On 06/15/2016 06:05 AM, Richard Biener wrote:
So I've gone over the patches and gave mostly high-level comments.
The vectorizer is already in somewhat messy (aka not easy to follow)
state, this series doesn't improve the situation (heh). Esp. the
high-level structure for code generation and its documentation needs
work (where we do versioning / peeling and how we use the copies in
which condition and where, etc).
Expecting major improvements here may not be realistic. I think the
question we need to answer is whether or not the improvements from this
work justify the added complexity.
In an ideal world, I think we'd probably start over on the vectorizer,
but the infrastructure we have is what it is -- a tangled mess that is
difficult to understand without working in it regularly.
I'm still hoping to give this stuff a high level looksie before going on
PTO later this month.
Jeff
