RFR: 8271118: C2: StressGCM should have higher priority than frequency-based policy

[Yi Yang]

On Thu, 22 Jul 2021 05:34:40 GMT, Yi Yang <yyang at openjdk.org> wrote:

> I think StressGCM should have higher priority than frequency-based policy, otherwise StressGCM has no chance to choose because the frequency-based policy would choose block first if appropriately.

Hi,

> Seems reasonable. Were you working on some concrete issue to discover this?

This idea is originated from our offline discussion about GCM that we are recently interested in..

> If your goal is to update random number then I agree with @kelthuzadx suggestion to call it before check.

Yes. The goal of this change is: if we turn on StressGCM, then we hope to do stress testing on C2 GCM and choose randomly instead of using conventional strategies such as freqneucy-based policy and latency-based policy. Therefore, we should check StressGCM in advance to give him an opportunity to make the choice.

Besides, I think we can break the loop(i.e. from earliest block to LCA block) once we have chosen a random block. It is reasonable?

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

PR: https://git.openjdk.java.net/jdk/pull/4865