Understanding Floating-Point Computation on AIE-ML Architecture (XAPP1406) - Explains how single-precision floating-point (FP32) computation is emulated on AMD Versal™ devices featuring the AIE-ML architecture, which natively supports only bfloat16 format. - XAPP1406
- Document ID
- XAPP1406
- Release Date
- 2026-01-28
- Revision
- 1.0 English