As I said above, with setting exactly as crutschow, I got the same result. However, I was wondering why run time (4fs, 5m) affects the rise or fall time. That seems strange to me.

My guess is that the schematic capture tool is overriding your choices based on some set of rules that it has. Try manually creating (or editing) the circuit files.

 

My guess is that the schematic capture tool is overriding your choices based on some set of rules that it has. Try manually creating (or editing) the circuit files.

I looked at the LTspice netlist and the default values of 0 rise and fall times gives actual rise and fall times of about 10% of the pulse width when simulated, so the override is being done somewhere in the Spice code when it processes the netlist.
You can make the rise and fall times arbitrarily small but not 0 (as would also be true in a real circuit).

 

There is also the issue of the Window Size parameter (settable in Tools/Control Panel/Compression). The default value is only 300. You can increase that to improve the graphical resolution: but you'd need an awful lot of points to resolve 1fS in a 1mS sim .