[m-rev.] for review: optimization_options.m
zoltan.somogyi at runbox.com
Tue Sep 29 10:31:27 AEST 2020
2020-09-28 17:06 GMT+10:00 "Zoltan Somogyi" <zoltan.somogyi at runbox.com>:
>>> + ;
>>> + OptOption = oo_procs_per_c_function(N),
>>> + OldN = !.OptTuple ^ ot_procs_per_c_function,
>>> + !OptTuple ^ ot_procs_per_c_function := int.max(OldN, N)
>> > I think all the other integer values set by -O<n> increase with n,
>> except for this one.
>> > -O6 implies everything_in_one_c_function by internally setting
>> ot_procs_per_c_function to 0.
>> If the user passes "--procs-per-c-function 5 -O6" then
>> ot_procs_per_c_function will be set to 5 instead of 0.
>> This is a VERY minor issue.
> I think I will fix this by making "everything in one function" a separate
> boolean option, with the value of the existing procs_per_c_function
> option meaningful only if this boolean does not override it. However,
> I will do this in a separate change.
The attached committed diff does this.
More information about the reviews