I think that if programs support "time" control of x nodes for y moves then it may be productive for testing.
One of the problem in testing is that sometimes hardware can be slowed down by a significant factor and even without that problem results of games at fast time control may not be reproducable even if we play nunn match.
If programs support time control of x nodes for y moves or
x nodes per game+y node per move then it may solve the problem.
of course it is not fair to give 2 programs to search the same number of nodes if the target is to find which program is better but it is possible to find average number of nodes that a program searches.
The idea is that if program A search 900,000 nodes per second when program B searches 500,000 nodes per second then you can ask program A to search 900,000,000 nodes for 40 moves and ask program B to search 500,000,000 nodes for 40 moves.
Winboard does not need to support that time control and it can use very long time control when the programs will know x and y by special parameters that they get from ini files(program A may get 900,000,000 40 when program B may get 500,000,000 and 40).
The main question is if programmers are interested in supporting the time control of x nodes for y moves or x nodes per game+y nodes per move.
I am interested in doing it only if enough programmers are interested in it.
Uri