Youwei Wang has posted comments on this change.

Change subject: IMPALA-2809: Improve ByteSwap with builtin function or SSE or 
AVX2.
......................................................................


Patch Set 11:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/3081/11/be/src/benchmarks/bswap-benchmark.cc
File be/src/benchmarks/bswap-benchmark.cc:

Line 49: //                           SIMD               57.73               
2.18X
> I am surprised by this. Is a single branch making this code 20x-40x slower?
Hi Jim. I suspect the following code:
if (CpuInfo::IsSupported(CpuInfo::AVX2)) {

I guess since there is a function in the conditional branch, it may be the 
reason why such much performance is consumed here. I will step in deeper to 
investigate this.


Line 222:         MACRO_TEST_FBS_CASE(8, 5)
> remove the switch; instead:
Done


Line 260: /// FIXED_LEN_SIZE = 4: Decimal4Value, size of array element is 8x4 = 
32bit
> You've explain what FIXED_LEN_SIZE is, but I still don't know why we would 
Hi Jim. Previous codebase is written by Zuo Wang. I believe his idea is to 
illustrate the performance difference for different FIXED_LEN_SIZE here so to 
produce comprehensive and persuasive results. As for me, my idea is pretty 
simple and straightforward as you may have seen in the bit-util.inline.h file: 
I always try to apply the swap approach which can deal with the most bits when 
possible, not this enumerating-style approach. I am not sure which one is 
preferable for a benchmark case.


-- 
To view, visit http://gerrit.cloudera.org:8080/3081
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I392ed5a8d5683f30f161282c228c1aedd7b648c1
Gerrit-PatchSet: 11
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Youwei Wang <[email protected]>
Gerrit-Reviewer: Alex Behm <[email protected]>
Gerrit-Reviewer: Jim Apple <[email protected]>
Gerrit-Reviewer: Marcel Kornacker <[email protected]>
Gerrit-Reviewer: Mostafa Mokhtar <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Youwei Wang <[email protected]>
Gerrit-HasComments: Yes

Reply via email to