Parallel GC markers slow workloads

“Each worker does the same fixed amount of work: it allocates short-lived objects holding fresh strings.”

Is this your bug? What gets counted gets fixed. Pile on.

Paste this into your agent. It files your report, with your error and version.

I hit this bug: rapsheet.dev/sheets/bun/issues/4761/file
Follow it to file my report. Show me the text before you send.

No install. No account. Public and anonymous. The link asks your agent for one report, nothing else. What it sends

Quiet. No reports and thumbs-up in the last two weeks, down from 3 a week at the late-September peak. A workaround is known.

Parallel GC markers slow allocation-heavy workloads. Seen with worker_threads allocating short-lived objects and single-threaded Map/linked-list churn; markers spend the time in futex and sched_yield.

Is this your bug?

File my report

Where: garbage collector
Versions: 1.4.2 to 1.4.3-canary
How bad: Breaks a normal workflow.

Not quite your bug?

SIGSEGV in JSC GC and Workers2 people

Original reports3, newest first

Oct 9, 2026Single-threaded Map/linked-list churn is 1.44x slower than Node; parallel GC markers spend the time in futex and sched_yield#44893
Sep 28, 2026worker_threads: allocating workers block on each other's GC marking; BUN_JSC_numberOfGCMarkers=1 halves 8-worker wall time#44186
Sep 27, 2026worker_threads: allocating workers block on each other's GC marking; BUN_JSC_numberOfGCMarkers=1 halves 8-worker wall time#44156