[stdlib] Switch Dictionary.subscript(_:default:) to use _modify

[airspeedswift]

Getting rid of another addressor.

[airspeedswift]

@swift-ci please smoke benchmark

[swift-ci]

Build comment file:

Performance: -O

TEST	OLD	NEW	DELTA	SPEEDUP
Regression
DictionarySubscriptDefaultMutation	256	294	+14.8%	0.87x
Improvement
DictionarySubscriptDefaultMutationArray	603	557	-7.6%	1.08x

Code size: -O

TEST	OLD	NEW	DELTA	SPEEDUP
Regression
DictionarySubscriptDefault.o	23894	24549	+2.7%	0.97x

Performance: -Osize

TEST	OLD	NEW	DELTA	SPEEDUP
Regression
DictionaryKeysContainsNative	30	34	+13.3%	0.88x
ObjectiveCBridgeFromNSArrayAnyObjectForced	4756	5244	+10.3%	0.91x (?)
Improvement
DictionarySubscriptDefaultMutationArray	659	593	-10.0%	1.11x

Code size: -Osize

TEST	OLD	NEW	DELTA	SPEEDUP
Regression
DictionarySubscriptDefault.o	22366	22813	+2.0%	0.98x

Performance: -Onone

TEST	OLD	NEW	DELTA	SPEEDUP
Improvement
DictionaryKeysContainsNative	61	52	-14.8%	1.17x

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: 12-Core Intel Xeon E5
  Processor Speed: 2.7 GHz
  Number of Processors: 1
  Total Number of Cores: 12
  L2 Cache (per Core): 256 KB
  L3 Cache: 30 MB
  Memory: 64 GB

---

[airspeedswift]

@swift-ci please smoke test

[airspeedswift]

@eeckstein @rjmccall any ideas about the regression in DictionarySubscriptDefaultMutation?

[lorentey]

👍

[eeckstein]

I'll take a look

[palimondo]

@eeckstein, @lorentey I was searching through PRs with benchmarks to see if a regression was caught in DictionarySubscriptDefaultMutation* family of benchmarks. I'm seeing 0.5x slowdown between 2018-09-18 and 2018-10-02 (those were my local build times, I'm not sure if they were synced to the tree at that time) on both O and Onone builds. Have you seen that?

edit: Is this what's being addressed by #19664?

[lorentey]

#19664 is an unrelated low-level optimizer issue. We worked around the defaulted subscript regression in #19663.