Closed morrisonlevi closed 1 week ago
Benchmark execution time: 2024-06-19 04:38:06
Comparing candidate commit db79e24c26af9a07b6f2f8465153963345a5ac6d in PR branch levi/no-execute-internal-crash
with baseline commit e731d2e06aed929735b0400768780175ab882852 in branch levi/randomized-tests-php8.3
.
Found 3 performance improvements and 0 performance regressions! Performance is the same for 24 metrics, 9 unstable metrics.
execution_time
[-60.036ms; -32.544ms] or [-6.243%; -3.384%]cpu_user_time
[-128.127ms; -42.189ms] or [-8.600%; -2.832%]execution_time
[-83.839ms; -47.077ms] or [-8.202%; -4.606%]Seems like... no. Something else in the patch must be messing it up.
Description
In another PR I seemed to be getting crashes in the sequence of events:
dd_uhook_end
on an internal functionddog_php_prof_interrupt_function
Profiler ::collect_time
And then I OOM, sometimes in Zend Memory Manager but sometimes in Rust.
I'm trying to isolate this from the other changes to see if there's just an issue here somewhow.
Reviewer checklist