[PATCH] D14257: [AsmParser] Generalize matching for grammars without mnemonic-lead statements.

[Craig Topper via llvm-commits]

craig.topper added a comment.

I know this revision went in a while ago, but this seems like a really terribly performing option for Hexagon. It does a linear scan through the entire match table (which at the time of this writing is 2160 entries) for any instruction that starts with a non-token. There appear to be about ~1500 such mnemonic-less entries in table so about 3/4 of the entire table.


Repository:
  rL LLVM

http://reviews.llvm.org/D14257