Youwei Wang has uploaded a new patch set (#2).
Change subject: IMPALA-2361: Using AVX2 intrinsic to accelerate the sort
operation Use AVX2 intrinsic to accelerate the sort operation
(corresponding to the "order by" clause). Only single column is supported
due to AVX2 restriction. Supported data types are
......................................................................
IMPALA-2361: Using AVX2 intrinsic to accelerate the sort operation
Use AVX2 intrinsic to accelerate the sort operation (corresponding to the
"order by" clause).
Only single column is supported due to AVX2 restriction.
Supported data types are INT32 (and any integer type whose width is less
than 32 bits), as well as FLOAT.
Overall runtime improvement is about 27% without significant memory
consumption increasing.
Change-Id: I0412fab9760a20bd29d7f8e971033063d1d63f2d
---
M be/src/runtime/sorter.cc
M be/src/util/CMakeLists.txt
A be/src/util/avx-sort-lib-test.cc
A be/src/util/avx-sort-lib.cc
A be/src/util/avx-sort-lib.h
M be/src/util/cpu-info.cc
M be/src/util/cpu-info.h
7 files changed, 3,199 insertions(+), 3 deletions(-)
git pull ssh://gerrit.cloudera.org:29418/Impala refs/changes/74/2774/2
--
To view, visit http://gerrit.cloudera.org:8080/2774
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I0412fab9760a20bd29d7f8e971033063d1d63f2d
Gerrit-PatchSet: 2
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Youwei Wang <[email protected]>
