You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Triggered by a discussion during the internal review of Zilsd (riscv/riscv-zilsd#15), I would like to suggest the following:
The P extension will be the only extension that is not overlapping with anything and has encodings that allow defining a pair-move instruction. Zdinx also has an instruction that can be used to move a pair, but its encoding is not available when D is present.
My proposal would be for P to define a sub-extension, which contains Zilsd, Zcmlsd, and the pair move instruction.
The text was updated successfully, but these errors were encountered:
Triggered by a discussion during the internal review of Zilsd (riscv/riscv-zilsd#15), I would like to suggest the following:
The P extension will be the only extension that is not overlapping with anything and has encodings that allow defining a pair-move instruction. Zdinx also has an instruction that can be used to move a pair, but its encoding is not available when D is present.
My proposal would be for P to define a sub-extension, which contains Zilsd, Zcmlsd, and the pair move instruction.
The text was updated successfully, but these errors were encountered: