Hard Processor System Technical Reference Manual: Agilex™ 5 SoCs

ID 814346
Date 7/19/2024
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

3.5.3.8. Advanced Single Instruction Multiple Data and Floating Point Support

The Cortex* -A76 core supports the advanced single instruction, multiple data (SIMD) and scalar floating-point instructions in the A64 instruction set and the advanced SIMD and floating-point instructions in the A32 and T32 instruction sets.

The Cortex* -A76 floating-point implementation:

  • Does not generate floating-point exceptions.
  • Implements all scalar operations in hardware with support for all combinations of:
    • Rounding modes
    • Flush-to-zero
    • Default not a number (NaN) modes

Software can identify the advanced SIMD and floating-point features using the feature identification registers in the AArch64 execution state only. The Cortex* -A76 core only supports AArch32 in EL0, therefore none of the feature identification registers are accessible in the AArch32 execution state.