Use AccessPath in LICM and split loads for combine load/store hoisting

[atrick]

Still waiting to merge #33121
And creating more thorough test cases for this.

Fixes regressions caused by rdar://66791257 (Print statement provokes "Can't unsafeBitCast between types of different sizes" when optimizations enabled)

Step 1:
Use AccessPath in LICM.

The LICM algorithm was not robust with respect to address projection
because it identifies a projected address by its SILValue. This should
never be done! Use AccessPath instead.

Step 2:
LICM: split loads that are wider than the loop-stored value.

For combined load-store hoisting, split loads that contain the
loop-stored value into a single load from the same address as the
loop-stores, and a set of loads disjoint from the loop-stores. The
single load will be hoisted while sinking the stores to the same
address. The disjoint loads will be hoisted normally in a subsequent
iteration on the same loop.

loop:
  load %outer
  store %inner1
exit:

Will be split into

loop:
  load %inner1
  load %inner2
  store %inner1
exit:

Then, combined load/store hoisting will produce:

  load %inner1
loop:
  load %inner2
exit:
  store %inner1

[eeckstein]

First step for my review: I'd like to fully understand how this should work. So I just reviewed the documentation for now. Once I have a clear picture, I continue with the implementation.

[atrick]

@eeckstein documentation feedback was all very helpful. The rewritten documentation is here:

https://github.com/apple/swift/blob/d5590801e44b51f96c371dc8d81e619687ae71b0/docs/SILProgrammersManual.md#accessedstorage-and-accesspath

Keep in mind, I'm keeping it as concise as I can since I think details belong in file or class level comments.

This PR is only supposed to cover the LICM load splitting optimization.

[shahmishal]

Please update the base branch to main by Oct 5th otherwise the pull request will be closed automatically.

• How to change the base branch: (Link)
• More detail about the branch update: (Link)

[atrick]

@swift-ci test

[atrick]

@swift-ci test source compatibility

[atrick]

@swift-ci benchmark

[swift-ci]

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
StringAdder	224	241	+7.6%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
Data.hash.Medium	31	27	-12.9%	1.15x (?)

Code size: -O

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
FlattenListLoop	930	1048	+12.7%	0.89x (?)
Improvement	OLD	NEW	DELTA	RATIO
Chars2	3700	3300	-10.8%	1.12x (?)
Calculator	166	149	-10.2%	1.11x (?)
Data.hash.Medium	31	28	-9.7%	1.11x
ObjectiveCBridgeStubToNSStringRef	83	75	-9.6%	1.11x (?)

Code size: -Osize

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
ArrayAppendAsciiSubstring	28656	31176	+8.8%	0.92x (?)
DataToStringSmall	3600	3900	+8.3%	0.92x (?)
ArrayAppendLatin1Substring	29304	31716	+8.2%	0.92x (?)
Improvement	OLD	NEW	DELTA	RATIO
Data.hash.Medium	40	37	-7.5%	1.08x (?)
LuhnAlgoEager	2525	2349	-7.0%	1.07x (?)

Code size: -swiftlibs

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 mini
  Model Identifier: Macmini8,1
  Processor Name: 6-Core Intel Core i7
  Processor Speed: 3.2 GHz
  Number of Processors: 1
  Total Number of Cores: 6
  L2 Cache (per Core): 256 KB
  L3 Cache: 12 MB
  Memory: 64 GB

[swift-ci]

Build failed
Swift Test Linux Platform
Git Sha - c37402993f3b6ede773ac016fdb865b2ef041648

[atrick]

I finished writing tests for corner cases and caught a potential miscompile.

[atrick]

@swift-ci test

[atrick]

@swift-ci test source compatibility

[swift-ci]

Build failed
Swift Test OS X Platform
Git Sha - 437765e

[atrick]

I'm seeing a PR testing issue:
Failed test:
lldb-api lang/swift/unknown_self/TestSwiftUnknownSelf.py

libc++abi.dylib: terminating with uncaught exception of type std::__1::system_error: recursive_mutex lock failed: Invalid argument

[atrick]

And another PR testing failure with the Debug SCK

<unknown>:0: error: fatal error encountered during compilation; please submit a bug report (https://swift.org/contributing/#reporting-bugs) and include the project
15:01:47 <unknown>:0: note: Cannot read '/Users/buildnode/jenkins/workspace-private/swift-PR-source-compat-suite-debug/build/compat_macos/swift-macosx-x86_64/./lib/swift/macosx/layouts-x86_64.yaml'

[atrick]

@swift-ci smoke test OS X

[compnerd]

This seems to have failed to build on Windows, can we revert this?