Returns the constant folding rule to fold |OpFDiv| with two constants.

Define the folding rules for ordered and unordered comparison for floating point values.

Define the folding rules for subtraction, addition, multiplication, and division for floating point values.

This macro defines a |UnaryScalarFoldingRule| that performs float to integer conversion. TODO(greg-lunarg): Support for 64-bit integer types.

This function defines a |UnaryScalarFoldingRule| that performs integer to float conversion. TODO(greg-lunarg): Support for 64-bit integer types.

Folds an OpcompositeInsert where input is a composite constant.

Fold OpLogicalNot instructions that follow a comparison, if the comparison is only used by that instruction. !(a == b) = (a != b) !(a != b) = (a == b

Folds an OpDot where all of the inputs are constants to a constant. A new constant is created if necessary.

This defines a |UnaryScalarFoldingRule| that performs |OpQuantizeToF16|.

Returns the result of performing an operation on scalar constant operands. This function extracts the operand values as 32 bit words and returns the r

Function declaration validation is performed by making sure that the FunctionParameter and FunctionEnd instructions only appear inside of functions. I

Creates a new vector which removes the string if the substr is found in the instructions vector and reinserts it in the location specified by order. N

Gather the set of blocks for all the path from |entry| to |root|.

Probabilities associated with applying various transformations. Keep them in alphabetical order.