[PATCH] D84883: [AMDGPU] prefer non-mfma in post-RA schedule

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec added a comment.

In D84883#2182909 <https://reviews.llvm.org/D84883#2182909>, @kerbowa wrote:

> Actually MFMA would need more height to ensure post-RA tries to schedule them first and together. You could have their results be implicit uses on later instruction like S_ENDPGM or something.

W/o this patch they are scheduled back to back:

S_ADD_U32
S_ADDC_U32
V_MFMA_F32_32X32X1F32
V_MFMA_F32_32X32X1F32
S_CMP_LG_U32


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D84883/new/

https://reviews.llvm.org/D84883