[webkit-dev] jit, unordered compare
ggaren at apple.com
Fri Oct 16 09:23:25 PDT 2009
> fortunately, no, I don't need double branch here. "cmpvs reg0, reg0"
> is a
> conditional instruction, which only executes if the v flag is set.
> My proposal would be to add a new DoubleCondition called
> or something similar to clarify what we expect from the conditional
> branch. Since I feel (and by IEEE as you mentioned) a comparison to
> should be false, not true as x86 actually does.
Yes, I think this is a great approach.
I would slightly prefer "DoubleEqualOrUnordered" to "NaN", since
"unordered" matches the terminology of assembly languages.
I haven't thought this through fully, but you'll probably need
DoubleNotEqualAndNotUnordered as well.
More information about the webkit-dev