On Wed, 20 May 2026 20:27:03 GMT, jengebr <[email protected]> wrote:

> ## Opportunity
> 
> JDK-836829 describes runtime overhead due to global megamorphism, which can 
> be overcome by considering only local *morphism. This change eliminates the 
> overhead from a sometimes ultra-hot method, Writer.append(CharSequence).
> 
> ## Solution
> The core change is moving from String.valueOf(Object) (with global profiling) 
> to a method-local call to CharSequence.toString() (with local profiling and a 
> clear optimization target, String). The included benchmark shows an 
> improvement around 60% when passing a String (table below, benchmark in PR).
> 
> Should be backportable.
> 
> | input | pre-fix ns | post-fix ns | delta (ns) | delta (%) |
> |---|---|---|---|---|
> | string | 4.611 | 1.724 | −2.89 | −62.6% |
> 
> 
> ---------
> - [x] I confirm that I make this contribution in accordance with the [OpenJDK 
> Interim AI Policy](https://openjdk.org/legal/ai).

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/jdk/pull/31224

Reply via email to