Optimize spill placement in builtins

sethbrenith · V8 LUCI CQ · commit 08c87ab24ddf · 2023-11-03T14:27:35.000Z
This change enables spill placement optimization for all values in builtins, rather than only for loop-top phi values. There is no correctness reason that this optimization is limited to loop-top phi values; I just chose that heuristic because I didn't have much evidence of benefits outside loops. See [1] for an overview of how spill placement optimization works. This change is beneficial not only because it saves a store instruction along some control-flow paths, but also because it gives the frame elider more freedom to avoid creating stack frames along those paths. I suspect that frame elision is more impactful for builtins than for JIT-compiled code, because builtins tend to be small functions called often, rather than large functions with loops. In Edge, I see about 0.3% improvement on Speedometer with this change. [1] https://docs.google.com/document/d/1n9ADWnDI-sw0OvdSmrthf61prmDqbDmQq-NSrQw2MVI/edit?usp=sharing Change-Id: Ibea4d63dda522cd8b1b95bbf4c67e10a1b3186e9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/5002146 Commit-Queue: Seth Brenith <seth.brenith@microsoft.com> Reviewed-by: Thibaud Michaud <thibaudm@chromium.org> Cr-Commit-Position: refs/heads/main@{#90729}
diff --git a/BUILD.gn b/BUILD.gn
@@ -2347,6 +2347,7 @@ template("run_mksnapshot") {
                                         "root_out_dir") + "/mksnapshot",
                          root_build_dir),
       "--turbo_instruction_scheduling",
+      "--stress-turbo-late-spilling",
 
       # In cross builds, the snapshot may be generated for both the host and
       # target toolchains.  The same host binary is used to generate both, so
