On 6/23/2026 12:55 PM, Robin Dapp wrote:
Hi Lino,
On whether we expect a non-min Zvfofp8 with FMA or similar: I don't know
of any concrete plan. The only FP8 vector item in the RISC-V spec
database today is Zvfofp8min (RVS-3748, currently in Freeze), which is
conversion-only (FP8<->BF16 and FP32->FP8), and I don't see a fuller
Zvfofp8 proposed yet. That said, given the precedent of the minimal
extensions (e.g. Zfhmin / Zvfhmin alongside the full Zfh / Zvfh), it
seems plausible a fuller FP8 extension could follow later.
That's my reasoning for the current approach. I'm happy for you and the
maintainers to take the mode question up at the patchwork call, and I'll
go with whatever you decide.
we discussed this today in the patchwork call and the consensus was that
separate modes seem like the more future-proof approach. I would hope that
some of the more involved parts of the approach will just disappear and make
the new pieces more similar to what we already have. (The vsetvl part which
tracks the altfmt mode is obviously still necessary).
To add a bit more color. Multiple people expressed a belief that at
some point there will likely be some limited computation instructions on
FP8 values. To support computation instructions we're going to need
real modes. Given that's seen as a likely scenario we should try to
future proof this work a bit.
jeff
