Dan Hecht has posted comments on this change. (
http://gerrit.cloudera.org:8080/8025 )
Change subject: IMPALA-5844: use a MemPool for expr result allocations
......................................................................
Patch Set 11:
(12 comments)
This is looking very good. Todo for me:
partitioned-aggregation-node.cc
sorter.{h,cc}
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/exec-node.h
File be/src/exec/exec-node.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/exec-node.h@379
PS11, Line 379: Status QueryMaintenance(RuntimeState* state)
WARN_UNUSED_RESULT;
I still think we should get rid of this but okay to leave in this change.
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/hash-table.h
File be/src/exec/hash-table.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/hash-table.h@396
PS11, Line 396: Cloes
Close
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/hash-table.h@525
PS11, Line 525: /// results of expr evaluation.
Not owned? Which also means the exec node will do the maintenance on them,
right?
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h
File be/src/exec/partitioned-aggregation-node.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h@214
PS11, Line 214: if this is a non-grouping aggregation used
: /// for the non-grouping cases
did that get garbled?
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h@219
PS11, Line 219: /// are allocated from 'expr_perm_pool_' and
'expr_results_pool_' respectively.
and what about for grouping case?
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h@419
PS11, Line 419: expr_managed
should this be talking about permanent?
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h@420
PS11, Line 420: per partition
I don't get that comment, given this is a single pool.
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/partitioned-aggregation-node.h@421
PS11, Line 421: Local allocations
results
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/scanner-context.h
File be/src/exec/scanner-context.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/exec/scanner-context.h@337
PS11, Line 337: /// evaluation to prevent memory from accumulating.
who owns this?
and why does this make sense in ScannerContext? (reading the class comment it's
not clear why a lot of this stuff is here).
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/udf/udf.h
File be/src/udf/udf.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/udf/udf.h@350
PS11, Line 350: For allocations that back the intermediate results of
aggregation
: /// returned by Init(), Update() and Merge() should be
allocated by
a bit garbled
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/udf/udf.h@616
PS11, Line 616: temporary
this is kind of new terminology. above i think you said "temporary results".
maybe that, or just "results"?
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/util/tuple-row-compare.h
File be/src/util/tuple-row-compare.h:
http://gerrit.cloudera.org:8080/#/c/8025/11/be/src/util/tuple-row-compare.h@77
PS11, Line 77: Any
: /// the evaluators with use
garbled
--
To view, visit http://gerrit.cloudera.org:8080/8025
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I4ba5a7542ed90a49a4b5586c040b5985a7d45b61
Gerrit-Change-Number: 8025
Gerrit-PatchSet: 11
Gerrit-Owner: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: Michael Ho <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Comment-Date: Tue, 03 Oct 2017 20:26:10 +0000
Gerrit-HasComments: Yes
