vpksdus
// CHECK-PPC: warning: implicit declaration of function 'vec_vpksdus'
/* vec_vpkudum */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vi = vec_vpkudum(vsll, vsll);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: vperm
// CHECK-LE: vperm
// CHECK-PPC: warning: implicit declaration of function 'vec_vpkudum'
res_vui = vec_vpkudum(vull, vull);
// CHECK: vperm
// CHECK-LE: vperm
res_vui = vec_vpkudus(vull, vull);
// CHECK: llvm.ppc.altivec.vpkudus
// CHECK-LE: llvm.ppc.altivec.vpkudus
// CHECK-PPC: warning: implicit declaration of function 'vec_vpkudus'
/* vec_vupkhsw */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_vupkhsw(vi);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vupkhsw
// CHECK-LE: llvm.ppc.altivec.vupklsw
// CHECK-PPC: warning: implicit declaration of function 'vec_vupkhsw'
res_vbll = vec_vupkhsw(vbi);
// CHECK: llvm.ppc.altivec.vupkhsw
// CHECK-LE: llvm.ppc.altivec.vupklsw
/* vec_vupklsw */
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
res_vsll = vec_vupklsw(vi);
[PPC64] Add vector pack/unpack support from ISA 2.07 This patch adds support for the following new instructions in the Power ISA 2.07: vpksdss vpksdus vpkudus vpkudum vupkhsw vupklsw These instructions are available through the vec_packs, vec_packsu, vec_unpackh, and vec_unpackl built-in interfaces. These are lane-sensitive instructions, so the built-ins have different implementations for big- and little-endian, and the instructions must be marked as killing the vector swap optimization for now. The first three instructions perform saturating pack operations. The fourth performs a modulo pack operation, which means it can be represented with a vector shuffle, and conversely the appropriate vector shuffles may cause this instruction to be generated. The other instructions are only generated via built-in support for now. I noticed during patch preparation that the macro __VSX__ was not previously predefined when the power8-vector or direct-move features are requested. This is an error, and I've corrected that here as well. Appropriate tests have been added. There is a companion patch to llvm for the rest of this support. llvm-svn: 237500
2015-05-16 09:02:25 +08:00
// CHECK: llvm.ppc.altivec.vupklsw
// CHECK-LE: llvm.ppc.altivec.vupkhsw
// CHECK-PPC: warning: implicit declaration of function 'vec_vupklsw'
res_vbll = vec_vupklsw(vbi);
// CHECK: llvm.ppc.altivec.vupklsw
// CHECK-LE: llvm.ppc.altivec.vupkhsw
[PowerPC] This revision adds 68 of the missing "Predefined Functions for Vector Programming" from appendix A of the OpenPOWER ABI for Linux Supplement document. I also added tests for the new functions and updated another test that was looking for specific line numbers in error messages from altivec.h. https://llvm.org/bugs/show_bug.cgi?id=23679 http://reviews.llvm.org/D10131 llvm-svn: 239066
2015-06-05 02:45:44 +08:00
/* vec_max */
res_vsll = vec_max(vsll, vsll);
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
res_vsll = vec_max(vbll, vsll);
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
res_vsll = vec_max(vsll, vbll);
// CHECK: @llvm.ppc.altivec.vmaxsd
// CHECK-LE: @llvm.ppc.altivec.vmaxsd
res_vull = vec_max(vull, vull);
// CHECK: @llvm.ppc.altivec.vmaxud
// CHECK-LE: @llvm.ppc.altivec.vmaxud
res_vull = vec_max(vbll, vull);
// CHECK: @llvm.ppc.altivec.vmaxud
// CHECK-LE: @llvm.ppc.altivec.vmaxud
/* vec_min */
res_vsll = vec_min(vsll, vsll);
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
res_vsll = vec_min(vbll, vsll);
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
res_vsll = vec_min(vsll, vbll);
// CHECK: @llvm.ppc.altivec.vminsd
// CHECK-LE: @llvm.ppc.altivec.vminsd
res_vull = vec_min(vull, vull);
// CHECK: @llvm.ppc.altivec.vminud
// CHECK-LE: @llvm.ppc.altivec.vminud
res_vull = vec_min(vbll, vull);
// CHECK: @llvm.ppc.altivec.vminud
// CHECK-LE: @llvm.ppc.altivec.vminud
Clang support for vector quad bit permute and gather instructions through builtins This patch corresponds to review: http://reviews.llvm.org/D10095 This is for just two instructions and related builtins: vbpermq vgbbd llvm-svn: 239506
2015-06-11 14:25:36 +08:00
/* vec_vbpermq */
res_vsll = vec_vbpermq(vsc, vsc);
// CHECK: llvm.ppc.altivec.vbpermq
// CHECK-LE: llvm.ppc.altivec.vbpermq
res_vull = vec_vbpermq(vuc, vuc);
// CHECK: llvm.ppc.altivec.vbpermq
// CHECK-LE: llvm.ppc.altivec.vbpermq
// CHECK-PPC: warning: implicit declaration of function 'vec_vbpermq'
/* vec_vgbbd */
res_vsc = vec_vgbbd(vsc);
// CHECK: llvm.ppc.altivec.vgbbd
// CHECK-LE: llvm.ppc.altivec.vgbbd
res_vuc = vec_vgbbd(vuc);
// CHECK: llvm.ppc.altivec.vgbbd
// CHECK-LE: llvm.ppc.altivec.vgbbd
// CHECK-PPC: warning: implicit declaration of function 'vec_vgbbd'
Add builtins for the 64-bit vector integer arithmetic instructions added in POWER8. These are the Clang-related changes for the instructions added to LLVM in http://reviews.llvm.org/D7959. Phabricator review: http://reviews.llvm.org/D8041 llvm-svn: 231931
2015-03-11 23:57:19 +08:00
}