[llvm-bugs] [Bug 36618] New: Vector decomposing of unsupported widths unsupported.

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=36618

            Bug ID: 36618
           Summary: Vector decomposing of unsupported widths unsupported.
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: GlobalISel
          Assignee: unassignedbugs at nondot.org
          Reporter: Sonicadvance1 at gmail.com
                CC: llvm-bugs at lists.llvm.org

Created attachment 20007
  --> https://bugs.llvm.org/attachment.cgi?id=20007&action=edit
Basic test case that will fail on AArch64 or x86-64 depending

GlobalISel doesn't support decomposing of unsupported vector widths in to
scalar operations.

This can be seen by attempting to do an operation on a <2 x i32> vector on
x86-64.
This can also be seen when attempting to do a <2 x float> vector operation on
AArch64, which is just an oversight in the original implementation.

I've attached a test case that can be compiled with llc to demonstrate the
issue.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20180306/40fc8026/attachment.html>