It turns out that Litmus has a tendency to thrash almost-endlessly if given a number of threads equal to the number of logical cores in the target machine. As such, it would be useful to have a tunable cap on the number of threads generated, that is then respected by program.make.empty.
I've already been considering adding another tunable cap on the number of actions (it's hardcoded at the moment). As such, it makes sense to pool the implementation of both.
A sketch of syntax would look like this:
fuzz {
# action weights go here
set param cap.threads to 3
set param cap.actions to 20
}
This is similar to (and complements) #134 but covers integer parameters rather than Boolean parameters.
It turns out that Litmus has a tendency to thrash almost-endlessly if given a number of threads equal to the number of logical cores in the target machine. As such, it would be useful to have a tunable cap on the number of threads generated, that is then respected by
program.make.empty
.I've already been considering adding another tunable cap on the number of actions (it's hardcoded at the moment). As such, it makes sense to pool the implementation of both.
A sketch of syntax would look like this:
This is similar to (and complements) #134 but covers integer parameters rather than Boolean parameters.