Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[CORE] Enable SortShuffleSuite with ColumnarShuffleManager #5816

Merged
merged 6 commits into from
May 29, 2024

Conversation

acvictor
Copy link
Contributor

@acvictor acvictor commented May 20, 2024

What changes were proposed in this pull request?

Enable Spark's SortShuffleSuite with shuffle manager set to ColumnarShuffleManager to ensure that shuffle is always compatible with existing Spark cases.

(Please explain how this patch was tested. E.g. unit tests, integration tests, manual tests)
Added a new test suite.

Copy link

Thanks for opening a pull request!

Could you open an issue for this pull request on Github Issues?

https://github.com/apache/incubator-gluten/issues

Then could you also rename commit message and pull request title in the following format?

[GLUTEN-${ISSUES_ID}][COMPONENT]feat/fix: ${detailed message}

See also:

Copy link

Run Gluten Clickhouse CI

@acvictor acvictor marked this pull request as ready for review May 20, 2024 12:47
Copy link

Run Gluten Clickhouse CI

@acvictor acvictor changed the title Enable SortShuffleSuite with ColumnarShuffleManager [CORE] Enable SortShuffleSuite with ColumnarShuffleManager May 20, 2024
@acvictor
Copy link
Contributor Author

@rui-mo @zhli1142015 can you please review?

@rui-mo
Copy link
Contributor

rui-mo commented May 21, 2024

Hi @zzcclp, do we need to exclude this suite for CH backend? Thanks.

@rui-mo
Copy link
Contributor

rui-mo commented May 21, 2024

@acvictor Could you rebase this PR to try the CH CI again? Thanks.

Copy link

Run Gluten Clickhouse CI

@acvictor
Copy link
Contributor Author

@rui-mo I think the Clickhouse CI has failed but I'm not sure how to access it. Should I disable the suite for CH?

@rui-mo
Copy link
Contributor

rui-mo commented May 22, 2024

@acvictor We provide an account to log into CH CI system at https://github.com/apache/incubator-gluten/blob/main/docs/get-started/ClickHouse.md#new-ci-system.

name: gluten, password: hN2xX3uQ4m

But I find the failure looks irrelevant to this change. @zzcclp Could you help confirm?

@zzcclp
Copy link
Contributor

zzcclp commented May 24, 2024

please help to exclude the GlutenSortShuffleSuite for CH backend first, thanks.

@acvictor
Copy link
Contributor Author

please help to exclude the GlutenSortShuffleSuite for CH backend first, thanks.

Done. Can you please review?

Copy link

Run Gluten Clickhouse CI

Copy link

Run Gluten Clickhouse CI

@acvictor
Copy link
Contributor Author

@zzcclp @rui-mo can you please review? CH CI is passing.

Copy link
Contributor

@rui-mo rui-mo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks. Will merge after CI passes.

@rui-mo rui-mo merged commit 588faae into apache:main May 29, 2024
40 checks passed
@GlutenPerfBot
Copy link
Contributor

===== Performance report for TPCH SF2000 with Velox backend, for reference only ====

query log/native_5816_time.csv log/native_master_05_28_2024_d7374bd8f_time.csv difference percentage
q1 35.85 34.04 -1.812 94.95%
q2 23.74 23.62 -0.111 99.53%
q3 38.90 36.93 -1.965 94.95%
q4 32.32 32.10 -0.218 99.33%
q5 67.01 69.91 2.902 104.33%
q6 7.51 5.97 -1.545 79.43%
q7 79.84 80.25 0.415 100.52%
q8 84.52 84.79 0.266 100.31%
q9 120.83 121.55 0.725 100.60%
q10 45.32 43.64 -1.678 96.30%
q11 19.84 22.27 2.435 112.27%
q12 23.83 26.90 3.075 112.91%
q13 54.56 52.40 -2.152 96.05%
q14 17.68 21.59 3.913 122.14%
q15 31.52 31.46 -0.058 99.82%
q16 14.13 13.97 -0.159 98.87%
q17 103.18 102.83 -0.354 99.66%
q18 145.11 145.92 0.812 100.56%
q19 13.52 14.82 1.297 109.59%
q20 26.96 31.06 4.099 115.20%
q21 262.22 261.69 -0.534 99.80%
q22 12.14 12.27 0.128 101.06%
total 1260.49 1269.97 9.480 100.75%

@GlutenPerfBot
Copy link
Contributor

===== Performance report for TPCDS SF2000 with Velox backend, for reference only ====

