[AutoDiff] Add differentiation benchmarks

[rxwei]

Revives #31108 by @marcrasi.

[rxwei]

@swift-ci benchmark

[atrick]

lgtm

[rxwei]

@swift-ci benchmark

[rxwei]

@swift-ci benchmark

[eeckstein]

lgtm, though it seems that the benchmarks are crashing during run.

[rxwei]

Interesting... I'm unable to reproduce this on macOS.

[rxwei]

@swift-ci benchmark

[rxwei]

@swift-ci benchmark

[rxwei]

@edymtt any concerns with 3e86dc9?

[rxwei]

@swift-ci please test

[swift-ci]

Performance: -O

Improvement	OLD	NEW	DELTA	RATIO
CharacterLiteralsLarge	97	87	-10.3%	1.11x (?)
UTF8Decode_InitFromBytes_ascii_as_ascii	452	407	-10.0%	1.11x (?)
UTF8Decode_InitFromData_ascii_as_ascii	644	581	-9.8%	1.11x (?)
CharacterLiteralsSmall	305	284	-6.9%	1.07x (?)
Added	MIN	MAX	MEAN	MAX_RSS
DifferentiationArraySum	1799	2178	1927	—
DifferentiationIdentity	304	306	305	—
DifferentiationSquare	590	591	591	—

Code size: -O

Regression	OLD	NEW	DELTA	RATIO
TestsUtils.o	28270	28810	+1.9%	0.98x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
Breadcrumbs.MutatedIdxToUTF16.ASCII	3	4	+33.3%	0.75x
NSStringConversion.UTF8	652	770	+18.1%	0.85x (?)
Array2D	6272	6752	+7.7%	0.93x (?)
ObjectiveCBridgeFromNSArrayAnyObjectForced	4220	4540	+7.6%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
ProtocolDispatch	333	307	-7.8%	1.08x (?)
Added	MIN	MAX	MEAN	MAX_RSS
DifferentiationArraySum	1681	1862	1743	—
DifferentiationIdentity	311	312	312	—
DifferentiationSquare	600	600	600	—

Code size: -Osize

Regression	OLD	NEW	DELTA	RATIO
TestsUtils.o	22409	23197	+3.5%	0.97x

Performance: -Onone

Added	MIN	MAX	MEAN	MAX_RSS
DifferentiationArraySum	1673	1887	1745	—
DifferentiationIdentity	380	380	380	—
DifferentiationSquare	703	704	703	—

Code size: -swiftlibs

✅	Benchmark Check Report
⚠️⏱	DifferentiationArraySum execution took at least 1653 μs. Decrease the workload of DifferentiationArraySum by a factor of 2 (10), to be less than 1000 μs.

How to read the data

The tables contain differences in performance which are larger than 8% and differences in code size which are larger than 1%.

If you see any unexpected regressions, you should consider fixing the
regressions before you merge the PR.

Noise: Sometimes the performance results (not code size!) contain false
alarms. Unexpected regressions which are marked with '(?)' are probably noise.
If you see regressions which you cannot explain you can try to run the
benchmarks again. If regressions still show up, please consult with the
performance team (@eeckstein).

Hardware Overview

  Model Name: Mac Pro
  Model Identifier: MacPro6,1
  Processor Name: 8-Core Intel Xeon E5
  Processor Speed: 3 GHz
  Number of Processors: 1
  Total Number of Cores: 8
  L2 Cache (per Core): 256 KB
  L3 Cache: 25 MB
  Memory: 64 GB