Tim Armstrong has posted comments on this change. Change subject: IMPALA-3066: Lazy materialization of LLVM module bitcode. ......................................................................
Patch Set 1: (3 comments) http://gerrit.cloudera.org:8080/#/c/3220/1/be/src/codegen/llvm-codegen.cc File be/src/codegen/llvm-codegen.cc: Line 76: DEFINE_bool(disable_lazy_materialization, false, > Sure, I think we can remove this flag. Just thought it would be safe to add I spent a little time trying to think of corner cases that we might be missing coverage of, but couldn't think of anything. I think we cover a good combination of exprs and plan shapes. Were there any corner cases you were concerned about? Line 97: static StringRef llvm_module_ir; > Will it be better if they are class static ? One way to look at them is tha It's not that bad but the local variable just seems simpler to me when there's no real perf concern. Line 565: if (called_fn != NULL) MaterializeFunctionHelper(called_fn); > Not really. We do check if the function has been materialized already above Oh, so isMaterializable() is false if the function is already materialized? That's confusing. -- To view, visit http://gerrit.cloudera.org:8080/3220 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I6ed7862fc5e86005ecea83fa2ceb489e737d66b2 Gerrit-PatchSet: 1 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Michael Ho <[email protected]> Gerrit-Reviewer: Michael Ho <[email protected]> Gerrit-Reviewer: Mostafa Mokhtar <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
