Marcel Kornacker has posted comments on this change. Change subject: IMPALA-4397,IMPALA-3259: reduce codegen time and memory ......................................................................
Patch Set 11: (9 comments) http://gerrit.cloudera.org:8080/#/c/4956/11/be/src/codegen/llvm-codegen.cc File be/src/codegen/llvm-codegen.cc: Line 939: if (is_corrupt_) { did the formatter do this? Line 988: bytes_allocated), odd indentation Line 1044: estimated_memory), odd indentation http://gerrit.cloudera.org:8080/#/c/4956/11/be/src/codegen/llvm-codegen.h File be/src/codegen/llvm-codegen.h: Line 127: // augment with brief comment describing memory tracking. Line 467: /// expressions), avoid inlining the individual expression evaluation into the could we make it "avoid inlining for the exprs exceeding this threshold"? Line 595: void DestroyIRModule(); DestroyIrModule although above we simply refer to it as 'module', so maybe simply DestroyModule()? http://gerrit.cloudera.org:8080/#/c/4956/11/be/src/exec/hash-join-node.cc File be/src/exec/hash-join-node.cc: Line 140: false, std::logical_or<bool>()); odd indentation http://gerrit.cloudera.org:8080/#/c/4956/11/be/src/exec/hdfs-scan-node-base.cc File be/src/exec/hdfs-scan-node-base.cc: Line 302: if (IsNodeCodegenDisabled()) return; i find this counterintuitive - why call codegen() if it's disabled (ie, i'd expect those lines to be reversed). http://gerrit.cloudera.org:8080/#/c/4956/11/fe/src/main/java/org/apache/impala/planner/DistributedPlanner.java File fe/src/main/java/org/apache/impala/planner/DistributedPlanner.java: Line 824: mergeFragment.getPlanRoot().setDisableCodegen(true); what is this for? does exchange ever do codegen? -- To view, visit http://gerrit.cloudera.org:8080/4956 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id10015b49da182cb181a653ac8464b4a18b71091 Gerrit-PatchSet: 11 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Reviewer: Marcel Kornacker <mar...@cloudera.com> Gerrit-Reviewer: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Silvius Rus <s...@cloudera.com> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-HasComments: Yes
