Re: [GHC] #7116: Missing optimisation: strength reduction of floating-point multiplication