maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/Thumb2/mve-frint.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

529 lines
17 KiB
LLVM
Raw Normal View History

[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-arm-none-eabi -mattr=+mve,+fullfp16 -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-MVE
; RUN: llc -mtriple=thumbv8.1m.main-arm-none-eabi -mattr=+mve.fp -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-MVEFP
define arm_aapcs_vfpcc <4 x float> @fceil_float32_t(<4 x float> %src) {
; CHECK-MVE-LABEL: fceil_float32_t:
; CHECK-MVE: @ %bb.0: @ %entry
; CHECK-MVE-NEXT: vrintp.f32 s7, s3
; CHECK-MVE-NEXT: vrintp.f32 s6, s2
; CHECK-MVE-NEXT: vrintp.f32 s5, s1
; CHECK-MVE-NEXT: vrintp.f32 s4, s0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: fceil_float32_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintp.f32 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.ceil.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @fceil_float16_t(<8 x half> %src) {
; CHECK-MVE-LABEL: fceil_float16_t:
; CHECK-MVE: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s4, s0
; CHECK-MVE-NEXT: vrintp.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vrintp.f16 s4, s4
; CHECK-MVE-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintp.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s0, s3
; CHECK-MVE-NEXT: vmov.16 q1[0], r1
; CHECK-MVE-NEXT: vrintp.f16 s0, s0
; CHECK-MVE-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[2], r0
; CHECK-MVE-NEXT: vrintp.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintp.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[3], r0
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[4], r0
; CHECK-MVE-NEXT: vrintp.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintp.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[5], r0
; CHECK-MVE-NEXT: vmov r0, s8
; CHECK-MVE-NEXT: vmov.16 q1[6], r0
; CHECK-MVE-NEXT: vmov r0, s0
; CHECK-MVE-NEXT: vmov.16 q1[7], r0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: fceil_float16_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintp.f16 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.ceil.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @fceil_float64_t(<2 x double> %src) {
; CHECK-LABEL: fceil_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl ceil
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl ceil
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.ceil.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
define arm_aapcs_vfpcc <4 x float> @ftrunc_float32_t(<4 x float> %src) {
; CHECK-MVE-LABEL: ftrunc_float32_t:
; CHECK-MVE: @ %bb.0: @ %entry
; CHECK-MVE-NEXT: vrintz.f32 s7, s3
; CHECK-MVE-NEXT: vrintz.f32 s6, s2
; CHECK-MVE-NEXT: vrintz.f32 s5, s1
; CHECK-MVE-NEXT: vrintz.f32 s4, s0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: ftrunc_float32_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintz.f32 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.trunc.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @ftrunc_float16_t(<8 x half> %src) {
; CHECK-MVE-LABEL: ftrunc_float16_t:
; CHECK-MVE: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s4, s0
; CHECK-MVE-NEXT: vrintz.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vrintz.f16 s4, s4
; CHECK-MVE-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintz.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s0, s3
; CHECK-MVE-NEXT: vmov.16 q1[0], r1
; CHECK-MVE-NEXT: vrintz.f16 s0, s0
; CHECK-MVE-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[2], r0
; CHECK-MVE-NEXT: vrintz.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintz.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[3], r0
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[4], r0
; CHECK-MVE-NEXT: vrintz.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintz.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[5], r0
; CHECK-MVE-NEXT: vmov r0, s8
; CHECK-MVE-NEXT: vmov.16 q1[6], r0
; CHECK-MVE-NEXT: vmov r0, s0
; CHECK-MVE-NEXT: vmov.16 q1[7], r0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: ftrunc_float16_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintz.f16 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.trunc.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @ftrunc_float64_t(<2 x double> %src) {
; CHECK-LABEL: ftrunc_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl trunc
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl trunc
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.trunc.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
define arm_aapcs_vfpcc <4 x float> @frint_float32_t(<4 x float> %src) {
; CHECK-MVE-LABEL: frint_float32_t:
; CHECK-MVE: @ %bb.0: @ %entry
; CHECK-MVE-NEXT: vrintx.f32 s7, s3
; CHECK-MVE-NEXT: vrintx.f32 s6, s2
; CHECK-MVE-NEXT: vrintx.f32 s5, s1
; CHECK-MVE-NEXT: vrintx.f32 s4, s0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: frint_float32_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintx.f32 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.rint.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @frint_float16_t(<8 x half> %src) {
; CHECK-MVE-LABEL: frint_float16_t:
; CHECK-MVE: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s4, s0
