Hi, I am trying to understand how to use the @turbo macro from LoopVectorization.jl efficiently. I am starting with a generalised dot product, as implemented in LinearAlgebra. I basically took the code from LinearAlgebra and tried to apply the @turbo macro following the examples in its official do…

Efficient use of @turbo for linear algebra operations (LoopVectorization.jl)

learned_fool August 21, 2021, 12:46pm 3

yes, code inside @turbo annotated expressions need to be free of branches; the iterations also need to be independent to each other.

1 Like

Topic		Replies	Views
@turbo speeds routine, slows down everything else Performance loopvectorization	16	2657	June 5, 2021
@turbo macro gives incorrect results Performance loopvectorization	4	476	October 26, 2022
Utilizing @turbo / @tturbo in Performance Critical Code Performance question	4	908	September 8, 2022
@turbo on sets of operations Performance	15	1060	August 10, 2021
@turbo macro giving slightly different results General Usage loopvectorization	6	541	January 28, 2023