Why tonemapping operation is not baked into LUT when using a LDR tonemapping?(URP ppstack)

In URP's color grading, when color grading mode is set to "High Dynamic Range", tonemapping will be applied in LUT building pass, which means these complex color transform calculations will be only be applied to 16x16x16 pixels, and in postprocess pass, only several LUT sample will be applied on full screen image. But when color grading mode is set to "Low Dynamic Range", tonemapping will be applied in postprocess pass, which means these complex color transform calculations will be applied on 1920x1080 pixels and that is 500+ times more than in HDR mode. I wonder why not baking these tonemapping transform into LUT and get a better performance like it is in HDR mode?