On Wed, Dec 16, 2015 at 11:57 AM, Nanley Chery <nanleych...@gmail.com> wrote:
> From: Nanley Chery <nanley.g.ch...@intel.com>
>
> Commit 64880d073ab21ae1abad0c049ea2d6a1169a3cfa consolidated two
> DIV_ROUND_UP() definitions to one, but chose the more
> compute-intensive version in the process. Use the simpler version
> instead. Reduces .text size by 1360 bytes.
>
> Output of `size lib/i965_dri.so`:
> text data bss dec hex filename
> 7850440 219264 27240 8096944 7b8cb0 lib/i965_dri.so (before)
> 7849080 219264 27240 8095584 7b8760 lib/i965_dri.so (after)
>
> Cc: Axel Davy <axel.d...@ens.fr>
> Signed-off-by: Nanley Chery <nanley.g.ch...@intel.com>
> ---
> src/util/macros.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/util/macros.h b/src/util/macros.h
> index 0c8958f..53a98a0 100644
> --- a/src/util/macros.h
> +++ b/src/util/macros.h
> @@ -211,6 +211,6 @@ do { \
> #endif
>
> /** Compute ceiling of integer quotient of A divided by B. */
> -#define DIV_ROUND_UP( A, B ) ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )
> +#define DIV_ROUND_UP(A, B) (((A) + (B) - 1) / (B))
This *is* the definition the Linux kernel uses, so I think it's
probably the right thing to do.
As a data point (and not necessarily an objection) this actually
increases .text size for me:
text data bss dec hex filename
5314834 221988 27912 5564734 54e93e lib/i965_dri.so (before)
5320490 221988 27912 5570390 54ff56 lib/i965_dri.so (after)
by about 5.5k. I've configured with -O2 -march=native -pipe on a
Haswell with gcc 4.9.3. I'm curious what the discrepancy is.
Reviewed-by: Matt Turner <matts...@gmail.com>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
