Hello Michael Ho, Alex Behm, Dan Hecht,
I'd like you to reexamine a change. Please visit
to look at the new patch set (#22).
Change subject: IMPALA-3567 Part 2, IMPALA-3899: factor out PHJ builder
IMPALA-3567 Part 2, IMPALA-3899: factor out PHJ builder
The main outcome of this patch is to split out PhjBuilder from
PartitionedHashJoinNode, which manages the build partitions for the
join and implements the DataSink interface.
A lot of this work is fairly mechanical refactoring: dividing the state
between the two classes and duplicating it where appropriate. One major
change is that probe partitions need to be separated from build
This required some significant algorithmic changes to memory management
for probe partition buffers: memory management for probe partitions was
entangled with the build partitions: small buffers were allocated for
the probe partitions even for non-spilling joins and the join could
spill additional partitions during probing if the probe partitions needed
to switch from small to I/O buffers. The changes made were:
- Probe partitions are only initialized after the build is partitioned, and
only for spilled build partitions.
- Probe partitions never use small buffers: once the initial write
buffer is allocated, appending to the probe partition never fails.
- All probe partition allocation is done after partitioning the build
and before processing the probe input during the same phase as hash
table building. (Aside from NAAJ partitions which are allocated
The probe partition changes necessitated a change in
BufferedTupleStream: allocation of write blocks is now explicit via the
Ran exhaustive build and local stress test.
Ran stress test binary search locally for TPC-DS SF-1 and TPC-H SF-20.
No regressions on TPC-DS. TPC-H either stayed the same or improved in
min memory requirement without spilling, but the min memory requirement
with spilling regressed in some cases. I investigated each of the
significant regressions on TPC-H and determined that they were all due
to exec nodes racing for spillable or non-spillable memory. None of them
were cases where exec nodes got their minimum reservation and failed to
execute the spilling algorithm correctly.
27 files changed, 2,264 insertions(+), 1,490 deletions(-)
git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/73/3873/22
To view, visit http://gerrit.cloudera.org:8080/3873
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Owner: Tim Armstrong <tarmstr...@cloudera.com>
Gerrit-Reviewer: Alex Behm <alex.b...@cloudera.com>
Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com>
Gerrit-Reviewer: Michael Ho
Gerrit-Reviewer: Michael Ho <k...@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com>