Mersenne Twister 64: origin of the "d" tempering parameter?

Question

Short version: common implementations of MT (and MT64) have an additional drop parameter d. The documents that I considered when describing the MT are not taken into account. Where is it from? When was it added?

Long version! :)

I am considering using the MT family with different parameters (for reasons). Nishimura published this document describing the implementation of 64-bit MT and a number of alternatives for matrix A and (u, s, t, l, b, c) tempering parameters. Like the original 32-bit MT paper , this describes the tempering procedure as :

y := x xor (x >> u) y := x xor ((y << s) and b) y := x xor ((y << t) and c) z := x xor (y >> l) 

However, real implementations (<a href = "> a b ) and the parameters described on the MT Wikipedia page have an additional mask mask parameter (called" d ") applied to the first stage of the vacation procedure:

 x ^= (x >> 29) & 0x5555555555555555ULL; x ^= (x << 17) & 0x71D67FFFEDA60000ULL; x ^= (x << 37) & 0xFFF7EEE000000000ULL; x ^= (x >> 43); 

I don’t want to just “blindly” use vacation parameters that exclude the “d” parameter, not understanding what it is for ... but I can’t find links to why this “d” parameter was added. (Presumably this is an improvement ...)

Any ideas on what this is for, any links to why it was added?