[SR-5875] GSB performance regression

[davezarzycki]

Previous ID	SR-5875
Radar	None
Original Reporter	@davezarzycki
Type	Bug
Status	Resolved
Resolution	Done

Environment

My smoke test debug build script:

#!/bin/sh
 
time ./utils/build-script \
--build-subdir="$(dirname $PWD)/d" \
--llvm-targets-to-build X86 \
--skip-ios --skip-tvos --skip-watchos \
--skip-build-benchmarks true \
--build-swift-static-stdlib false \
--build-swift-static-sdk-overlay false \
--build-swift-dynamic-sdk-overlay false \
--build-swift-stdlib-unittest-extra false \
-r \
--debug-swift \
--debug-swift-stdlib \
"$@"
 

Additional Detail from JIRA

Votes	0
Component/s	Compiler
Labels	Bug
Assignee	@DougGregor
Priority	Medium

md5: 3014dcefe990cf7a1732b00bd3ac9e96

Issue Description:

The stdlib takes about 2x longer to build after pull request #11769 and #11751 were merged. (From about 5+ minutes to 10.5 minutes on a MacBook Pro.)

[DougGregor]

Yeah, working on it. Lots of work to do for performance of the GenericSignatureBuilder. Thanks for the bug!

[davezarzycki]

I'm about to file a bug that PR #11751 also caused about a 40% regression too, thus causing a multiplicative 2x slowdown. Want two bugs, or just one?

[DougGregor]

Just the one is fine.

[davezarzycki]

I think we can close this as fixed, right?

[DougGregor]

Through a long series of improvements to the compiler and standard library, we've closed this performance gap.