on 10-20-2013 12:00 PM
Dear HANA Guru!
In this lesson Hasso explain scan performance principles.
https://openhpi.de/videos/348?module_item_id=847&type=flash
How HANA perform column (in column storage table) scan?
In parallel (many CPU cores on each column) or max one core on each column?
How HANA perform column (in column storage table) scan?
In parallel (many CPU cores on each column) or max one core on each column?
The exact details of the column store scan operation are not publicly documented and certainly are part of the core SAP HANA algorithms (aka crown jewels). So don't expect to find anything more detailed than: SAP HANA uses SIMD/AltiVec processor functionality to access the internal data structures for columns (e.g. 'main vector').
Concerning the parallel degree the answer is: it depends on the operation performed, In many cases it would be one core works on a single column at any given time. Things like table partitioning can of course change that.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello Michael,
may be for you can be helpful info at
Week_4,
Script "parallel_data_processing (1)" (a script, not slides! from a download section),
Page 7,
Fig. 17.6: Single Instruction Multiple Data Parallelism:
on the right side:
Parallel Add
(Low Level Parallel Column Access)
//----
//----
I hope this answers your question.
P.S.
Good luck with achievement tomorrow, tёzka 😃
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
84 | |
24 | |
11 | |
9 | |
7 | |
6 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.