You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The @turbo and @avx macros will be replaced by @inbounds @fastmath macro, which definitely can't SIMD complex-complex multiplication, and from my experience it will make the downconverter kernels run 40x slower, to the point that I even removed the fallback version completely in my fork: JuliaGNSS/Acquisition.jl@ac096f5
Hopefully someone will make a replacement for LoopVectorization before 1.11
The text was updated successfully, but these errors were encountered:
LoopVectorization will be deprecated in Julia 1.11: JuliaSIMD/LoopVectorization.jl#519 because of too much maintenance burden. This is currently used in the downconverter kernels at Tracking.jl and Acquisition.jl
The
@turboand@avxmacros will be replaced by@inbounds @fastmathmacro, which definitely can't SIMD complex-complex multiplication, and from my experience it will make the downconverter kernels run 40x slower, to the point that I even removed the fallback version completely in my fork: JuliaGNSS/Acquisition.jl@ac096f5Hopefully someone will make a replacement for LoopVectorization before 1.11
The text was updated successfully, but these errors were encountered: