How big is the difference really? Another month or two travel time? Humans have spent over a year (Mir) and routinely do six month stints on the ISS. If you've got the tech to get to Mars and land there, then an extra month or two flight time is not going to kill you.
I say go as soon as we have the flight hardware built and qualified.
If you're sending humans, that's another month or two of the most expensive supplies in the history of human endeavor. It's not exactly a detail that we can currently afford to wave away.
(And I do mean afford. If we had better propulsion tech, like, oh, say, anything nuclear, we wouldn't have to worry so much about such details. But if we're going to hobble ourselves by rejecting that out of hand, we can't afford much mass.)
It's not just the problem of having astronauts hanging out in their ship playing zero-gravity poker for an extra couple months. You also need to have adequate supplies for that time and make sure the crew is protected from any additional radiation exposure they'd experience during that time.
There's something wrong with the math here. 100 km/s delta-V means you go from standstill to 100 km/s. Solar system's escape velocity is 42 km/s at Earth's orbit. By the very definition, any A->B travel for any A and B within the solar system, requires less delta-V if Hoffman's orbits are used.
In reality the delta-V for Earth_surface -> Mars_surface journey is between 19 and 21 km/s (depending on the relative position of the two bodies) for Hoffman's orbits. Without using aerobreaking. If using aerobreaking, it's less by about 3-4km/s [1]
No matter when, you are going to burn extra fuel to shorten the trip for humans. You can get down to about 6 months travel time before it starts being crazy to spend extra fuel to shave extra days off the trip.
I say go as soon as we have the flight hardware built and qualified.