[PATCH] D36849: [AMDGPU] Port of HSAIL inliner

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: lib/Target/AMDGPU/AMDGPUInline.cpp:30-32
+static cl::opt<int>
+HintThreshold("amdgpu-inlinehint-threshold", cl::Hidden, cl::init(12000),
+              cl::desc("Threshold for inlining functions with inline hint"));
----------------
rampitec wrote:
> arsenm wrote:
> > We should avoid having a custom version of the exact same flag that the regular inliner uses. This would result in some surprises.
> Stock inline hint threshold is only 44% higher than regular threshold. Here I have it 6 times higher. The exact number may and will change, but when I return 9 from getInliningThresholdMultiplier() and multiply hint by the very same number it still will be a very low number, far less than we would really like to have.
If that is really a problem we should find a different solution. I really don't want to have the set of flags for debugging this changing.


https://reviews.llvm.org/D36849