Skip to content

Commit 40d66bd

Browse files
committed
Follow Benoit D. comments
1 parent 06ea5a1 commit 40d66bd

File tree

2 files changed

+3
-3
lines changed

2 files changed

+3
-3
lines changed

doc/contributor/updating-ruby-installers.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -46,7 +46,7 @@ script/update-truffleruby-graalvm $VERSION ../ruby-versions/pkg
4646

4747
### Create a Release
4848

49-
Once pushed/merged to the main branch, it's a good idea to create a `ruby-build` release,
49+
Once pushed/merged to the default branch, it's a good idea to create a `ruby-build` release,
5050
so `ruby-build` Homebrew users can get the new TruffleRuby release too:
5151

5252
```bash

doc/user/reporting-performance-problems.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -84,9 +84,9 @@ That seems like a lot - and what is actually happening here is that TruffleRuby
8484
away your benchmark.
8585
The effect is less pronounced with complex code that cannot be optimized away.
8686

87-
### Final Technical Note: Gaps and Value Profiling
87+
### Final Technical Note: Blackholes and Value Profiling
8888

89-
Some benchmarking tools for other languages have features called "gaps."
89+
Some benchmarking tools for other languages have features called "blackholes."
9090
These surround a value and make it appear to be a variable at runtime, even if it is in fact a constant, so that the optimizer does not remove it and actually performs any computations that use it.
9191
However, TruffleRuby uses extensive value profiling (caching of values and turning them into constants), so that even if you make a value appear to be a variable at its source, it is likely to be value-profiled at an intermediate stage.
9292
In general, more complex benchmarks that naturally defeat value profiling are preferable, rather than manually adding annotations to turn off important features.

0 commit comments

Comments
 (0)