[X86] Cast atomic vectors in IR to support floats

This commit casts floats to ints in an atomic load during AtomicExpand to support
floating point types. It also is required to support 128 bit vectors in SSE/AVX.

Author: jofrn

llvm/lib/Target/X86/X86ISelLowering.cpp

@@ -32136,6 +32136,13 @@ X86TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   }
 }
 
+TargetLowering::AtomicExpansionKind
+X86TargetLowering::shouldCastAtomicLoadInIR(LoadInst *LI) const {
+  if (LI->getType()->getScalarType()->isFloatingPointTy())
+    return AtomicExpansionKind::CastToInteger;
+  return AtomicExpansionKind::None;
+}
+
 LoadInst *
 X86TargetLowering::lowerIdempotentRMWIntoFencedLoad(AtomicRMWInst *AI) const {
   unsigned NativeWidth = Subtarget.is64Bit() ? 64 : 32;

llvm/lib/Target/X86/X86ISelLowering.h

@@ -1841,6 +1841,8 @@ namespace llvm {
     shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;
     TargetLoweringBase::AtomicExpansionKind
     shouldExpandLogicAtomicRMWInIR(AtomicRMWInst *AI) const;
+    TargetLoweringBase::AtomicExpansionKind
+    shouldCastAtomicLoadInIR(LoadInst *LI) const override;
     void emitBitTestAtomicRMWIntrinsic(AtomicRMWInst *AI) const override;
     void emitCmpArithAtomicRMWIntrinsic(AtomicRMWInst *AI) const override;
 

llvm/lib/Target/X86/X86InstrCompiler.td

@@ -1220,6 +1220,21 @@ def : Pat<(v2i64 (scalar_to_vector (i64 (atomic_load_64 addr:$src)))),
 def : Pat<(v2i64 (scalar_to_vector (i64 (atomic_load_64 addr:$src)))),
   (VMOV64toPQIZrm addr:$src)>, Requires<[HasAVX512]>;
 
+// load atomic <2 x i64>
+def : Pat<(v2i64 (atomic_load_128_v2i64 addr:$src)),
+          (MOVAPDrm addr:$src)>, Requires<[UseSSE2]>;
+def : Pat<(v2i64 (atomic_load_128_v2i64 addr:$src)),
+          (VMOVAPDrm addr:$src)>, Requires<[UseAVX]>;
+def : Pat<(v2i64 (atomic_load_128_v2i64 addr:$src)),
+          (VMOVAPDZ128rm addr:$src)>, Requires<[HasAVX512]>;
+// load atomic <4 x i32>
+def : Pat<(v4i32 (atomic_load_128_v4i32 addr:$src)),
+          (MOVAPDrm addr:$src)>, Requires<[UseSSE2]>;
+def : Pat<(v4i32 (atomic_load_128_v4i32 addr:$src)),
+          (VMOVAPDrm addr:$src)>, Requires<[UseAVX]>;
+def : Pat<(v4i32 (atomic_load_128_v4i32 addr:$src)),
+          (VMOVAPDZ128rm addr:$src)>, Requires<[HasAVX512]>;
+
 // Floating point loads/stores.
 def : Pat<(atomic_store_32 (i32 (bitconvert (f32 FR32:$src))), addr:$dst),
           (MOVSSmr addr:$dst, FR32:$src)>, Requires<[UseSSE1]>;