; CHECK-MVE-NEXT: vrintx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vrintx.f16 s4, s4
; CHECK-MVE-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintx.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s0, s3
; CHECK-MVE-NEXT: vmov.16 q1[0], r1
; CHECK-MVE-NEXT: vrintx.f16 s0, s0
; CHECK-MVE-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[2], r0
; CHECK-MVE-NEXT: vrintx.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[3], r0
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[4], r0
; CHECK-MVE-NEXT: vrintx.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintx.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[5], r0
; CHECK-MVE-NEXT: vmov r0, s8
; CHECK-MVE-NEXT: vmov.16 q1[6], r0
; CHECK-MVE-NEXT: vmov r0, s0
; CHECK-MVE-NEXT: vmov.16 q1[7], r0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: frint_float16_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintx.f16 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.rint.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @frint_float64_t(<2 x double> %src) {
; CHECK-LABEL: frint_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl rint
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl rint
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.rint.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
define arm_aapcs_vfpcc <4 x float> @fnearbyint_float32_t(<4 x float> %src) {
; CHECK-LABEL: fnearbyint_float32_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vrintr.f32 s7, s3
; CHECK-NEXT: vrintr.f32 s6, s2
; CHECK-NEXT: vrintr.f32 s5, s1
; CHECK-NEXT: vrintr.f32 s4, s0
; CHECK-NEXT: vmov q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.nearbyint.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @fnearbyint_float16_t(<8 x half> %src) {
; CHECK-LABEL: fnearbyint_float16_t:
; CHECK: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vmovx.f16 s4, s0
; CHECK-NEXT: vrintr.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vrintr.f16 s4, s4
; CHECK-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vrintr.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vmovx.f16 s0, s3
; CHECK-NEXT: vmov.16 q1[0], r1
; CHECK-NEXT: vrintr.f16 s0, s0
; CHECK-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov.16 q1[2], r0
; CHECK-NEXT: vrintr.f16 s8, s8
; CHECK-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vrintr.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov.16 q1[3], r0
; CHECK-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov.16 q1[4], r0
; CHECK-NEXT: vrintr.f16 s8, s8
; CHECK-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-NEXT: vrintr.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-NEXT: vmov.16 q1[5], r0
; CHECK-NEXT: vmov r0, s8
; CHECK-NEXT: vmov.16 q1[6], r0
; CHECK-NEXT: vmov r0, s0
; CHECK-NEXT: vmov.16 q1[7], r0
; CHECK-NEXT: vmov q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.nearbyint.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @fnearbyint_float64_t(<2 x double> %src) {
; CHECK-LABEL: fnearbyint_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl nearbyint
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl nearbyint
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.nearbyint.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
define arm_aapcs_vfpcc <4 x float> @ffloor_float32_t(<4 x float> %src) {
; CHECK-MVE-LABEL: ffloor_float32_t:
; CHECK-MVE: @ %bb.0: @ %entry
; CHECK-MVE-NEXT: vrintm.f32 s7, s3
; CHECK-MVE-NEXT: vrintm.f32 s6, s2
; CHECK-MVE-NEXT: vrintm.f32 s5, s1
; CHECK-MVE-NEXT: vrintm.f32 s4, s0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: ffloor_float32_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintm.f32 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.floor.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @ffloor_float16_t(<8 x half> %src) {
; CHECK-MVE-LABEL: ffloor_float16_t:
; CHECK-MVE: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s4, s0
; CHECK-MVE-NEXT: vrintm.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vrintm.f16 s4, s4
; CHECK-MVE-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintm.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s0, s3
; CHECK-MVE-NEXT: vmov.16 q1[0], r1
; CHECK-MVE-NEXT: vrintm.f16 s0, s0
; CHECK-MVE-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[2], r0
; CHECK-MVE-NEXT: vrintm.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintm.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[3], r0
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[4], r0
; CHECK-MVE-NEXT: vrintm.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrintm.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[5], r0
; CHECK-MVE-NEXT: vmov r0, s8
; CHECK-MVE-NEXT: vmov.16 q1[6], r0
; CHECK-MVE-NEXT: vmov r0, s0
; CHECK-MVE-NEXT: vmov.16 q1[7], r0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: ffloor_float16_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrintm.f16 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.floor.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @ffloor_float64_t(<2 x double> %src) {
; CHECK-LABEL: ffloor_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl floor
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl floor
