Symmetric support for syr and syrk

[goerch]

After having the (for now) fruitless discussion I'd like to propose that at least something like

using BenchmarkTools, Random, LinearAlgebra
elt = Float64
n = 32
Random.seed!(1234)
# A = rand(elt, n, n)
A = rand(elt, n, n)
b = rand(elt, n)
println(BLAS.get_config())
t2 = BLAS.syr!('U', elt(1), b, Symmetric(copy(A)))

and this

using BenchmarkTools, Random, LinearAlgebra

elt = Float64
n = 32
Random.seed!(1234)
# A = rand(elt, n, n)
A = rand(elt, n, n)
b = Symmetric(rand(elt, n, n))
println(BLAS.get_config())
t2 = BLAS.syrk!('U', 'N', elt(1), b, elt(1), Symmetric(copy(A)))

ought to work. See JuliaLang/LinearAlgebra.jl#887.

[goerch]

Solved by comment https://discourse.julialang.org/t/mul-dispatch-to-blas-incomplete/71831/14

[goerch]

@N5N3, just for future reference, this works

using BenchmarkTools, Random, LinearAlgebra
println(BLAS.get_config())
elt = Float64
for n in (32, 64, 128, 256, 512, 1024, 2048)
    Random.seed!(1234)
    # A = rand(elt, n, n)
    A = rand(elt, n, n)
    b = rand(elt, n)
    # t1 = @benchmark BLAS.ger!(elt(1), $b, $b, a) setup = (a = copy($A))
    t1 = @benchmark BLAS.syr!('U', elt(1), $b, a) setup = (a = copy($A))
    # t2 = @benchmark BLAS.gemm!('N', 'C', true, $b, $b, true, a) setup = (a = copy($A))
    t2 = @benchmark mul!(a, $b, $b', elt(1), elt(1)) setup = (a = Symmetric(copy($A)).data)
    println(elt, '|', n, ':', ratio(median(t1), median(t2)))
end
using MKL
println(BLAS.get_config())
elt = Float64
for n in (32, 64, 128, 256, 512, 1024, 2048)
    Random.seed!(1234)
    A = rand(elt, n, n)
    b = rand(elt, n)
    
    # t1 = @benchmark BLAS.ger!(elt(1), $b, $b, a) setup = (a = copy($A))
    t1 = @benchmark BLAS.syr!('U', elt(1), $b, a) setup = (a = copy($A))
    # t2 = @benchmark BLAS.gemm!('N', 'C', true, $b, $b, true, a) setup = (a = copy($A))
    t2 = @benchmark mul!(a, $b, $b', elt(1), elt(1)) setup = (a = Symmetric(copy($A)).data)
    println(elt, '|', n, ':', ratio(median(t1), median(t2)))
end

and shows

LBTConfig([ILP64] libopenblas64_.dll)
Float64|32:TrialRatio(1814.75%)
Float64|64:TrialRatio(1595.74%)
Float64|128:TrialRatio(797.57%)
Float64|256:TrialRatio(121.17%)
Float64|512:TrialRatio(82.48%)
Float64|1024:TrialRatio(48.37%)
Float64|2048:TrialRatio(40.52%)
LBTConfig([ILP64] mkl_rt.1.dll)
Float64|32:TrialRatio(49.47%)
Float64|64:TrialRatio(51.50%)
Float64|128:TrialRatio(65.33%)
Float64|256:TrialRatio(132.03%)
Float64|512:TrialRatio(68.13%)
Float64|1024:TrialRatio(65.42%)
Float64|2048:TrialRatio(51.49%)

on my system.