[PATCH] D46356: [TableGen] Emit a fatal error on inconsistencies in resource units vs cycles.

[Clement Courbet via Phabricator via llvm-commits]

courbet added inline comments.


================
Comment at: lib/Target/X86/X86ScheduleBtVer2.td:371
 defm : JWriteResYMMPair<WriteCvtPS2IY,     [JFPU1, JSTC], 3, [2,2], 2>;
-defm : JWriteResFpuPair<WriteCvtSD2I,      [JFPU1, JSTC, JFPA, JALU0], 7, [1,1], 2>;
+defm : JWriteResFpuPair<WriteCvtSD2I,      [JFPU1, JSTC, JFPA, JALU0], 7, [1,1,1,1], 4>;
 defm : JWriteResFpuPair<WriteCvtPD2I,      [JFPU1, JSTC], 3, [1,1], 1>;
----------------
RKSimon wrote:
> courbet wrote:
> > RKSimon wrote:
> > > These are 2 uops not 4, btver2 splits the issue and execution resource.
> > OK cool, thanks. So this is an NFC.
> No, you need to set it back to 2
Sorry, yes, that's what I meant; I just did not re-export to phabricator. I submitted r334178 with 2 here.


Repository:
  rL LLVM

https://reviews.llvm.org/D46356