[11u] RFR: 8059241: C2: Excessive RemoveUseless passes during incremental inlining

[Hohensee, Paul]

Hi, Andrew,

We backported this patch because we noticed increased compilation overhead after we increased MaxInlineDepth to 15. We've been running with it (and the follow-on patches) in production starting with 11.0.9 (last October) with no issues. At our scale, even relatively small improvements help reduce fleet costs.

Thanks,
Paul

-----Original Message-----
From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> on behalf of Andrew Haley <aph at redhat.com>
Date: Monday, March 29, 2021 at 7:46 AM
To: Roland Westrelin <rwestrel at redhat.com>, "jdk-updates-dev at openjdk.java.net" <jdk-updates-dev at openjdk.java.net>
Subject: RE: [11u] RFR: 8059241: C2: Excessive RemoveUseless passes during incremental inlining

On 3/29/21 12:03 PM, Roland Westrelin wrote:
>
>> Okay. Will you review this, please? And please consider the likely
>> performance gain v.s churn and risk.
>
> I reviewed it. I think we would need to know what motivates the
> backport to make a judgement.

OK, thanks. Over to you, Dan.

--
Andrew Haley  (he/him)
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
https://keybase.io/andrewhaley
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671