[m-rev.] for review: let simplification introduce ground_term_const cons_ids
zoltan.somogyi at runbox.com
Tue Jun 1 12:36:48 AEST 2021
2021-06-01 12:09 GMT+10:00 "Peter Wang" <novalazy at gmail.com>:
> That looks okay.
Thanks for that. I followed all your suggestions.
My benchmarking was inconclusive. The effect of this diff
on the size of the rodata section was minimal, and any effects
on the speed of the compiler on speedtest were smaller
than the effects of random thermal throttling on my laptop :-(
The main reason why I think this diff is nevertheless desirable
is that it should help us to eventually phase out the use of
the other ways the compiler generates static data.
More information about the reviews