You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues-all@impala.apache.org by "Quanlong Huang (Jira)" <ji...@apache.org> on 2022/02/21 01:34:00 UTC

[jira] [Commented] (IMPALA-11037) Bump ORC to 1.7-p4 to contain the improvement of ORC-1020

    [ https://issues.apache.org/jira/browse/IMPALA-11037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17495273#comment-17495273 ] 

Quanlong Huang commented on IMPALA-11037:
-----------------------------------------

Ran perf A/B test on TPCH dataset: [https://jenkins.impala.io/job/perf-AB-test/292]
{code:java}
Report Generated on 2022-02-20
Run Description: "4077bc849ae14bb92a463aeeb6c8f5c1fca658c9 vs d467a2f96d0ca03c77aae1b30b2bcacfff20a8e1"

Cluster Name: UNKNOWN
Lab Run Info: UNKNOWN
Impala Version:          impalad version 4.1.0-SNAPSHOT RELEASE (2021-12-01)
Baseline Impala Version: impalad version 4.1.0-SNAPSHOT RELEASE (2021-12-01)

+----------+-------------------+---------+------------+------------+----------------+
| Workload | File Format       | Avg (s) | Delta(Avg) | GeoMean(s) | Delta(GeoMean) |
+----------+-------------------+---------+------------+------------+----------------+
| TPCH(42) | orc / def / block | 6.53    | -3.75%     | 4.95       | -5.15%         |
+----------+-------------------+---------+------------+------------+----------------+

