875afa939d
This change is part of this proposal: https://discourse.llvm.org/t/rfc-all-the-math-intrinsics/78294 Based on example PR #96222 and fix PR #101268, with some differences due to 2-arg intrinsic and intermediate refactor (RuntimeLibCalls.cpp). - Add llvm.experimental.constrained.atan2 - Intrinsics.td, ConstrainedOps.def, LangRef.rst - Add to ISDOpcodes.h and TargetSelectionDAG.td, connect to intrinsic in BasicTTIImpl.h, and LibFunc_ in SelectionDAGBuilder.cpp - Update LegalizeDAG.cpp, LegalizeFloatTypes.cpp, LegalizeVectorOps.cpp, and LegalizeVectorTypes.cpp - Update isKnownNeverNaN in SelectionDAG.cpp - Update SelectionDAGDumper.cpp - Update libcalls - RuntimeLibcalls.def, RuntimeLibcalls.cpp - TargetLoweringBase.cpp - Expand for vectors, promote f16 - X86ISelLowering.cpp - Expand f80, promote f32 to f64 for MSVC Part 4 for Implement the atan2 HLSL Function #70096.
81 lines
2.7 KiB
LLVM
81 lines
2.7 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck %s
|
|
|
|
define half @use_atan2f16(half %a, half %b) nounwind {
|
|
; CHECK-LABEL: use_atan2f16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill
|
|
; CHECK-NEXT: movaps %xmm1, %xmm0
|
|
; CHECK-NEXT: callq __extendhfsf2@PLT
|
|
; CHECK-NEXT: movss %xmm0, (%rsp) # 4-byte Spill
|
|
; CHECK-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 4-byte Reload
|
|
; CHECK-NEXT: # xmm0 = mem[0],zero,zero,zero
|
|
; CHECK-NEXT: callq __extendhfsf2@PLT
|
|
; CHECK-NEXT: movss (%rsp), %xmm1 # 4-byte Reload
|
|
; CHECK-NEXT: # xmm1 = mem[0],zero,zero,zero
|
|
; CHECK-NEXT: callq atan2f@PLT
|
|
; CHECK-NEXT: callq __truncsfhf2@PLT
|
|
; CHECK-NEXT: popq %rax
|
|
; CHECK-NEXT: retq
|
|
%x = call half @llvm.atan2.f16(half %a, half %b)
|
|
ret half %x
|
|
}
|
|
|
|
define float @use_atan2f32(float %a, float %b) nounwind {
|
|
; CHECK-LABEL: use_atan2f32:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: jmp atan2f@PLT # TAILCALL
|
|
%x = call float @llvm.atan2.f32(float %a, float %b)
|
|
ret float %x
|
|
}
|
|
|
|
define double @use_atan2f64(double %a, double %b) nounwind {
|
|
; CHECK-LABEL: use_atan2f64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: jmp atan2@PLT # TAILCALL
|
|
%x = call double @llvm.atan2.f64(double %a, double %b)
|
|
ret double %x
|
|
}
|
|
|
|
define x86_fp80 @use_atan2f80(x86_fp80 %a, x86_fp80 %b) nounwind {
|
|
; CHECK-LABEL: use_atan2f80:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: subq $40, %rsp
|
|
; CHECK-NEXT: fldt {{[0-9]+}}(%rsp)
|
|
; CHECK-NEXT: fldt {{[0-9]+}}(%rsp)
|
|
; CHECK-NEXT: fstpt {{[0-9]+}}(%rsp)
|
|
; CHECK-NEXT: fstpt (%rsp)
|
|
; CHECK-NEXT: callq atan2l@PLT
|
|
; CHECK-NEXT: addq $40, %rsp
|
|
; CHECK-NEXT: retq
|
|
%x = call x86_fp80 @llvm.atan2.f80(x86_fp80 %a, x86_fp80 %b)
|
|
ret x86_fp80 %x
|
|
}
|
|
|
|
define fp128 @use_atan2fp128(fp128 %a, fp128 %b) nounwind {
|
|
; CHECK-LABEL: use_atan2fp128:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: jmp atan2f128@PLT # TAILCALL
|
|
%x = call fp128 @llvm.atan2.f128(fp128 %a, fp128 %b)
|
|
ret fp128 %x
|
|
}
|
|
|
|
define ppc_fp128 @use_atan2ppc_fp128(ppc_fp128 %a, ppc_fp128 %b) nounwind {
|
|
; CHECK-LABEL: use_atan2ppc_fp128:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: callq atan2l@PLT
|
|
; CHECK-NEXT: popq %rax
|
|
; CHECK-NEXT: retq
|
|
%x = call ppc_fp128 @llvm.atan2.ppcf128(ppc_fp128 %a, ppc_fp128 %b)
|
|
ret ppc_fp128 %x
|
|
}
|
|
|
|
declare half @llvm.atan2.f16(half, half)
|
|
declare float @llvm.atan2.f32(float, float)
|
|
declare double @llvm.atan2.f64(double, double)
|
|
declare x86_fp80 @llvm.atan2.f80(x86_fp80, x86_fp80)
|
|
declare fp128 @llvm.atan2.f128(fp128, fp128)
|
|
declare ppc_fp128 @llvm.atan2.ppcf128(ppc_fp128, ppc_fp128)
|