[VL] Daily Update Velox Version (2024_06_30)

[GlutenPerfBot]

Upstream Velox's New Commits:

0ef0ac8e4 by Jia Ke, Enable right join in smj (10148)
c54e59dbb by wypb, Fix HashStringAllocator::clear() and cumulativeBytes_ (10260)
4963d7116 by duanmeng, Add recursive spill for RowNumber (8654)
e3de4ea9d by Sandino Flores, Add support for Protobuf v22+ (10294)
0d8022846 by PHILO-HE, Support finding installed arrow libraries from system (9992)
fd955bff4 by liangyongyuan, Add float/double types support for Spark mod function (9848)
0ced9e5f0 by NEUpanning, Fix typo in expression evaluation documentation (10304)
8803bfbd1 by lingbin, Fix typo in SIMD document (10319)
bcfc8f8c3 by PHILO-HE, Allow returning Status from callNullable and callNullFree methods (10274)
258db516d by PHILO-HE, Use legacySizeOfNull argument to determine the behavior of Spark size function (10100)

[github-actions]

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:

• Other pull requests

[philo-he]

@marin-ma, we still need to apply some changes for arrow build (if arrow lib is not found, velox will resolve the dependency by
building from source as before). Its cmake file is moved here:
https://github.com/facebookincubator/velox/tree/main/CMake/resolve_dependency_modules/arrow

[philo-he]

ARROW_HOME needs to be set to

"${VELOX_HOME}/_build/release/CMake/resolve_dependency_modules/arrow/arrow_ep/" (replace release with debug for debug mode).

https://github.com/apache/incubator-gluten/blob/main/cpp/CMakeLists.txt#L71-L73

[GlutenPerfBot]

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

query	log/native_master_07_01_2024_time.csv	log/native_master_06_30_2024_86449d0c8_time.csv	difference	percentage
q1	19.66	14.83	-4.822	75.47%
q2	13.98	14.85	0.872	106.24%
q3	4.21	4.21	0.007	100.16%
q4	63.74	63.51	-0.229	99.64%
q5	7.86	7.38	-0.481	93.88%
q6	2.39	2.32	-0.079	96.70%
q7	5.24	6.03	0.791	115.08%
q8	3.70	5.14	1.448	139.16%
q9	22.95	21.68	-1.265	94.49%
q10	11.68	11.58	-0.097	99.17%
q11	35.50	36.42	0.927	102.61%
q12	2.44	2.59	0.153	106.26%
q13	5.53	9.04	3.517	163.63%
q14a	44.64	41.82	-2.822	93.68%
q14b	40.59	38.81	-1.773	95.63%
q15	3.93	3.62	-0.314	92.01%
q16	38.53	39.65	1.118	102.90%
q17	7.25	5.07	-2.185	69.88%
q18	6.66	6.20	-0.459	93.11%
q19	2.08	2.34	0.258	112.40%
q20	1.49	1.40	-0.089	94.05%
q21	1.06	1.02	-0.041	96.15%
q22	7.67	7.79	0.119	101.56%
q23a	84.05	84.02	-0.026	99.97%
q23b	100.62	101.13	0.513	100.51%
q24a	69.49	69.39	-0.103	99.85%
q24b	68.46	72.20	3.747	105.47%
q25	8.18	4.35	-3.835	53.14%
q26	4.43	2.85	-1.573	64.46%
q27	2.86	3.15	0.286	109.98%
q28	20.85	28.74	7.892	137.86%
q29	9.59	7.11	-2.482	74.12%
q30	5.23	4.19	-1.039	80.13%
q31	6.16	6.70	0.549	108.91%
q32	1.14	2.13	0.989	186.97%
q33	4.80	4.74	-0.063	98.68%
q34	3.66	3.65	-0.010	99.73%
q35	6.73	6.57	-0.163	97.58%
q36	3.29	3.35	0.068	102.06%
q37	4.02	4.05	0.027	100.68%
q38	11.92	11.85	-0.068	99.43%
q39a	3.33	5.69	2.361	170.99%
q39b	2.86	3.27	0.418	114.63%
q40	8.52	3.73	-4.793	43.77%
q41	0.64	0.62	-0.021	96.71%
q42	0.96	0.94	-0.019	98.05%
q43	5.34	3.89	-1.447	72.89%
q44	8.81	8.62	-0.190	97.84%
q45	3.54	3.40	-0.137	96.12%
q46	3.40	3.22	-0.178	94.76%
q47	14.29	14.40	0.108	100.75%
q48	4.59	4.58	-0.006	99.87%
q49	10.03	9.80	-0.233	97.68%
q50	23.46	24.23	0.776	103.31%
q51	8.63	8.51	-0.116	98.66%
q52	1.03	1.04	0.008	100.74%
q53	1.99	2.01	0.029	101.45%
q54	3.37	3.23	-0.138	95.90%
q55	1.05	0.99	-0.062	94.08%
q56	5.48	4.53	-0.950	82.67%
q57	8.61	9.28	0.673	107.82%
q58	2.71	2.59	-0.118	95.63%
q59	17.09	16.43	-0.663	96.12%
q60	4.95	4.83	-0.122	97.53%
q61	7.46	5.67	-1.794	75.95%
q62	4.34	4.76	0.416	109.58%
q63	2.24	2.32	0.080	103.57%
q64	48.99	49.99	1.001	102.04%
q65	13.85	13.70	-0.141	98.98%
q66	3.43	4.25	0.817	123.81%
q67	355.58	355.08	-0.502	99.86%
q68	3.79	3.65	-0.140	96.32%
q69	6.55	6.25	-0.295	95.49%
q70	8.73	8.60	-0.127	98.54%
q71	4.20	2.45	-1.748	58.38%
q72	189.48	188.05	-1.430	99.25%
q73	2.37	2.26	-0.111	95.31%
q74	21.60	21.26	-0.336	98.44%
q75	23.29	23.20	-0.097	99.59%
q76	12.12	9.42	-2.697	77.75%
q77	2.10	2.21	0.109	105.20%
q78	39.33	41.97	2.638	106.71%
q79	3.45	3.51	0.056	101.64%
q80	11.00	11.16	0.154	101.40%
q81	5.29	5.19	-0.100	98.11%
q82	7.99	10.14	2.157	127.01%
q83	1.48	1.55	0.077	105.24%
q84	2.99	2.68	-0.309	89.65%
q85	6.89	6.54	-0.354	94.86%
q86	3.17	3.21	0.042	101.34%
q87	13.43	12.09	-1.340	90.03%
q88	24.68	24.76	0.076	100.31%
q89	7.29	3.25	-4.039	44.58%
q90	4.32	9.55	5.230	221.11%
q91	2.68	5.39	2.714	201.44%
q92	1.41	1.40	-0.011	99.25%
q93	28.70	28.51	-0.191	99.33%
q94	21.24	21.60	0.359	101.69%
q9	85.76	83.91	-1.853	97.84%
q5	3.47	3.78	0.318	109.18%
q96	12.28	12.34	0.060	100.49%
q97	2.08	1.96	-0.129	93.80%
q98	9.15	9.18	0.024	100.26%
q99	9.15	9.18	0.024	100.26%
total	1921.10	1914.13	-6.975	99.64%

