[VL] Enable sort-based shuffle in micro benchmark

[marin-ma]

Removed ShuffleSplitBenchmark.cc as it's no longer maintained. We can use the GenericBenchmark.cc to test the shuffle write performance. Steps are added to the doc.

[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

[Yohahaha]

👍

[GlutenPerfBot]

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

query	log/native_5942_time.csv	log/native_master_06_16_2024_a08a57c61_time.csv	difference	percentage
q1	35.73	37.13	1.402	103.92%
q2	23.95	24.00	0.046	100.19%
q3	39.62	37.23	-2.389	93.97%
q4	32.50	32.75	0.253	100.78%
q5	68.52	71.56	3.040	104.44%
q6	8.52	8.16	-0.357	95.81%
q7	78.89	80.98	2.097	102.66%
q8	87.04	86.78	-0.263	99.70%
q9	122.55	121.57	-0.978	99.20%
q10	46.25	44.17	-2.080	95.50%
q11	20.49	21.42	0.930	104.54%
q12	22.25	25.32	3.065	113.78%
q13	40.59	39.84	-0.747	98.16%
q14	18.40	19.23	0.834	104.53%
q15	33.21	33.05	-0.160	99.52%
q16	14.16	14.54	0.382	102.70%
q17	104.62	102.23	-2.388	97.72%
q18	148.37	142.93	-5.436	96.34%
q19	14.00	15.44	1.438	110.27%
q20	30.05	27.99	-2.057	93.15%
q21	259.29	261.41	2.123	100.82%
q22	12.42	12.41	-0.006	99.95%
total	1261.40	1260.15	-1.252	99.90%