query log/native_master_05_29_2024_time.csv log/native_master_05_28_2024_d7374bd8f5_time.csv difference percentage
q1 16.97 14.54 -2.429 85.69%
q2 14.94 15.27 0.332 102.22%
q3 4.96 4.95 -0.011 99.79%
q4 64.00 63.51 -0.493 99.23%
q5 7.81 8.07 0.264 103.38%
q6 2.42 2.31 -0.111 95.41%
q7 5.64 5.15 -0.492 91.28%
q8 5.59 3.29 -2.303 58.79%
q9 18.71 16.20 -2.506 86.60%
q10 9.65 9.74 0.086 100.89%
q11 35.04 36.46 1.425 104.07%
q12 1.39 1.35 -0.040 97.12%
q13 6.52 6.67 0.155 102.38%
q14a 44.10 43.33 -0.768 98.26%
q14b 43.05 41.66 -1.385 96.78%
q15 2.47 3.69 1.215 149.12%
q16 39.09 39.54 0.453 101.16%
q17 5.64 5.39 -0.241 95.72%
q18 6.47 6.21 -0.263 95.93%
q19 1.91 3.56 1.645 186.03%
q20 1.21 1.59 0.379 131.33%
q21 2.41 1.09 -1.318 45.22%
q22 8.78 7.79 -0.989 88.74%
q23a 81.60 83.10 1.498 101.84%
q23b 102.04 102.20 0.164 100.16%
q24a 76.07 69.27 -6.810 91.05%
q24b 68.64 68.74 0.102 100.15%
q25 4.14 4.01 -0.133 96.80%
q26 3.07 4.40 1.337 143.60%
q27 3.01 2.93 -0.086 97.16%
q28 22.76 20.32 -2.436 89.29%
q29 6.60 7.82 1.225 118.57%
q30 4.28 9.08 4.794 211.95%
q31 5.97 6.79 0.821 113.75%
q32 0.96 1.23 0.275 128.76%
q33 4.84 4.80 -0.034 99.29%
q34 6.25 4.81 -1.432 77.08%
q35 6.40 6.41 0.013 100.21%
q36 2.98 3.08 0.108 103.64%
q37 3.86 3.70 -0.164 95.76%
q38 12.00 11.80 -0.200 98.34%
q39a 3.22 3.14 -0.078 97.59%
q39b 2.75 2.90 0.145 105.28%
q40 4.75 3.64 -1.103 76.76%
q41 0.58 0.58 -0.006 98.96%
q42 0.86 0.84 -0.026 97.02%
q43 3.47 3.54 0.072 102.06%
q44 6.87 10.58 3.707 153.97%
q45 3.47 3.39 -0.086 97.53%
q46 5.11 3.12 -1.992 61.03%
q47 15.72 14.59 -1.130 92.81%
q48 4.50 4.29 -0.210 95.33%
q49 7.64 7.27 -0.370 95.15%
q50 22.54 21.31 -1.229 94.55%
q51 8.32 10.20 1.876 122.55%
q52 1.01 0.97 -0.043 95.70%
q53 1.74 2.85 1.112 164.10%
q54 3.06 3.11 0.059 101.92%
q55 0.98 0.98 -0.002 99.84%
q56 4.27 4.34 0.069 101.62%
q57 8.43 8.63 0.198 102.34%
q58 2.52 2.47 -0.046 98.19%
q59 13.85 14.13 0.281 102.03%
q60 4.58 4.59 0.010 100.22%
q61 6.91 5.15 -1.758 74.55%
q62 5.46 3.70 -1.759 67.77%
q63 1.91 1.83 -0.083 95.65%
q64 48.49 50.79 2.298 104.74%
q65 15.70 15.15 -0.550 96.50%
q66 3.11 6.39 3.273 205.17%
q67 353.82 348.77 -5.051 98.57%
q68 3.62 3.54 -0.080 97.78%
q69 9.38 6.62 -2.765 70.53%
q70 8.84 8.43 -0.413 95.33%
q71 2.28 2.23 -0.053 97.68%
q72 187.72 187.81 0.096 100.05%
q73 2.29 2.21 -0.076 96.67%
q74 21.11 21.75 0.638 103.02%
q75 23.43 23.68 0.251 101.07%
q76 7.32 6.98 -0.344 95.30%
q77 1.85 1.73 -0.120 93.51%
q78 40.94 37.38 -3.562 91.30%
q79 3.38 3.49 0.113 103.34%
q80 10.89 10.51 -0.381 96.50%
q81 4.55 4.53 -0.026 99.42%
q82 6.34 6.09 -0.250 96.06%
q83 1.49 1.47 -0.017 98.84%
q84 4.82 2.47 -2.349 51.30%
q85 6.85 6.83 -0.020 99.71%
q86 3.24 3.11 -0.129 96.00%
q87 12.48 12.27 -0.210 98.31%
q88 16.41 17.80 1.385 108.44%
q89 2.61 2.60 -0.007 99.75%
q90 3.25 3.35 0.104 103.20%
q91 2.57 2.68 0.101 103.93%
q92 2.46 1.15 -1.310 46.66%
q93 30.56 28.51 -2.049 93.29%
q94 21.42 25.26 3.844 117.95%
q9 81.92 80.22 -1.698 97.93%
q5 5.38 2.56 -2.821 47.53%
q96 12.10 12.20 0.105 100.87%
q97 1.83 2.02 0.196 110.71%
q98 9.29 9.16 -0.131 98.59%
q99 9.29 9.16 -0.131 98.59%
total 1882.45 1859.70 -22.754 98.79%

@GlutenPerfBot
Copy link
Contributor

===== Performance report for TPCH SF2000 with Velox backend, for reference only ====

query log/native_master_05_29_2024_time.csv log/native_master_05_28_2024_d7374bd8f_time.csv difference percentage
q1 33.56 34.04 0.483 101.44%
q2 23.85 23.62 -0.227 99.05%
q3 37.37 36.93 -0.441 98.82%
q4 32.42 32.10 -0.317 99.02%
q5 69.83 69.91 0.078 100.11%
q6 7.54 5.97 -1.577 79.09%
q7 81.75 80.25 -1.499 98.17%
q8 86.57 84.79 -1.779 97.94%
q9 118.35 121.55 3.201 102.71%
q10 44.16 43.64 -0.516 98.83%
q11 22.23 22.27 0.041 100.19%
q12 26.72 26.90 0.181 100.68%
q13 53.60 52.40 -1.198 97.76%
q14 17.59 21.59 3.999 122.73%
q15 32.89 31.46 -1.435 95.64%
q16 13.41 13.97 0.565 104.22%
q17 103.62 102.83 -0.797 99.23%
q18 144.53 145.92 1.388 100.96%
q19 13.59 14.82 1.232 109.07%
q20 29.66 31.06 1.402 104.73%
q21 260.43 261.69 1.264 100.49%
q22 13.84 12.27 -1.574 88.63%
total 1267.50 1269.97 2.473 100.20%

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants