Use globalization invariant mode + prefer speed for AOT compilers

[MichalStrehovsky]

Last time I checked OptimizationPreference=Speed is a free 3% throughput improvement. (And 6% size regression but we're not too sensitive to that.)
Globalization invariant mode removes ICU initialization and that also improves perf. Also makes size smaller.

Cc @dotnet/ilc-contrib

[ghost]

Tagging subscribers to this area: @agocke, @MichalStrehovsky, @jkotas
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Last time I checked OptimizationPreference=Speed is a free 3% throughput improvement. (And 6% size regression but we're not too sensitive to that.)
Globalization invariant mode removes ICU initialization and that also improves perf. Also makes size smaller.

Cc @dotnet/ilc-contrib

Author:	MichalStrehovsky
Assignees:	-
Labels:	area-NativeAOT-coreclr
Milestone:	-

[MichalStrehovsky]

/azp run runtime-extra-platforms

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[EgorBo]

For crossgen it should make sense to enable DOTNET_TC_CallCountingDelayMs=0 - we've seen up to 20% improvement with it (but tiered compilation has to be enabled too)

[MichalStrehovsky]

For crossgen it should make sense to enable DOTNET_TC_CallCountingDelayMs=0 - we've seen up to 20% improvement with it (but tiered compilation has to be enabled too)

That can be a separate change. I don't know how to set that. We only run crossgen on top of a jitted runtime if NativeAOT is not supported, so it's basically just Windows x86 and Linux ARM32. We get like a 60% improvement from NativeAOT otherwise :).