Re: Misc: BGBCC targeting RV64G, initial results...

On 9/30/24 1:52 AM, MitchAlsup1 wrote:

On Sat, 28 Sep 2024 1:44:10 +0000, Paul A. Clayton wrote:

[snip]

Another weird concept that came to mind would be providing an
8-bit (e.g.) field that enumerated a set of interesting
conditions.

 I use a 64-bit container of conditions

A enumeration of conditions is different from a bitmask of
conditions. An enumeration could support N-way branching in a
single instruction rather than a tree of single bit-condition
branches.
My 66000's compare result has unused space for multiple such
enumerations.
I do not know of any enumeration of conditions that would be
commonly useful. Less than, equal to, greater than might be
somewhat useful for a three-way branch. Relation to zero as well
as an explicit comparison value might be useful for some multi-way
choices.
Lack of density is also a problem for multi-way branches; the
encoding will waste space if multiple enumerated states share a
target.
The concept seemed worth mentioning even if I thought it unlikely
to be practically useful.