Michael Ho has posted comments on this change.

Change subject: IMPALA-3115: Hoist some variables out of loops in 
cross-compiled code.
......................................................................


Patch Set 6:

(7 comments)

http://gerrit.cloudera.org:8080/#/c/2661/6/be/src/exec/partitioned-hash-join-node-ir.cc
File be/src/exec/partitioned-hash-join-node-ir.cc:

Line 53: next_probe_row
> would it be simpler to not have this variable, but instead just use probe_b
Done


Line 183: next_probe_row;
> i.e. this could be probe_batch_iterator.Get()
Done


http://gerrit.cloudera.org:8080/#/c/2661/6/be/src/runtime/row-batch.h
File be/src/runtime/row-batch.h:

Line 164:         row_ = NULL;
> why do we have to special case the full case? can't we just require that ro
It was done so that we won't hit the DCHECK() in GetRow(). As we discussed in 
person, I will duplicate the logic of GetRow() here instead. Also added some 
DCHECKs for sanity check.


Line 170:     TupleRow* IR_ALWAYS_INLINE Next() {
> add function comments for these methods.
Done


Line 172:       DCHECK(row_ == NULL ||
> row_ can't be NULL here.
It can before as we also set num_tuples_per_row_ to 0. It's irrelevant in the 
new patch now.


Line 173:           (row_ - parent_->tuple_ptrs_) / num_tuples_per_row_ <= 
parent_->capacity_);
> Isn't this the same as DCHECK_LE(row_, row_batch_end_)?
Not really. This is testing against the capacity of the row batch which is 
useful for write iterator.


Line 398: 
> Why not have the syntax closer to BOOST_FOREACH?
Done.


-- 
To view, visit http://gerrit.cloudera.org:8080/2661
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ic7152b1fb094b3c3574d203e3774f4297f2225dc
Gerrit-PatchSet: 6
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Michael Ho <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: Michael Ho <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to