[m-rev.] for preliminary review: bug 493 & 532 alterative fix

Peter Wang novalazy at gmail.com
Tue May 18 11:29:11 AEST 2021


On Tue, 18 May 2021 02:18:14 +1000 "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> The attached diff fixes the problem, but I am not sure it is worth
> turning it on. It does eliminate the double row in state_var.c,
> but applying "stat -A" and then grep for .rodata to all the .o files
> in the compiler directory (except the two for the two source files
> touched by this change), I get 160,636 bytes before this diff
> and 160,540 bytes after, with all compiler options being the same.
> That is a small fraction of a percent improvement in space usage.
> I can't measure the impact on execution time, because I am pretty
> sure that any benchmarking will measure more the thermal throttling
> done by my laptop than the impact of this change :-(
> 
> If Peter approves the correctness of this change, I propose
> to commit the change to mlds_to_c_type.m as is, but comment out
> the change to ml_global_data.m with a comment on why. In other words,
> add this space optimization, but leave it turned off.

That's fine.

Peter


More information about the reviews mailing list