tree 969b49f35e380e41b0af7f904e62c9f6f71849a0
parent 46135ade9ec0b7fd975b5d3923aea95f4d416296
author Jeffrey Byrnes <jeffrey.byrnes@amd.com> 1744410889 -0700
committer GitHub <noreply@github.com> 1744410889 -0700
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsFcBAABCAAQBQJn+ZkJCRC1aQ7uu5UhlAAAnDwQAClTenbDPgMf5D694bwvkVJQ
 x5epE7eCPssdaGFnb94zjAwyr47kpnGpXvU4AnhBUXQENnsdGyqlK3RrniqofCmL
 MR/bNyWgVWedErP4YDhu0iRaLf/d0nUn7ur1PE0mfBGqIhEVdfEwnQ6LXeENnys5
 6kZAvODWeoHp44uM6UofYFnXLNFa5W3mVFj8VoFnRl9JfuWhD5ilB4pLf/mQgk1S
 DEK6lceusISyXrMhsr6GPuRN+Ps4JuLdMO+u3whG4cMYxbm8WTSxJcCpPcTQjqXS
 BlzGP5s+cM3GoZz48UtZJdTwlNnsr1jJek2DXWNyd8tBcCbBJOO+jTsEOjbH/9Sy
 Fb8TwsLAPxfXB6L8elccjOvg6toQ5ongDmCalXU3hgIAaae5gvG6mUCXU/C4te26
 rL9nXfkyQIOYf7sUUSxclSb0jZT9m9Lsgwk0C/zfwXp2c6CfXu2YNKHGH6GGYR7j
 AFDv5FDHROb3nZR2TVHvlpDbRjri1KW03YXMm4hJXznb+qJO+gN8IqaohTulkfrP
 wnlSMhvjSweB2dFJNmajX9Zr+fD2qtXw9FUfyDAjLr+KwexURM/aHJgXPjXKd4rB
 P/7s00i0NeW5Eqyj699MDGkgMTEcGjy80SrYpqdJcxmGem3cXvpmkUE8N8K52zNc
 PJSn42hvGs3Z3QTTuOSk
 =AIIz
 -----END PGP SIGNATURE-----
 

[AMDGPU] Teach iterative schedulers about IGLP (#134953)

This adds IGLP mutation to the iterative schedulers
(`gcn-iterative-max-occupancy-experimental`, `gcn-iterative-minreg`, and
`gcn-iterative-ilp`).

The `gcn-iterative-minreg` and `gcn-iterative-ilp` schedulers never
actually applied the mutations added, so this also has the effect of
teaching them about mutations in general. The
`gcn-iterative-max-occupancy-experimental` scheduler has calls to
`ScheduleDAGMILive::schedule()`, so, before this, mutations were applied
at this point. Now this is done during calls to `BuildDAG`, with IGLP
superseding other mutations (similar to the other schedulers). We may
end up scheduling regions multiple times, with mutations being applied
each time, so we need to track for
`AMDGPU::SchedulingPhase::PreRAReentry`