Chris Elrod

Disappointing, I really expected this to be good enough: ```julia function test!(x, a, beta) @inbounds @simd ivdep for n = 1:size(a,2) x[n] += beta * conj(a[n]) end end N =...

I added the ability to manually specify the unroll factor in v0.3.8, via specifying `unroll = U` or `tile = (U,T)` where U and T are integer literals (ie, they...

`unroll=8` should be unnecessary. It should work for multidimensional arrays.

> which could be made more convenient by using Note that as of Julia 1.6, you can use `reinterpret(reshape, real(typeof(x)), x)` instead, to combine it into one step. Doing this...

> Hi, on "which falls behind MKL's gemm beyond 70x70 or so" is that mostly outdated text? Yes, see how far things have come since the [ANN: LoopVectorization](https://discourse.julialang.org/t/ann-loopvectorization/32843) post on...

There's another issue like this somewhere, but I couldn't find it with a brief check. You can do some variant of: ```julia function f(X::Matrix{Float64}) retval = 1 @turbo for i...

This is reminiscent of: https://github.com/chriselrod/LoopVectorization.jl/issues/7 But I'll make that one about supporting `StaticArrays`, and this one about not throwing an error and instead using a fallback. The problem here is...

For some reason I had restricted `vmaterialize!` to only accept `T @avx a[indices] .= items ERROR: conversion to pointer not defined for SubArray{Int64,1,Array{Int64,1},Tuple{Array{Int64,1}},false} Stacktrace: [1] error(::String) at ./error.jl:33 [2] unsafe_convert(::Type{Ptr{Int64}},...

I'll probably hold off on this until the triangular loops update, which is probably at least several months away. But if anyone else would like to take a stab at...

`K` is not defined. I assume you're missing the line ```julia K = size(q,1) ``` ? I'll have to look more later, but I get ```julia julia> @time f(g, qq,...