[Mesa-dev] Reducing GLSL compiler optimization time

[Ian Romanick]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/11/2011 04:06 PM, Eric Anholt wrote:
> With the conversion of texenvprogram to producing GLSL IR, all those
> nice little optimization passes end up slowing down our
> combinatorial-explosion ff tests pretty significantly.  This patch
> series reduces glean texCombine runtime overall by 20% on my gen6
> system.  The lowest-hanging fruit at this point looks like the
> register allocator setup, which I think a 1-entry cache would cut down
> pretty easily.
> 
> Surprisingly enough, making the refcount visitor use a hash table
> didn't pay off in reducing its couple of percent, as the cost of
> alloc/free of the hash table was too high.  This branch is present as
> glsl-refcount-ht of my tree if you might have a workload where it
> helps -- it was a non-statistically-significant change here.

1, 2, 4 (lol), 6, 7 look good as-is.  Reviewed-by: Ian Romanick
<ian.d.romanick at intel.com>

I posted a reply to 3 that may make 3 and 5 irrelevant.  We may decid to
commit them anyway.  Dunno.

I shared Ken's issues with 8 and 9, but those appear to be resolved.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAk1+id4ACgkQX1gOwKyEAw+fWQCgh92ZK1dZut0VGF6dDy3CfYYV
b/8An3jAxco1AHA0JXUseaD61c/hhZ77
=7s76
-----END PGP SIGNATURE-----