Application failure with traversal

[Lennart Börjeson]

> 9 feb. 2018 kl. 17:13 skrev Aleksey Shipilev <shade at redhat.com>:
> 
> On 02/09/2018 05:07 PM, Lennart Börjeson wrote:
>> I'd like some advice on how I should collect some useful info for you. The test involves a
>> multi-server setup, so I cannot easily provide any submittable testcase.
> 
> Must be a compiler bug then.
> 
> See our Wiki under "Functional diagnostics"
> 
> Try to dissect if a particular compiler fails?
>  a) interpreter: -Xint

Runs OK.

>  b) C1: -XX:TieredStopAtLevel=1

Runs OK.

>  c) C2: -XX:-TieredCompilation

Fails.

> 
>> Bound=20000 -XX:-UseBiasedLocking -XX:+DoEscapeAnalysis -XX:+UseNUMA -XX:+UseShenandoahMatrix
>> -XX:+UnlockExperimentalVMOptions -XX:ShenandoahGCHeuristics=traversal -DRoundRobinPrio=0
> 
> Drop UseShenandoahMatrix, it is not compatible with "traversal". It is automatically enabled with
> "partial", and is not actually user-settable.

Done in all test.

> 
> -Aleksey
> 


/Lennart