+----------+----------+---------------------+--------+-------------+------------+-----------+----------------+-------+----------------+---------+---------+
| Workload | Query    | File Format         | Avg(s) | Base Avg(s) | Delta(Avg) | StdDev(%) | Base StdDev(%) | Iters | Median Diff(%) | MW Zval | Tval    |
+----------+----------+---------------------+--------+-------------+------------+-----------+----------------+-------+----------------+---------+---------+
| TPCH(42) | TPCH-Q1  | orc / block / block | 6.79   | 6.71        |   +1.06%   |   7.27%   |   0.71%        | 50    |   -0.10%       | -1.53   | 1.01    |
| TPCH(42) | TPCH-Q13 | orc / block / block | 7.48   | 7.46        |   +0.20%   |   1.10%   |   0.93%        | 50    |   +0.02%       | 0.99    | 1.00    |
| TPCH(42) | TPCH-Q21 | orc / block / block | 31.42  | 31.60       |   -0.59%   |   6.16%   |   6.44%        | 50    |   +0.12%       | 0.03    | -0.47   |
| TPCH(42) | TPCH-Q18 | orc / block / block | 12.20  | 12.26       |   -0.53%   |   1.55%   |   1.92%        | 50    |   -0.41%       | -2.01   | -1.53   |
| TPCH(42) | TPCH-Q9  | orc / block / block | 14.76  | 15.00       |   -1.60%   |   0.56%   |   0.51%        | 50    |   -1.59%       | -8.32   | -15.10  |
| TPCH(42) | TPCH-Q22 | orc / block / block | 2.36   | 2.39        |   -1.29%   |   2.19%   |   2.04%        | 50    |   -2.04%       | -3.18   | -3.06   |
| TPCH(42) | TPCH-Q10 | orc / block / block | 5.63   | 5.75        |   -2.07%   |   0.81%   |   1.10%        | 50    |   -1.86%       | -7.53   | -10.81  |
| TPCH(42) | TPCH-Q2  | orc / block / block | 1.60   | 1.64        |   -2.73%   |   3.55%   |   2.72%        | 50    |   -3.13%       | -3.66   | -4.39   |
| TPCH(42) | TPCH-Q3  | orc / block / block | 6.45   | 6.66        |   -3.25%   |   0.83%   |   0.72%        | 50    |   -3.23%       | -8.48   | -21.22  |
| TPCH(42) | TPCH-Q16 | orc / block / block | 2.35   | 2.44        |   -3.57%   |   1.75%   |   1.76%        | 50    |   -4.22%       | -7.00   | -10.36  |
| TPCH(42) | TPCH-Q19 | orc / block / block | 4.80   | 5.03        |   -4.49%   |   0.60%   |   0.46%        | 50    |   -4.60%       | -8.53   | -43.03  |
| TPCH(42) | TPCH-Q5  | orc / block / block | 5.00   | 5.30        | I -5.56%   |   0.95%   |   1.06%        | 50    | I -5.89%       | -8.53   | -28.37  |
| TPCH(42) | TPCH-Q14 | orc / block / block | 4.47   | 4.73        | I -5.68%   |   1.22%   |   1.05%        | 50    | I -6.25%       | -8.52   | -25.80  |
| TPCH(42) | TPCH-Q6  | orc / block / block | 2.46   | 2.61        | I -5.91%   |   0.97%   |   0.87%        | 50    | I -6.13%       | -8.53   | -33.10  |
| TPCH(42) | TPCH-Q7  | orc / block / block | 5.66   | 6.07        | I -6.80%   |   1.24%   |   0.77%        | 50    | I -7.18%       | -8.53   | -34.51  |
| TPCH(42) | TPCH-Q11 | orc / block / block | 1.20   | 1.29        | I -7.03%   |   2.22%   |   1.98%        | 50    | I -8.31%       | -8.50   | -17.35  |
| TPCH(42) | TPCH-Q8  | orc / block / block | 6.07   | 6.55        | I -7.42%   |   1.16%   |   1.02%        | 50    | I -8.26%       | -8.53   | -35.49  |
| TPCH(42) | TPCH-Q4  | orc / block / block | 4.03   | 4.37        | I -7.91%   |   0.93%   |   0.90%        | 50    | I -8.77%       | -8.53   | -44.83  |
| TPCH(42) | TPCH-Q17 | orc / block / block | 5.60   | 6.16        | I -9.17%   |   4.54%   |   5.88%        | 50    | I -9.95%       | -7.16   | -9.02   |
| TPCH(42) | TPCH-Q15 | orc / block / block | 5.63   | 6.33        | I -11.04%  |   0.51%   |   0.43%        | 50    | I -12.54%      | -8.53   | -124.63 |
| TPCH(42) | TPCH-Q12 | orc / block / block | 3.86   | 4.43        | I -12.89%  |   1.39%   |   1.24%        | 50    | I -14.48%      | -8.53   | -52.56  |
| TPCH(42) | TPCH-Q20 | orc / block / block | 3.85   | 4.43        | I -13.02%  |   0.79%   |   0.86%        | 50    | I -14.58%      | -8.53   | -83.66  |
+----------+----------+---------------------+--------+-------------+------------+-----------+----------------+-------+----------------+---------+---------+
{code}
Attached the report details ( [^perf-report-ORC-1020-tpch-42.txt] ). CC [~drorke], [~rizaon]

> Bump ORC to 1.7-p4 to contain the improvement of ORC-1020
> ---------------------------------------------------------
>
>                 Key: IMPALA-11037
>                 URL: https://issues.apache.org/jira/browse/IMPALA-11037
>             Project: IMPALA
>          Issue Type: Improvement
>          Components: Backend
>            Reporter: Quanlong Huang
>            Assignee: Quanlong Huang
>            Priority: Major
>             Fix For: Impala 4.1.0
>
>         Attachments: orc_1.7.0-p3_random_int32.svg, orc_1.7.0-p3_random_int64.svg, orc_1.7.0-p4_random_int32.svg, orc_1.7.0-p4_random_int64.svg, perf-report-ORC-1020-tpch-42.txt
>
>
> ORC-1020 improves read performance of the ORC library in scanning random integers. Columns that encoded into integers, e.g. dictionary encoded strings, will also benifit from this.
> This Jira aims to add ORC-1020 to our native-toolchain and bump our orc version to 1.7-p4 to contain it.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscribe@impala.apache.org
For additional commands, e-mail: issues-all-help@impala.apache.org