mirror of
https://github.com/luau-lang/luau.git
synced 2024-11-15 14:25:44 +08:00
9aa82c6fb9
When the input is a constant, we use a fairly inefficient sequence of fmov+fcvt+dup or, when the double isn't encodable in fmov, adr+ldr+fcvt+dup. Instead, we can use the same lowering as X64 when the input is a constant, and load the vector from memory. However, if the constant is encodable via fmov, we can use a vector fmov instead (which is just one instruction and doesn't need constant space). Fortunately the bit encoding of fmov for 32-bit floating point numbers matches that of 64-bit: the decoding algorithm is a little different because it expands into a larger exponent, but the values are compatible, so if a double can be encoded into a scalar fmov with a given abcdefgh pattern, the same pattern should encode the same float; due to the very limited number of mantissa and exponent bits, all values that are encodable are also exact in both 32-bit and 64-bit floats. This strategy is ~same as what gcc uses. For complex vectors, we previously used 4 instructions and 8 bytes of constant storage, and now we use 2 instructions and 16 bytes of constant storage, so the memory footprint is the same; for simple vectors we just need 1 instruction (4 bytes). clang lowers vector constants a little differently, opting to synthesize a 64-bit integer using 4 instructions (mov/movk) and then move it to the vector register - this requires 5 instructions and 20 bytes, vs ours/gcc 2 instructions and 8+16=24 bytes. I tried a simpler version of this that would be more compact - synthesize a 32-bit integer constant with mov+movk, and move it to vector register via dup.4s - but this was a little slower on M2, so for now we prefer the slightly larger version as it's not a regression vs current implementation. On the vector approximation benchmark we get: - Before this PR (flag=false): ~7.85 ns/op - After this PR (flag=true): ~7.74 ns/op - After this PR, with 0.125 instead of 0.123 in the benchmark code (to use fmov): ~7.52 ns/op - Not part of this PR, but the mov/dup strategy described above: ~8.00 ns/op |
||
|---|---|---|
| .. | ||
| conformance | ||
| require | ||
| AssemblyBuilderA64.test.cpp | ||
| AssemblyBuilderX64.test.cpp | ||
| AstJsonEncoder.test.cpp | ||
| AstQuery.test.cpp | ||
| AstQueryDsl.cpp | ||
| AstQueryDsl.h | ||
| AstVisitor.test.cpp | ||
| Autocomplete.test.cpp | ||
| BuiltinDefinitions.test.cpp | ||
| ClassFixture.cpp | ||
| ClassFixture.h | ||
| CodeAllocator.test.cpp | ||
| Compiler.test.cpp | ||
| Config.test.cpp | ||
| Conformance.test.cpp | ||
| ConstraintGeneratorFixture.cpp | ||
| ConstraintGeneratorFixture.h | ||
| ConstraintSolver.test.cpp | ||
| CostModel.test.cpp | ||
| DataFlowGraph.test.cpp | ||
| DenseHash.test.cpp | ||
| DiffAsserts.cpp | ||
| DiffAsserts.h | ||
| Differ.test.cpp | ||
| Error.test.cpp | ||
| Fixture.cpp | ||
| Fixture.h | ||
| Frontend.test.cpp | ||
| InsertionOrderedMap.test.cpp | ||
| IostreamOptional.h | ||
| IrBuilder.test.cpp | ||
| IrCallWrapperX64.test.cpp | ||
| IrLowering.test.cpp | ||
| IrRegAllocX64.test.cpp | ||
| JsonEmitter.test.cpp | ||
| Lexer.test.cpp | ||
| Linter.test.cpp | ||
| LValue.test.cpp | ||
| main.cpp | ||
| Module.test.cpp | ||
| NonstrictMode.test.cpp | ||
| NonStrictTypeChecker.test.cpp | ||
| Normalize.test.cpp | ||
| NotNull.test.cpp | ||
| Parser.test.cpp | ||
| RegisterCallbacks.cpp | ||
| RegisterCallbacks.h | ||
| Repl.test.cpp | ||
| RequireByString.test.cpp | ||
| RequireTracer.test.cpp | ||
| RuntimeLimits.test.cpp | ||
| ScopedFlags.h | ||
| Set.test.cpp | ||
| Simplify.test.cpp | ||
| StringUtils.test.cpp | ||
| Subtyping.test.cpp | ||
| Symbol.test.cpp | ||
| ToDot.test.cpp | ||
| TopoSort.test.cpp | ||
| ToString.test.cpp | ||
| Transpiler.test.cpp | ||
| TxnLog.test.cpp | ||
| TypeFamily.test.cpp | ||
| TypeInfer.aliases.test.cpp | ||
| TypeInfer.annotations.test.cpp | ||
| TypeInfer.anyerror.test.cpp | ||
| TypeInfer.builtins.test.cpp | ||
| TypeInfer.cfa.test.cpp | ||
| TypeInfer.classes.test.cpp | ||
| TypeInfer.definitions.test.cpp | ||
| TypeInfer.functions.test.cpp | ||
| TypeInfer.generics.test.cpp | ||
| TypeInfer.intersectionTypes.test.cpp | ||
| TypeInfer.loops.test.cpp | ||
| TypeInfer.modules.test.cpp | ||
| TypeInfer.negations.test.cpp | ||
| TypeInfer.oop.test.cpp | ||
| TypeInfer.operators.test.cpp | ||
| TypeInfer.primitives.test.cpp | ||
| TypeInfer.provisional.test.cpp | ||
| TypeInfer.refinements.test.cpp | ||
| TypeInfer.singletons.test.cpp | ||
| TypeInfer.tables.test.cpp | ||
| TypeInfer.test.cpp | ||
| TypeInfer.tryUnify.test.cpp | ||
| TypeInfer.typePacks.test.cpp | ||
| TypeInfer.typestates.test.cpp | ||
| TypeInfer.unionTypes.test.cpp | ||
| TypeInfer.unknownnever.test.cpp | ||
| TypePack.test.cpp | ||
| TypePath.test.cpp | ||
| TypeVar.test.cpp | ||
| Unifier2.test.cpp | ||
| Variant.test.cpp | ||
| VecDeque.test.cpp | ||
| VisitType.test.cpp | ||