[Enhancement] Improve performance of strings::memcpy_inlined

[dirtysalt]

What type of PR is this：

• BugFix
• Feature
• Enhancement
• Refactor
• UT
• Doc
• Tool

Which issues of this PR fixes ：

Fixes #

Problem Summary(Required) ：

This PR is to improve memcpy to accelerate block cache:

• use overlapped move to optimize the case when size <= 16
• use avx(256bit reg) move to optimize the case when size <= 256
• use erms(enhanced repeat movsb/stosb) to optimize the case when size in [512KB,2MB]
• use unrolled avx(256bit reg) move to optimze the rest cases.

---

The overlapped move is very useful when copying irregular size like 5,7,11bytes.

• if we don't use overlapped move, 11 bytes will be copied in (movq, mov ax, mov al) (8+2+1)
• but if we use overlapped move, 11 bytes can be copied by (movq, movl) (8+4)

---

erms version works very wll when size in [512KB, 2MB], in 1 thread or 8 threads

---

avx unrolled version works well when size is large. this version works better than glibc __memcpy_avx_unaligned version when thread =1, but is worse when thread = 8. so it's really difficult to write a very general good performance memcpy for all workload.

---

	native ssb	parquet ssb	io tuned	memcpy	memcpy(io tuned)
Q01	109	173	169	179	163
Q02	91	126	128	126	122
Q03	87	118	121	122	115
Q04	578	905	773	813	739
Q05	524	1015	883	708	654
Q06	484	774	674	688	649
Q07	766	1450	1100	1214	1073
Q08	517	926	786	802	748
Q09	505	888	754	757	716
Q10	139	182	183	177	173
Q11	1000	1567	1315	1418	1263
Q12	376	590	568	539	502
Q13	301	389	388	378	362
SUM	5477	9103	7842	7921	7279

---

benchmark comparison between this impl and previous impl

-----------------------------------------------------------------
Benchmark                       Time             CPU   Iterations
-----------------------------------------------------------------
memcpy_sr/16                 2.01 ns         2.01 ns    347320268
memcpy_sr/32                 2.26 ns         2.26 ns    309545548
memcpy_sr/64                 3.66 ns         3.66 ns    191179564
memcpy_sr/128                4.61 ns         4.61 ns    151792072
memcpy_sr/256                6.84 ns         6.84 ns    102082429
memcpy_sr/512                9.67 ns         9.67 ns     72437612
memcpy_sr/1024               15.0 ns         15.0 ns     46503179
memcpy_sr/2048               29.0 ns         29.0 ns     24160369
memcpy_sr/4096               38.5 ns         38.5 ns     18135859
memcpy_sr/8192               63.9 ns         63.9 ns     10980053
memcpy_sr/16384               132 ns          132 ns      5287391
memcpy_sr/32768               781 ns          781 ns       889899
memcpy_sr/65536              1582 ns         1582 ns       450258
memcpy_sr/131072             3112 ns         3112 ns       224624
memcpy_sr/262144             6292 ns         6291 ns       111086
memcpy_sr/524288            17364 ns        17364 ns        39997
memcpy_sr/1048576           70380 ns        70378 ns        10529
memcpy_sr/2097152          256405 ns       256397 ns         2647
memcpy_sr/4194304          682078 ns       682056 ns         1021
memcpy_sr/8388608         1494222 ns      1494169 ns          469
memcpy_sr/16777216        3031167 ns      3031001 ns          231
memcpy_sr/33554432        6110517 ns      6110305 ns          115
memcpy_sr/67108864       12198513 ns     12198063 ns           58
memcpy_sr/134217728      23999964 ns     23998751 ns           29
memcpy_gutil/16              2.10 ns         2.10 ns    333638027
memcpy_gutil/32              4.68 ns         4.68 ns    149708142
memcpy_gutil/64              4.68 ns         4.68 ns    149662819
memcpy_gutil/128             4.99 ns         4.99 ns    140184350
memcpy_gutil/256             6.89 ns         6.89 ns    101552625
memcpy_gutil/512             11.9 ns         11.9 ns     58626829
memcpy_gutil/1024            21.7 ns         21.7 ns     32212684
memcpy_gutil/2048            42.0 ns         42.0 ns     16680581
memcpy_gutil/4096            84.7 ns         84.7 ns      8264949
memcpy_gutil/8192             163 ns          163 ns      4285315
memcpy_gutil/16384            361 ns          361 ns      1933720
memcpy_gutil/32768           1042 ns         1042 ns       670798
memcpy_gutil/65536           2069 ns         2069 ns       338344
memcpy_gutil/131072          4157 ns         4157 ns       165195
memcpy_gutil/262144          9016 ns         9015 ns        80531
memcpy_gutil/524288         34986 ns        34984 ns        22445
memcpy_gutil/1048576       149037 ns       149032 ns         5211
memcpy_gutil/2097152       386921 ns       386908 ns         1843
memcpy_gutil/4194304       923590 ns       923556 ns          753
memcpy_gutil/8388608      1904319 ns      1904252 ns          367
memcpy_gutil/16777216     3549960 ns      3549845 ns          196
memcpy_gutil/33554432     7838210 ns      7837956 ns           91
memcpy_gutil/67108864    15422806 ns     15422286 ns           46
memcpy_gutil/134217728   30822498 ns     30821511 ns           23

Checklist:

• I have added test cases for my bug fix or my new feature
• This pr will affect users' behaviors
• This pr needs user documentation (for new or modified features or behaviors)
  • I have added documentation for my new feature or new function

[github-actions]

clang-tidy made some suggestions

[kangkaisen]

Why not improve the memcpy by ourself?

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[dirtysalt]

@Mergifyio rebase

[mergify]

rebase

✅ Branch has been successfully rebased

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
0.7% Duplication

[dirtysalt]

run starrocks_admit_test

[wanpengfei-git]

run starrocks_admit_test