Package org.graalvm.compiler.lir.amd64

Interface Summary

Interface	Description
AMD64ArithmeticLIRGeneratorTool	This interface can be used to generate AMD64 LIR for arithmetic operations.

Class Summary

Class	Description
AMD64AddressValue
AMD64Arithmetic.FPDivRemOp
AMD64ArrayCompareToOp	Emits code which compares two arrays lexicographically.
AMD64ArrayCopyWithConversionsOp	Arraycopy operation for arbitrary source and destination arrays, with arbitrary byte offset, with support for arbitrary compression and inflation of 8 bit, 16 bit or 32 bit array elements.
AMD64ArrayEqualsOp	Emits code which compares two arrays of the same length.
AMD64ArrayIndexOfOp
AMD64ArrayRegionCompareToOp	Emits code to compare two array regions lexicographically, using SIMD instructions where possible.
AMD64Binary	AMD64 LIR instructions that have two inputs and one output.
AMD64Binary.CommutativeTwoOp	Commutative instruction that has two AllocatableValue operands.
AMD64Binary.ConstOp	Instruction that has one AllocatableValue operand and one 32-bit immediate operand.
AMD64Binary.DataTwoOp	Instruction that has one AllocatableValue operand and one DataSectionReference operand.
AMD64Binary.MemoryTwoOp	Instruction that has one AllocatableValue operand and one memory operand.
AMD64Binary.RMIOp	Instruction with a separate result operand, one AllocatableValue input and one 32-bit immediate input.
AMD64Binary.TwoOp	Instruction that has two AllocatableValue operands.
AMD64BinaryConsumer	AMD64 LIR instructions that have two input operands, but no output operand.
AMD64BinaryConsumer.ConstOp	Instruction that has one AllocatableValue operand and one 32-bit immediate operand.
AMD64BinaryConsumer.DataOp	Instruction that has one AllocatableValue operand and one DataSectionReference operand.
AMD64BinaryConsumer.MemoryConstOp	Instruction that has one memory operand and one 32-bit immediate operand.
AMD64BinaryConsumer.MemoryMROp	Instruction that has a memory operand as first input and an AllocatableValue as second input.
AMD64BinaryConsumer.MemoryRMOp	Instruction that has an AllocatableValue as first input and a memory operand as second input.
AMD64BinaryConsumer.MemoryVMConstOp	Instruction that has one memory operand and one 32-bit immediate operand that needs to be patched at runtime.
AMD64BinaryConsumer.Op	Instruction that has two AllocatableValue operands.
AMD64BinaryConsumer.VMConstOp	Instruction that has one AllocatableValue operand and one 32-bit immediate operand that needs to be patched at runtime.
AMD64BlockEndOp
AMD64BreakpointOp	Emits a breakpoint.
AMD64ByteSwapOp
AMD64CacheWritebackOp	Implements jdk.internal.misc.Unsafe.writeback0(long).
AMD64CacheWritebackPostSyncOp	Implements jdk.internal.misc.Unsafe.writebackPostSync0(long).
AMD64CalcStringAttributesOp	This intrinsic calculates the code range and codepoint length of strings in various encodings.
AMD64Call
AMD64Call.CallOp
AMD64Call.DirectCallOp
AMD64Call.DirectFarForeignCallOp
AMD64Call.DirectNearForeignCallOp
AMD64Call.ForeignCallOp
AMD64Call.IndirectCallOp
AMD64Call.MethodCallOp
AMD64CCall
AMD64ClearRegisterOp
AMD64ComplexVectorOp	Base class for AMD64 LIR instruction using AVX CPU features.
AMD64ControlFlow
AMD64ControlFlow.BranchOp
AMD64ControlFlow.CmpBranchOp
AMD64ControlFlow.CmpConstBranchOp
AMD64ControlFlow.CmpDataBranchOp
AMD64ControlFlow.CondMoveOp
AMD64ControlFlow.CondSetOp
AMD64ControlFlow.FloatBranchOp
AMD64ControlFlow.FloatCondMoveOp
AMD64ControlFlow.FloatCondSetOp
AMD64ControlFlow.HashTableSwitchOp
AMD64ControlFlow.RangeTableSwitchOp
AMD64ControlFlow.StrategySwitchOp
AMD64ControlFlow.TestBranchOp
AMD64ControlFlow.TestByteBranchOp
AMD64ControlFlow.TestConstBranchOp
AMD64EncodeArrayOp
AMD64FrameMap	AMD64 specific frame map.
AMD64FrameMapBuilder
AMD64HasNegativesOp
AMD64HotSpotHelper
AMD64LFenceOp
AMD64LIRInstruction	Convenience class to provide AMD64MacroAssembler for the AMD64LIRInstruction.emitCode(org.graalvm.compiler.lir.asm.CompilationResultBuilder) method.
AMD64MathCopySignOp
AMD64MathCosOp	ALGORITHM DESCRIPTION - COS() --------------------- 1.
AMD64MathExpOp	ALGORITHM DESCRIPTION - EXP() --------------------- Description: Let K = 64 (table size).
