herhut accepted this revision.
herhut added a comment.
This revision is now accepted and ready to land.
Separate pass works for me.
Comment at: mlir/include/mlir/Dialect/LLVMIR/Transforms/Passes.td:19
+def NVVMOptimize : Pass<"nvvm-optimize"> {
+ let summary = "Optimize NVVM
herhut added inline comments.
Comment at: mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp:158
+// by the same divisor.
+struct ExpandDivF16 : public ConvertOpToLLVMPattern {
+ using ConvertOpToLLVMPattern::ConvertOpToLLVMPattern;
This pattern is a bit mis
herhut added a comment.
Could you add the normalization back? This is in line with the comment to make
sure the old and new files align.
Comment at: clang/include/clang/Driver/Options.td:3455
+ HelpText<"Specify target triple (e.g. i686-apple-darwin9)">,
+ MarshallingInfoStr