Dan Hecht has posted comments on this change. Change subject: IMPALA-4008: Don't bake ExprContext pointers into IR code ......................................................................
Patch Set 6: Code-Review+2 (4 comments) http://gerrit.cloudera.org:8080/#/c/4390/5/be/src/codegen/gen_ir_descriptions.py File be/src/codegen/gen_ir_descriptions.py: Line 48: ["AGG_NODE_GET_FN_CTX", "GetAggFnCtx"], > Not really. Even with the class name, the function name is still a substrin Yeah, but there's no good reason for those to have different method names, right? I think it'd be better to name them consistently and include the class, but since this is the old agg, I don't feel too strongly so you can leave it if you prefer. http://gerrit.cloudera.org:8080/#/c/4390/5/be/src/exec/aggregation-node.h File be/src/exec/aggregation-node.h: PS5, Line 81: avoid loading agg_fn_evaluators_[i] at runtime. > It would be slightly faster to avoid double-dereferencing (i.e. aggregate_e So you are saying we avoid loading aggregate_evaluators[i] entirely on the codegen path? if that's the case, then okay. http://gerrit.cloudera.org:8080/#/c/4390/6/be/src/exec/aggregation-node.h File be/src/exec/aggregation-node.h: PS6, Line 81: agg_fn_evaluators_ does this mean to say aggregate_evaluators_? http://gerrit.cloudera.org:8080/#/c/4390/6/be/src/exec/partitioned-aggregation-node.h File be/src/exec/partitioned-aggregation-node.h: PS6, Line 203: agg_fn_evaluators_ aggregate_evaluators_? -- To view, visit http://gerrit.cloudera.org:8080/4390 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I42039eed803a39fa716b9ed647510b6440974ae5 Gerrit-PatchSet: 6 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-HasComments: Yes