• Run a floating-point accuracy subgroup test.

    Parameters

    • t: GPUTest

      The base test

    • seed: number

      A seed for the PRNG

    • wgSize: number[]

      An array for the workgroup size

    • operation: string

      The subgroup operation

    • type: "f32" | "f16"

      The type (f16 or f32)

    • identity: number

      The identity for the operation

    • intervalGen: ((x, y) => FPInterval)

      A functor to generate an appropriate FPInterval for a binary operation

    Returns Promise<void>

Generated using TypeDoc