git.osdn.net Git - android-x86/external-llvm.git/commit

(root) / android-x86 / external-llvm.git / commit

author	Matt Arsenault <Matthew.Arsenault@amd.com>
	Mon, 3 Jun 2019 18:41:34 +0000 (18:41 +0000)
committer	Matt Arsenault <Matthew.Arsenault@amd.com>
	Mon, 3 Jun 2019 18:41:34 +0000 (18:41 +0000)
commit	a69b8271d528b248e80415df7a65e4b254c86536
tree	da5f4e2fed9afbf77a826719b376b1ae933a4b75	tree \| snapshot
parent	327196792a52252fcc56fc4d57bdd411cc39904f	commit \| diff

TTI: Improve default costs for addrspacecast

For some reason multiple places need to do this, and the variant the
loop unroller and inliner use was not handling it.

Also, introduce a new wrapper to be slightly more precise, since on
AMDGPU some addrspacecasts are free, but not no-ops.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@362436 91177308-0d34-0410-b5e6-96231b3b80d8

external/llvm

RSS Atom

About OSDN

Site Announce
About OSDN
Privacy
Terms of Use
Advertise

Find Software

Search
Find by Categories
Download Ranking
Project Ranking

Develop Software

Create Project/Chamber
What's new
Recently Registered Project
Chambers Map
API

Help

Help Docs
Contact Us
Report Abuse/spam

include/llvm/CodeGen/BasicTTIImpl.h		diff \| blob \| history
include/llvm/CodeGen/TargetLowering.h		diff \| blob \| history
lib/CodeGen/CodeGenPrepare.cpp		diff \| blob \| history
lib/Target/AMDGPU/SIISelLowering.cpp		diff \| blob \| history
lib/Target/AMDGPU/SIISelLowering.h		diff \| blob \| history
test/Analysis/CostModel/AMDGPU/addrspacecast.ll		diff \| blob \| history
test/Transforms/LoopUnroll/AMDGPU/unroll-cost-addrspacecast.ll	[new file with mode: 0644]	blob