http://llvm.org/bugs/show_bug.cgi?id=15866
Bug ID: 15866
Summary: SROA builds needlessly complex shufflevectors
Product: libraries
Version: trunk
Hardware: PC
OS: All
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Classification: Unclassified
Created attachment 10439
--> http://llvm.org/bugs/attachment.cgi?id=10439&action=edit
fubarness inside!
For the attached input, SROA creates a series of vector shuffles that just end
up recreating the original vector before storing it. No other optimization
cleans it up and codegen doesn't either, so we end up with horrible scalarized
vector operations.
In the attached testcase, using the inline function causes the above nastiness.
Interestingly, just replacing the call with the actual stores results in code
that works fine. See if "#if 0/#else" in the testcase.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
LLVMbugs mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
