I personally don't feel like it's necessary to split the leaderboards, even though the VC version is, from my estimation, about ~1:10 faster (not including the Dedede timesave from the J v1.0 version). If other N64 runners feel strongly about this, please let me know.

As far as I can tell, from Peril1337's 2010 run onwards, the majority of top runners, including all record holders (I believe 0xwas used N64 for some of his older runs), used some VC version (i.e. regular Wii VC or Dream Collection). In this sense VC can be considered the main version to run the game on.

Looking at related games, the Kirby series does not have any leaderboards that are split, even ones where there has been competition at the top between runners of different versions with noticeably different loading times/lag (e.g. Super Star, Squeak Squad). The impression I got was that people of those respective communities either didn't care, or they were in favor of keeping the leaderboards combined. One could argue--in both directions--about how the leaderboards are arranged encourages or discourages prospective runners from picking up the game. Personally, I wanted to keep the leaderboards combined, but did runs of the N64 version to not only better gauge the time differences (e.g., I wanted to dispel the claim that VC saved possibly up to 2 minutes over N64) but to set a standard for people filtering for "N64 only."

For other N64 games, I found that more games with a VC version are actually not split (most notably Ocarina of Time, where people are doing version-exclusive glitches). The situation for some other games arose when the newly released Wii U VC version turned out to be faster, threatening to obsolete people's work put into the N64 version. If I recall correctly, those boards were not split prior to that, even when one of those games went through a period where VC was believed to be faster. One compromise I've seen, which is to have both a combined and split ranking, is purely cosmetic, in my opinion.