mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-24 03:33:20 +01:00
dbac87d1fc
Patch by Ana Pazos. - Completed implementation of instruction formats: AdvSIMD three same AdvSIMD modified immediate AdvSIMD scalar pairwise - Completed implementation of instruction classes (some of the instructions in these classes belong to yet unfinished instruction formats): Vector Arithmetic Vector Immediate Vector Pairwise Arithmetic - Initial implementation of instruction formats: AdvSIMD scalar two-reg misc AdvSIMD scalar three same - Intial implementation of instruction class: Scalar Arithmetic - Initial clang changes to support arm v8 intrinsics. Note: no clang changes for scalar intrinsics function name mangling yet. - Comprehensive test cases for added instructions To verify auto codegen, encoding, decoding, diagnosis, intrinsics. llvm-svn: 187567
57 lines
2.5 KiB
LLVM
57 lines
2.5 KiB
LLVM
; RUN: llc -mtriple=aarch64-none-linux-gnu -mattr=+neon < %s | FileCheck %s
|
|
|
|
declare <2 x i32> @llvm.arm.neon.vacged(<2 x float>, <2 x float>)
|
|
declare <4 x i32> @llvm.arm.neon.vacgeq(<4 x float>, <4 x float>)
|
|
declare <2 x i64> @llvm.aarch64.neon.vacgeq(<2 x double>, <2 x double>)
|
|
|
|
define <2 x i32> @facge_from_intr_v2i32(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facge_from_intr_v2i32:
|
|
%val = call <2 x i32> @llvm.arm.neon.vacged(<2 x float> %A, <2 x float> %B)
|
|
; CHECK: facge {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
|
|
ret <2 x i32> %val
|
|
}
|
|
define <4 x i32> @facge_from_intr_v4i32( <4 x float> %A, <4 x float> %B) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facge_from_intr_v4i32:
|
|
%val = call <4 x i32> @llvm.arm.neon.vacgeq(<4 x float> %A, <4 x float> %B)
|
|
; CHECK: facge {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
|
|
ret <4 x i32> %val
|
|
}
|
|
|
|
define <2 x i64> @facge_from_intr_v2i64(<2 x double> %A, <2 x double> %B) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facge_from_intr_v2i64:
|
|
%val = call <2 x i64> @llvm.aarch64.neon.vacgeq(<2 x double> %A, <2 x double> %B)
|
|
; CHECK: facge {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
|
|
ret <2 x i64> %val
|
|
}
|
|
|
|
declare <2 x i32> @llvm.arm.neon.vacgtd(<2 x float>, <2 x float>)
|
|
declare <4 x i32> @llvm.arm.neon.vacgtq(<4 x float>, <4 x float>)
|
|
declare <2 x i64> @llvm.aarch64.neon.vacgtq(<2 x double>, <2 x double>)
|
|
|
|
define <2 x i32> @facgt_from_intr_v2i32(<2 x float> %A, <2 x float> %B, <2 x float> %C) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facgt_from_intr_v2i32:
|
|
%val = call <2 x i32> @llvm.arm.neon.vacgtd(<2 x float> %A, <2 x float> %B)
|
|
; CHECK: facgt {{v[0-9]+}}.2s, {{v[0-9]+}}.2s, {{v[0-9]+}}.2s
|
|
ret <2 x i32> %val
|
|
}
|
|
define <4 x i32> @facgt_from_intr_v4i32( <4 x float> %A, <4 x float> %B) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facgt_from_intr_v4i32:
|
|
%val = call <4 x i32> @llvm.arm.neon.vacgtq(<4 x float> %A, <4 x float> %B)
|
|
; CHECK: facgt {{v[0-9]+}}.4s, {{v[0-9]+}}.4s, {{v[0-9]+}}.4s
|
|
ret <4 x i32> %val
|
|
}
|
|
|
|
define <2 x i64> @facgt_from_intr_v2i64(<2 x double> %A, <2 x double> %B) {
|
|
; Using registers other than v0, v1 and v2 are possible, but would be odd.
|
|
; CHECK: facgt_from_intr_v2i64:
|
|
%val = call <2 x i64> @llvm.aarch64.neon.vacgtq(<2 x double> %A, <2 x double> %B)
|
|
; CHECK: facgt {{v[0-9]+}}.2d, {{v[0-9]+}}.2d, {{v[0-9]+}}.2d
|
|
ret <2 x i64> %val
|
|
}
|
|
|