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.floor.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
define arm_aapcs_vfpcc <4 x float> @fround_float32_t(<4 x float> %src) {
; CHECK-MVE-LABEL: fround_float32_t:
; CHECK-MVE: @ %bb.0: @ %entry
; CHECK-MVE-NEXT: vrinta.f32 s7, s3
; CHECK-MVE-NEXT: vrinta.f32 s6, s2
; CHECK-MVE-NEXT: vrinta.f32 s5, s1
; CHECK-MVE-NEXT: vrinta.f32 s4, s0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: fround_float32_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrinta.f32 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <4 x float> @llvm.round.v4f32(<4 x float> %src)
ret <4 x float> %0
}
define arm_aapcs_vfpcc <8 x half> @fround_float16_t(<8 x half> %src) {
; CHECK-MVE-LABEL: fround_float16_t:
; CHECK-MVE: @ %bb.0: @ %entry
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s4, s0
; CHECK-MVE-NEXT: vrinta.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vrinta.f16 s4, s4
; CHECK-MVE-NEXT: vmov r0, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrinta.f16 s4, s0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r1, s4
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s0, s3
; CHECK-MVE-NEXT: vmov.16 q1[0], r1
; CHECK-MVE-NEXT: vrinta.f16 s0, s0
; CHECK-MVE-NEXT: vmov.16 q1[1], r0
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s1
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[2], r0
; CHECK-MVE-NEXT: vrinta.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrinta.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[3], r0
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vmovx.f16 s8, s2
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[4], r0
; CHECK-MVE-NEXT: vrinta.f16 s8, s8
; CHECK-MVE-NEXT: vmov r0, s8
[MVE] VMOVX patterns This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some adjustments for MVE. It allows us to move fp16 registers without going into and out of gprs. VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits of another register, zeroing the rest. This can be used for odd MVE register lanes. The top bits are not read by fp16 instructions, so no move is required there if we are dealing with even lanes. Differential revision: https://reviews.llvm.org/D66793 llvm-svn: 370184
2019-08-28 18:13:23 +08:00
; CHECK-MVE-NEXT: vrinta.f16 s8, s3
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
; CHECK-MVE-NEXT: vmov.16 q1[5], r0
; CHECK-MVE-NEXT: vmov r0, s8
; CHECK-MVE-NEXT: vmov.16 q1[6], r0
; CHECK-MVE-NEXT: vmov r0, s0
; CHECK-MVE-NEXT: vmov.16 q1[7], r0
; CHECK-MVE-NEXT: vmov q0, q1
; CHECK-MVE-NEXT: bx lr
;
; CHECK-MVEFP-LABEL: fround_float16_t:
; CHECK-MVEFP: @ %bb.0: @ %entry
; CHECK-MVEFP-NEXT: vrinta.f16 q0, q0
; CHECK-MVEFP-NEXT: bx lr
entry:
%0 = call fast <8 x half> @llvm.round.v8f16(<8 x half> %src)
ret <8 x half> %0
}
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
define arm_aapcs_vfpcc <2 x double> @fround_float64_t(<2 x double> %src) {
; CHECK-LABEL: fround_float64_t:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: .save {r7, lr}
; CHECK-NEXT: push {r7, lr}
; CHECK-NEXT: .vsave {d8, d9}
; CHECK-NEXT: vpush {d8, d9}
; CHECK-NEXT: vmov q4, q0
; CHECK-NEXT: vmov r0, r1, d9
; CHECK-NEXT: bl round
; CHECK-NEXT: vmov r2, r3, d8
; CHECK-NEXT: vmov d9, r0, r1
; CHECK-NEXT: mov r0, r2
; CHECK-NEXT: mov r1, r3
; CHECK-NEXT: bl round
; CHECK-NEXT: vmov d8, r0, r1
; CHECK-NEXT: vmov q0, q4
; CHECK-NEXT: vpop {d8, d9}
; CHECK-NEXT: pop {r7, pc}
entry:
%0 = call fast <2 x double> @llvm.round.v2f64(<2 x double> %src)
ret <2 x double> %0
}
[ARM] MVE VRINT support This adds support for the floor/ceil/trunc/... series of instructions, converting to various forms of VRINT. They use the same suffixes as their floating point counterparts. There is not VTINTR, so nearbyint is expanded. Also added a copysign test, to show it is expanded. Differential Revision: https://reviews.llvm.org/D63985 llvm-svn: 366003
2019-07-13 22:38:53 +08:00
declare <4 x float> @llvm.ceil.v4f32(<4 x float>)
declare <4 x float> @llvm.trunc.v4f32(<4 x float>)
declare <4 x float> @llvm.rint.v4f32(<4 x float>)
declare <4 x float> @llvm.nearbyint.v4f32(<4 x float>)
declare <4 x float> @llvm.floor.v4f32(<4 x float>)
declare <4 x float> @llvm.round.v4f32(<4 x float>)
declare <8 x half> @llvm.ceil.v8f16(<8 x half>)
declare <8 x half> @llvm.trunc.v8f16(<8 x half>)
declare <8 x half> @llvm.rint.v8f16(<8 x half>)
declare <8 x half> @llvm.nearbyint.v8f16(<8 x half>)
declare <8 x half> @llvm.floor.v8f16(<8 x half>)
declare <8 x half> @llvm.round.v8f16(<8 x half>)
[ARM] MVE vector for 64bit types We need to make sure that we are sensibly dealing with vectors of types v2i64 and v2f64, even if most of the time we cannot generate native operations for them. This mostly adds a lot of testing, plus fixes up a couple of the issues found. And, or and xor can be legal for v2i64, and shifts combining needs a slight fixup. Differential Revision: https://reviews.llvm.org/D64316 llvm-svn: 366106
2019-07-16 02:42:54 +08:00
declare <2 x double> @llvm.ceil.v2f64(<2 x double>)
declare <2 x double> @llvm.trunc.v2f64(<2 x double>)
declare <2 x double> @llvm.rint.v2f64(<2 x double>)
declare <2 x double> @llvm.nearbyint.v2f64(<2 x double>)
declare <2 x double> @llvm.floor.v2f64(<2 x double>)
declare <2 x double> @llvm.round.v2f64(<2 x double>)