[VL] [Test] Test rebasing upstream velox 11_8

[JkSelf]

What changes were proposed in this pull request?

In the upstream Velox, it is ensured that the sorting key is unique in TopNode here. Therefore, we add a check for this in this PR.

How was this patch tested?

Pass CI.

[github-actions]

Thanks for opening a pull request!

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

https://github.com/oap-project/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

[JkSelf]

/Benchmark Velox

[GlutenPerfBot]

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

query	log/native_3642_time.csv	log/native_master_11_07_2023_e3eff1d8f_time.csv	difference	percentage
q1	34.53	34.38	-0.149	99.57%
q2	24.72	25.03	0.309	101.25%
q3	38.61	38.14	-0.464	98.80%
q4	37.26	37.57	0.312	100.84%
q5	69.90	71.50	1.599	102.29%
q6	6.39	6.26	-0.126	98.03%
q7	86.78	82.22	-4.553	94.75%
q8	90.09	86.95	-3.143	96.51%
q9	121.62	119.81	-1.814	98.51%
q10	53.04	51.26	-1.780	96.64%
q11	20.53	19.73	-0.796	96.12%
q12	25.89	24.39	-1.498	94.21%
q13	49.19	50.30	1.104	102.24%
q14	17.91	17.67	-0.242	98.65%
q15	31.57	30.35	-1.212	96.16%
q16	16.25	16.20	-0.053	99.67%
q17	101.45	101.51	0.059	100.06%
q18	148.22	148.26	0.038	100.03%
q19	16.59	16.17	-0.420	97.47%
q20	29.62	30.31	0.688	102.32%
q21	226.25	224.88	-1.365	99.40%
q22	13.44	14.08	0.646	104.80%
total	1259.84	1246.98	-12.863	98.98%

[rui-mo]

How many tets are affected by Velox's change? If there are many of them, maybe we should consider removing duplicated keys in Gluten before sending to Velox.

[JkSelf]

How many tets are affected by Velox's change? If there are many of them, maybe we should consider removing duplicated keys in Gluten before sending to Velox.

@rui-mo Only Q72 in TPC-DS is affected. Maybe we can remove the duplicated keys in the follow PRs.

[rui-mo]

How does its fallback affect the performance of q72?

[zhouyuan]

👍

[JkSelf]

How does its fallback affect the performance of q72?

@rui-mo There is no performance impact.

[GlutenPerfBot]

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

query	log/native_3642_time.csv	log/native_master_11_08_2023_29b58997a_time.csv	difference	percentage
q1	32.99	34.11	1.124	103.41%
q2	22.98	25.13	2.150	109.36%
q3	38.64	38.62	-0.023	99.94%
q4	37.28	36.07	-1.204	96.77%
q5	71.86	71.84	-0.013	99.98%
q6	6.36	7.59	1.234	119.41%
q7	84.69	85.82	1.130	101.33%
q8	87.29	86.97	-0.323	99.63%
q9	120.32	121.86	1.543	101.28%
q10	54.33	54.80	0.476	100.88%
q11	19.88	20.59	0.707	103.56%
q12	24.75	24.63	-0.116	99.53%
q13	48.39	49.95	1.565	103.23%
q14	18.15	16.99	-1.153	93.65%
q15	30.88	30.97	0.093	100.30%
q16	16.22	16.51	0.289	101.78%
q17	103.34	102.40	-0.944	99.09%
q18	148.73	148.28	-0.457	99.69%
q19	15.27	15.13	-0.131	99.14%
q20	29.65	29.97	0.325	101.10%
q21	226.71	222.20	-4.508	98.01%
q22	13.06	13.55	0.489	103.74%
total	1251.75	1254.01	2.252	100.18%