[libre-riscv-dev] FP16 FCVT needed (int, uint, fp32, fp64)
luke.leighton at gmail.com
Fri Sep 13 00:41:07 BST 2019
On Friday, September 13, 2019 at 7:34:23 AM UTC+8, lkcl wrote:
> On Thursday, September 12, 2019 at 4:18:29 AM UTC+8, Bruce Hoult wrote:
> > We do intend to have half precision FP in the scalar instruction set,
> > as it doesn't make sense to have it in the vector instruction set but
> > not scalar.
> Ah before I forget:
> Search for FCVT
> I arbitrarily chose some funct5 allocations to illustrate and ensure that it's not forgotten that FCVT FP16 opcodes will be needed.
> Int/signed, 4 opcodes:
> FP, 6 opcodes:
Re-read my own document, missed FCLASS.H, FMV.H, *face-palm* another 4 funct5 reservations needed.
More information about the libre-riscv-dev