AMD64MathIntrinsicBinaryOp	AMD64MathIntrinsicBinaryOp assumes that the input values are stored in the xmm0 and xmm1 registers, and it will emit the output value into the xmm0 register.
AMD64MathIntrinsicUnaryOp	AMD64MathIntrinsicUnaryOp assumes that the input value is stored at the xmm0 register, and will emit the output value into the xmm0 register as well.
AMD64MathLog10Op	ALGORITHM DESCRIPTION - LOG10() --------------------- Let x=2^k * mx, mx in [1,2) Get B~1/mx based on the output of rcpss instruction (B0) B = int((B0*LH*2^7+0.5))/2^7 LH is a short approximation for log10(e) Reduced argument: r=B*mx-LH (computed accurately in high and low parts) Result: k*log10(2) - log(B) + p(r) p(r) is a degree 7 polynomial -log(B) read from data table (high, low parts) Result is formed from high and low parts.
AMD64MathLogOp	ALGORITHM DESCRIPTION - LOG() --------------------- x=2^k * mx, mx in [1,2) Get B~1/mx based on the output of rcpss instruction (B0) B = int((B0*2^7+0.5))/2^7 Reduced argument: r=B*mx-1.0 (computed accurately in high and low parts) Result: k*log(2) - log(B) + p(r) if |x-1| >= small value (2^-6) and p(r) is a degree 7 polynomial -log(B) read from data table (high, low parts) Result is formed from high and low parts.
AMD64MathPowOp	ALGORITHM DESCRIPTION - POW() --------------------- Let x=2^k * mx, mx in [1,2) log2(x) calculation: Get B~1/mx based on the output of rcpps instruction (B0) B = int((B0*LH*2^9+0.5))/2^9 LH is a short approximation for log2(e) Reduced argument, scaled by LH: r=B*mx-LH (computed accurately in high and low parts) log2(x) result: k - log2(B) + p(r) p(r) is a degree 8 polynomial -log2(B) read from data table (high, low parts) log2(x) is formed from high and low parts For |x| in [1-1/32, 1+1/16), a slower but more accurate computation based om the same table design is performed.
AMD64MathSignumOp
AMD64MathSinOp	ALGORITHM DESCRIPTION - SIN() --------------------- 1.
AMD64MathTanOp	ALGORITHM DESCRIPTION - TAN() --------------------- Polynomials coefficients and other constants.
AMD64Move
AMD64Move.AMD64MultiStackMove
AMD64Move.AMD64PushPopStackMove
AMD64Move.AMD64StackMove
AMD64Move.AtomicReadAndAddOp
AMD64Move.AtomicReadAndWriteOp
AMD64Move.CompareAndSwapOp
AMD64Move.CompressPointerOp
AMD64Move.ConvertNullToZeroOp
AMD64Move.ConvertZeroToNullOp
AMD64Move.LeaDataOp
AMD64Move.LeaOp
AMD64Move.MembarOp
AMD64Move.MoveFromConstOp
AMD64Move.MoveFromRegOp
AMD64Move.MoveToRegOp
AMD64Move.NullCheckOp
AMD64Move.PointerCompressionOp
AMD64Move.StackLeaOp
AMD64Move.UncompressPointerOp
AMD64MulDivOp	AMD64 mul/div operation.
AMD64PauseOp	Emits a pause.
AMD64PrefetchOp
AMD64ReadTimestampCounter	AMD64 rdtsc operation.
AMD64ReadTimestampCounterWithProcid	AMD64 rdtscp operation.
AMD64RestoreRegistersOp	Restores registers from stack slots.
AMD64RoundFloatToIntegerOp	Round float to integer.
AMD64SaveRegistersOp	Saves registers to stack slots.
AMD64ShiftOp	AMD64 shift/rotate operation.
AMD64SignExtendOp
AMD64StrideUtil	This class provides utility methods for "stride-agnostic" intrinsic ops such as AMD64ArrayRegionCompareToOp.
AMD64StringLatin1InflateOp
AMD64StringUTF16CompressOp
AMD64Ternary	AMD64 LIR instructions that have three inputs and one output.
AMD64Ternary.ThreeOp	Instruction that has two AllocatableValue operands.
AMD64Unary	AMD64 LIR instructions that have one input and one output.
AMD64Unary.MemoryOp	Instruction with a memory operand.
AMD64Unary.MOp	Instruction with a single operand that is both input and output.
AMD64Unary.MROp	Instruction with separate input and output operands, and an operand encoding of MR.
AMD64Unary.RMOp	Instruction with separate input and output operands, and an operand encoding of RM.
AMD64UnaryConsumer	AMD64 LIR instructions that have one input and no output.
AMD64UnaryConsumer.MemoryOp
AMD64VZeroUpper	vzeroupper is essential to avoid performance penalty during SSE-AVX transition.
AMD64ZapRegistersOp	Writes well known garbage values to registers.
AMD64ZapStackOp	Writes well known garbage values to stack slots.
AMD64ZeroMemoryOp	Zeros a chunk of memory using rep stosb.

Enum Summary

Enum	Description
AMD64Arithmetic
AMD64CalcStringAttributesOp.Op