Skip to content

Normalize machine-size names for 64-bit and 128-bit values - #1960

Open
relokin wants to merge 2 commits into
herd:masterfrom
relokin:machsize-fix
Open

Normalize machine-size names for 64-bit and 128-bit values#1960
relokin wants to merge 2 commits into
herd:masterfrom
relokin:machsize-fix

Conversation

@relokin

@relokin relokin commented Aug 11, 2026

Copy link
Copy Markdown
Member

This PR updates the machine-size terminology so that 64-bit values use Double/double and 128-bit values use Quad/quad.

Changes include:

  • Rename the old 64-bit Quad size to Double.
  • Rename the old 128-bit S128 size to Quad.
  • Update architecture, parser, generator, scalar, variant, and litmus references to use the new names.
  • Keep s128 accepted as a compatibility alias where machine-size/variant tags are parsed.
  • Update C instruction tests and configs to use the new Quad variant naming.

The machSize that was previously named Quad corresponds to 64bits. This
is really only double the size of Word which is 32bits. For this reason,
and to avoid confusion rename it to Double.

Assisted-by: OpenAI Codex <codex@openai.com>
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Follow the naming convention of the other machSize constructors and name
the 128bit wide type to Quad.

Assisted-by: OpenAI Codex <codex@openai.com>
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
@ShaleXIONG

ShaleXIONG commented Aug 11, 2026

Copy link
Copy Markdown
Collaborator

The diy test should be fixed in 3b0d2eb in #1962 , please feel free to cherry pick. It is due to annotation name change from old q0 to new d0.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants