[PATCH] D12826: [PM] Wire up optimization levels and default pipeline construction APIs in the PassBuilder.
Paul Robinson via llvm-commits
llvm-commits at lists.llvm.org
Mon Sep 14 14:28:01 PDT 2015
probinson added a subscriber: probinson.
probinson added a comment.
I appreciate the direction that O1 should not ruin the debugging experience.
One suggestion, and while I was in there anyway, pointed out some typos.
================
Comment at: include/llvm/Passes/PassBuilder.h:46
@@ +45,3 @@
+
+ /// Optimize quickly without destroying debugability.
+ ///
----------------
I think "debuggability" (two g's) here and elsewhere.
================
Comment at: include/llvm/Passes/PassBuilder.h:56
@@ +55,3 @@
+ /// immediately executed as part of testing. As a consequence, where
+ /// possibly, we would like to produce efficient to execute code, but not
+ /// if it significantly slows down compilation or would prevent even basic
----------------
possibly -> possible
================
Comment at: include/llvm/Passes/PassBuilder.h:58
@@ +57,3 @@
+ /// if it significantly slows down compilation or would prevent even basic
+ /// debugging of the resulting binary.
+ O1,
----------------
Is it worth providing one or two examples of things _not_ to do at O1?
"For example, this level should avoid transforms that reorder statements, such as LICM."
================
Comment at: include/llvm/Passes/PassBuilder.h:65
@@ +64,3 @@
+ /// The key idea is that optimizations at this level should "pay for
+ /// themselves". So if an optimization reduces compile time by 5% or
+ /// increases code size by 5% for a particular benchmark, that benchmark
----------------
reduces -> increases (as in, makes it worse)
================
Comment at: include/llvm/Passes/PassBuilder.h:84
@@ +83,3 @@
+ /// this mode should include any optimization that helps execution time on
+ /// balance across a diverse collection of benchmarks, even if increase
+ /// code size or compile time for some benchmarks without corresponding
----------------
if increase -> if it increases
http://reviews.llvm.org/D12826
More information about the llvm-commits
mailing list