impala-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <ger...@cloudera.org>
Subject [Impala-CR](cdh5-trunk) IMPALA-3629: Codegen TransferScratchTuples() in hdfs-parquet-scanner
Date Thu, 28 Jul 2016 23:02:34 GMT
Tim Armstrong has posted comments on this change.

Change subject: IMPALA-3629: Codegen TransferScratchTuples() in hdfs-parquet-scanner
......................................................................


Patch Set 3:

(2 comments)

http://gerrit.cloudera.org:8080/#/c/3774/3/be/src/exec/hdfs-parquet-scanner-ir.cc
File be/src/exec/hdfs-parquet-scanner-ir.cc:

PS3, Line 44:   if (tuple_size == 0) {
> Given that if (tuple_size == 0) { ...} is handling an entire scratch patch 
My thought was that tuple_size is a constant most of the function should be pruned out immediately
and the remainder should be cheap to compile and optimise (no branches, not much code).

Handling tuple_size == 0 outside of codegen'd code might give us a gain though so could be
worth doing if it's simple.


PS3, Line 61: while (scratch_tuple != scratch_tuple_end) {
> Can we factor the while loop (or the loop body only) out as a separate func
I thought about that but I'm not sure there's much gain once we prune out the branch above
with our num_tuples value - we need most of the variables computed above for this loop.


-- 
To view, visit http://gerrit.cloudera.org:8080/3774
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ic327e437c7cd2b3f92cdb11c1e907bfee2d44ee8
Gerrit-PatchSet: 3
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Thomas Tauber-Marshall <tmarshall@cloudera.com>
Gerrit-Reviewer: Michael Ho <kwho@cloudera.com>
Gerrit-Reviewer: Thomas Tauber-Marshall <tmarshall@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-HasComments: Yes

Mime
View raw message