[stdlib] Change _withUnsafeGuaranteedRef to use Builtin.convertUnownedUnsafeToGuaranteed.

[gottesmm]

This builtin (which lowers to raw SIL that doesn't use an actual builtin
instruction) allows us to access an unmanaged value at +0 with a language
guarantee rather than relying on the optimizer.

Previously, we did not do this directly since without OSSA, we were scared that
the frontend/optimizer would not be able to safely emit this code. Now that we
have ownership ssa, we are able to ensure that the frontend always copies the +0
value passed into the closure if the value +0 escapes from the closure (either
via a return, storing into memory, or by passing off as a +1 parameter to a
function).

rdar://59735604

[gottesmm]

@swift-ci test

[gottesmm]

@swift-ci benchmark

[gottesmm]

@swift-ci test source compatibility

[gottesmm]

@swift-ci test

[gottesmm]

@swift-ci benchmark

[gottesmm]

@swift-ci test source compatibility

[gottesmm]

@swift-ci test

[gottesmm]

@swift-ci benchmark

[aschwaighofer]

As a follow-up we should remove UnsafeGuaranteedPeephole.cpp

[gottesmm]

@aschwaighofer SGTM

[swift-ci]

Performance: -O

Code size: -O

Performance: -Osize

Improvement	OLD	NEW	DELTA	RATIO
PrefixWhileCountableRange	14	13	-7.1%	1.08x (?)

Code size: -Osize

Performance: -Onone

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: 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

[lorentey]

LGTM. Is there a way to add a test to catch future regressions, like the one that broke the original implementation?

[lorentey]

cc @aschwaighofer

[atrick]

I think you should explain what fakeBase is for, and why a "real" base isn't actually required in this case.

[gottesmm]

The idea is that the user is stating with this API that they are putting a _fixLifetime around as per the documentation.

[gottesmm]

@atrick I am happy to do that in a follow on commit.

[swift-ci]

Build failed
Swift Test Linux Platform
Git Sha - 7820ddc

[gottesmm]

@lorentey best way is by adding a benchmark

[gottesmm]

@swift-ci smoke test linux platform

[gottesmm]

linux failure was in sourcekit-lsp everything else passed.

[gottesmm]

@swift-ci smoke test linux platform

[gottesmm]

@lorentey I have the perfect benchmark for this actually already. I have a version of Prims that uses UnsafeUnowned directly using Unmanaged, but instead of using _wUGR, it uses takeUnretainedValue I will add a version that uses this API instead and land it before this change.

[lorentey]

@gottesmm Ah, this reminded me I still have a benchmark somewhere with three binary tree implementations (vanilla ADT with @indirect case, UnsafeMutablePointer, Unmanaged) that exercises this

[lorentey]

(But what I meant is a unit test that simply checks for unexpected retain/release calls, so that we get an immediate and clear signal. Benchmarks don't run on every merge...)

[gottesmm]

Ok. I’ll do it in a follow on commit. Clack needs this.

[gottesmm]

Or nm. Won’t let me do it on my phone.

[gottesmm]

I’m going to put in the test and retest

[gottesmm]

@lorentey The benchmark for Prims is here: #30049

[gottesmm]

I am just going to use Karoy's benchmark that just landed to show the impact here. Then I am going to merge and fix the other issues people brought up in a follow on commit.

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci smoke benchmark

[gottesmm]

@swift-ci benchmark

[gottesmm]

Finally, swift-ci does my bidding!

[gottesmm]

I'm going to merge this now. The benchmarks will show up when they do. Karoy says locally he validated the perf.

[swift-ci]

!!! Couldn't read commit file !!!

[swift-ci]

!!! Couldn't read commit file !!!