No it is unnecessary in my experience. The tool has a 'Target' dropdown where you can select either 'Luminance (CIE Y)' or ;RGB/K Components' as the the target for the process without needing to extract the luminance to a separate image.
A lot of the 'make a pseudo luminance' processing is driven by the use of PhotoShop and similar, whereas most of the operations that need to target a specific component can do so in PI without splitting the image.