Tim Armstrong has posted comments on this change.

Change subject: Use AVX2 operations to speedup Bloom filter insert by 10-50%.
......................................................................


Patch Set 2:

(3 comments)

Nice! Did you see any difference in end-to-end query time? I suspect the 
improvement may be partially masked by the interpreted hash evaluation thta's 
done at the same time (There's a "TODO: codegen expr evaluation and hashing" in 
exec/partitioned-hash-join-node-ir.cc)

http://gerrit.cloudera.org:8080/#/c/3338/2/be/src/util/bloom-filter.h
File be/src/util/bloom-filter.h:

Line 129:   // A faster SIMD version of Insert()
/// instead of //

Also combine the declaration and definition comments up here, I thnk - they're 
semi-redundant.


Line 154:     const uint32_t bucket_idx = HashUtil::Rehash32to32(hash) & 
directory_mask_;
Factor this out into InsertGeneric()


http://gerrit.cloudera.org:8080/#/c/3338/2/bin/impala-config.sh
File bin/impala-config.sh:

Line 395:   LD_LIBRARY_PATH="${IMPALA_TOOLCHAIN_GCC_LIB}:${LD_LIBRARY_PATH}"
This seems fine, but it may be worth running a private impala packaging build, 
just in case this breaks something on a particular OS.


-- 
To view, visit http://gerrit.cloudera.org:8080/3338
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I6fef4f6652876f8fd7e3f0e41431702380418c98
Gerrit-PatchSet: 2
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Jim Apple <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to