Skip to content
STIMSMITH

FLI.D

ISA WIKI v1 · 9/27/2026

FLI.D is a Zfa-extension instruction that loads a floating-point constant into the double-precision floating-point register rd. It is part of the Zfa extension, which defines FLI.[HSD] variants for half-, single-, and double-precision floating-point formats.

FLI.D

Overview

FLI.D is a floating-point load-immediate instruction introduced as part of the RISC-V Zfa extension. It writes a constant floating-point value to the double-precision (.D) floating-point register rd.

Extension Context

FLI.D is one of the FLI.[HSD] variants added when Zfa extension support (excluding quad-precision) was introduced into the sail-riscv formal model. The broader Zfa support added in that change set includes infrastructure for the Zfa extension (such as an existence macro) as well as implementations of:

  • FLI.[HSD] — load floating-point immediate (half / single / double precision).
  • FMINM.[HSD] / FMAXM.[HSD] — min/max with adjusted NaN-handling semantics.
  • FROUND.[HSD] / FROUNDNX.[HSD] — round-to-integer instructions, with the NX variant also signaling inexact.
  • FMVH.X.D / FMVP.D.X — RV32-only moves between pairs of integer registers and a double-precision FP register.
  • FLEQ.[HSD] / FLTQ.[HSD] — quiet comparison instructions that do not signal invalid on quiet NaN inputs.
  • FCVTMOD.W.D — bounded conversion from double to signed word.

Implementation

FLI.D is implemented in the sail-riscv formal specification as part of the Zfa instruction set. The commit that adds Zfa support to sail-riscv enumerates FLI.[HSD] among the instructions introduced, together with the rest of the Zfa instruction set listed above.

Relationships

  • Part of: Zfa (RISC-V extension for additional floating-point instructions).
  • Implemented in: sail-riscv — the Sail formal specification of RISC-V, which models FLI.D alongside the rest of the Zfa extension.

See Also

  • Zfa — the parent extension providing the FLI.[HSD] family.
  • sail-riscv — the formal-model code artifact in which FLI.D is implemented.

LINKED ENTITIES

2 links
Zfa PART_OF
sail-riscv IMPLEMENTS

CITATIONS

2 sources
2 citations
[1] FLI.D is a floating-point load-immediate instruction in the double-precision format, part of the FLI.[HSD] family. Add Zfa extension support (excl. quad-precision) — sail-riscv commit f2d8e1b9
[2] FLI.D was added to sail-riscv together with infrastructure for the Zfa extension and a set of Zfa instructions including FMINM/FMAXM, FROUND/FROUNDNX, FMVH.X.D, FMVP.D.X, FLEQ/FLTQ, and FCVTMOD.W.D. Add Zfa extension support (excl. quad-precision) — sail-riscv commit f2d8e1b9