ZhengSh

And I would like to make some PRs to solve the problem 2 firstly. There are three ways to relieve the implicit tail undisturbed assumptions in in-place accumulations. 1. Split...

I agree. Indeed we should keep the x280 kernels unchanged and apply changes in a new copy. Now the isa spec of rvv has been stable, but the intrinsic spec...

@ken-unger , thank you very much for explaining some details of the design. I think the design intention of vector-isa, is to operate the vectors in their "natural" length. So...

Yes, PR #4159 has updated the tail policy. Now almost all kernels have supported rvv, except for *trsm and c/zsymv kernels. Another issue is that, if we can merge ZVL256B...

Yes they are different realizations now, at least using different unrolling. Even so maybe it's better if they are in a unified framework in future, regardless of vlen.

Hi~, are there any comments, or further things to do about this PR?

Hi, @mgates3 , I have made some updates, including the dates, authorship and the doxygen doc. Could you please continue the review?

Thank you so much for the review @mgates3 @langou . I agree with that this pr should be divided into finer sub-prs to improve the clarity. I plan to arrange...

> Do we still need to encode into two letters a lot of meaning? Why not just `(s,d,c,z)skewsymm` ? Or even plain `(s,d,c,z)skew_hermitian` for that matter? Although the trend of...

> `skewsymm` and `skewhemm` make a lot of sense to me — very clear (assuming you already know `symm` and `hemm`), whereas `kymm` and `kemm` are pretty cryptic (s**K**ew-s**Y**mmetric, s**K**ew-h**E**rmitian)....