aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/Analysis/InlineCost.cpp
AgeCommit message (Expand)AuthorFilesLines
2016-08-29Fix a thinko in r278189.Easwaran Raman1-1/+1
2016-08-11Make more fields of InlineParams Optional.Easwaran Raman1-4/+8
2016-08-10Changed sign of LastCallToStaticBounsPiotr Padlewski1-1/+1
2016-08-10Do not directly use inline threshold cl options in cost analysis.Easwaran Raman1-57/+98
2016-08-05Remove cold callsite heuristic that is not necessary because of cold callee h...Dehao Chen1-7/+5
2016-08-05Replace hot-callsite based heuristic to use its own threshold parameter inste...Dehao Chen1-6/+17
2016-07-23Avoid using a raw AssumptionCacheTracker in various inliner functions.Sean Silva1-29/+29
2016-07-11Implement callsite-hotness based inline cost for Sample-based PGODehao Chen1-1/+8
2016-06-27Fix size computation of array allocation in inline cost analysisEaswaran Raman1-3/+4
2016-06-09Use ProfileSummaryInfo in inline cost analysis.Easwaran Raman1-39/+28
2016-05-19Allow -inline-threshold to override default threshold.Easwaran Raman1-4/+7
2016-05-10Revert r269131Easwaran Raman1-4/+2
2016-05-10Reapply r266477 and r266488Easwaran Raman1-2/+4
2016-05-09[Inliner] don't assume that a Constant alloca size is a ConstantInt (PR27277)Sanjay Patel1-4/+4
2016-04-28[Inliner] Formatting. NFC.Chad Rosier1-36/+41
2016-04-22Introduce llvm.load.relative intrinsic.Peter Collingbourne1-0/+5
2016-04-18Revert "Replace the use of MaxFunctionCount module flag"Eric Liu1-4/+2
2016-04-15Replace the use of MaxFunctionCount module flagEaswaran Raman1-2/+4
2016-04-15[TTI] Add getInliningThresholdMultiplier.Justin Lebar1-0/+4
2016-04-13Return immediately from analyzeCall if analyzeBlock returns false.Easwaran Raman1-14/+2
2016-04-08Refactor Threshold computation. NFC.Easwaran Raman1-22/+35
2016-04-08Don't IPO over functions that can be de-refinedSanjoy Das1-5/+6
2016-03-08Revert revisions 262636, 262643, 262679, and 262682.Easwaran Raman1-86/+16
2016-03-04Fix a memory leak.Easwaran Raman1-1/+4
2016-03-03Fix breakage caused by r262636.Easwaran Raman1-1/+1
2016-03-03Infrastructure for PGO enhancements in inlinerEaswaran Raman1-16/+83
2016-02-05CallAnalyzer::analyzeCall: change the condition back to "Cost < Threshold"Hans Wennborg1-1/+1
2016-02-01Avoid inlining call sites in unreachable-terminated blockJun Bum Lim1-6/+17
2016-01-29Annotate dump() methods with LLVM_DUMP_METHOD, addressing Richard Smith r2591...Yaron Keren1-1/+1
2016-01-28Lower inlining threshold when the caller has minsize attribute.Easwaran Raman1-8/+8
2016-01-21Change ConstantFoldInstOperands to take Instruction instead of opcode and typ...Manuel Jacob1-2/+1
2016-01-14Refactor threshold computation for inline cost analysisEaswaran Raman1-4/+106
2015-12-28Refactor inline costs analysis by removing the InlineCostAnalysis classEaswaran Raman1-36/+12
2015-12-22Provide a way to specify inliner's attribute compatibility and merging.Akira Hatanaka1-3/+1
2015-12-22Revert r256277 and r256279.Akira Hatanaka1-1/+3
2015-12-22Provide a way to specify inliner's attribute compatibility and merging.Akira Hatanaka1-3/+1
2015-12-07Use updated threshold for indirect call bonusEaswaran Raman1-2/+2
2015-12-03Test commit.Easwaran Raman1-2/+2
2015-11-13Revert r252990.Akira Hatanaka1-1/+10
2015-11-13Provide a way to specify inliner's attribute compatibility and merging.Akira Hatanaka1-10/+1
2015-11-12Revert r252949.Akira Hatanaka1-1/+10
2015-11-12Provide a way to specify inliner's attribute compatibility and mergingAkira Hatanaka1-10/+1
2015-11-10Inliner: Do zero-cost inlines even if above a negative threshold (PR24851)Hans Wennborg1-1/+1
2015-10-10Analysis: Remove implicit ilist iterator conversionsDuncan P. N. Exon Smith1-8/+7
2015-09-1580-cols; NFC80-cols; NFC80-cols; NFC80-cols; NFC80-cols; NFC80-cols; NFC80-co...Sanjay Patel1-4/+5
2015-08-23[WinEH] Require token linkage in EH pad/ret signaturesJoseph Tremoulet1-1/+1
2015-08-18[PM/AA] Remove the last relics of the separate IPA library from LLVM,Chandler Carruth1-0/+1451
2013-01-21Sink InlineCost.cpp into IPA -- it is now officially an interproceduralChandler Carruth1-1237/+0
2013-01-21Move the inline cost analysis's primary cost query to TTI instead of theChandler Carruth1-4/+4
2013-01-21Now that the inline cost analysis is a pass, we can easily have itChandler Carruth1-12/+20
generated by cgit v1.1 at 2025-10-14 22:00:29 +0000