Source/Packages

Runtime.NativePhysicalProgram

packages/execution/src/Runtime/NativePhysicalProgram.alpha

1,008 lines185 declarations40.0 KiBSHA-256 e6bb0cdfb8f4

family · lines 88–96

NativePhysicalFloat64Operation

Full file
The binary64 operations the host computes with (IEEE 754, round to nearest even; x86 SSE2): the sum, difference, product and quotient of the left and right operands' words read as binary64; the square root of the right one; the left word, a natural below 2^63, as the nearest binary64; the left binary64 rounded to the nearest binary32 (its word zero-extended); the left word's low half, a binary32, as the binary64 of the same value. (Binary32 +, -, *, / and square root are these binary64 operations rounded to binary32: 53 bits hold 2 x 24 + 2, so the double rounding is exact.)
88family NativePhysicalFloat64Operation : Type 0
89constructor NativePhysicalFloat64Add
90constructor NativePhysicalFloat64Subtract
91constructor NativePhysicalFloat64Multiply
92constructor NativePhysicalFloat64Divide
93constructor NativePhysicalFloat64SquareRoot
94constructor NativePhysicalFloat64FromNatural
95constructor NativePhysicalFloat64ToBinary32
96constructor NativePhysicalFloat64FromBinary32

The compiler supplied declaration spans and resolved links from this source snapshot. This page does not assert that this file belongs to a checked closure.