[FelixYBW]

@philo-he just to confirm, with this PR, how many times Arrow will be built and where it's built?

[FelixYBW]

@marin-ma why these operators test should be ignored?

[philo-he]

@FelixYBW, let me help clarify. This is a known test failure since we need some code change in Gluten. I have created a Gluten pr to do the code change and re-enable these tests: #6014.

[GlutenPerfBot]

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

query	log/native_master_07_01_2024_time.csv	log/native_master_06_30_2024_86449d0c8_time.csv	difference	percentage
q1	34.57	35.24	0.667	101.93%
q2	23.98	23.80	-0.185	99.23%
q3	39.13	40.97	1.838	104.70%
q4	32.81	32.74	-0.069	99.79%
q5	70.64	70.02	-0.618	99.12%
q6	8.16	7.93	-0.236	97.11%
q7	83.92	82.49	-1.427	98.30%
q8	85.80	83.59	-2.204	97.43%
q9	120.92	118.63	-2.295	98.10%
q10	46.13	49.82	3.684	107.99%
q11	21.88	20.68	-1.198	94.53%
q12	25.17	25.51	0.349	101.39%
q13	41.24	38.43	-2.802	93.20%
q14	20.05	17.87	-2.188	89.09%
q15	33.15	30.75	-2.394	92.78%
q16	13.35	14.20	0.845	106.33%
q17	104.45	104.33	-0.121	99.88%
q18	148.04	148.98	0.933	100.63%
q19	13.81	14.66	0.850	106.16%
q20	27.79	29.28	1.490	105.36%
q21	266.41	261.77	-4.637	98.26%
q22	13.75	12.54	-1.202	91.26%
total	1275.14	1264.22	-10.919	99.14%

[philo-he]

@philo-he just to confirm, with this PR, how many times Arrow will be built and where it's built?

@FelixYBW, currently, there are still two builds, since a build issue for Velox was found yesterday. We need wait for this velox pr: facebookincubator/velox#10355

[FelixYBW]

Can we add this PR to OAP/Velox temporarily?

In future if a PR from us to Velox may impact Gluten, let's submit the PR to Gluten also and make sure once Velox's PR is merged Gluten can work as expected.