[llvm-bugs] [Bug 31964] New: [test][builtins] __divsc3 (and more) tests fail depending on -march

[via llvm-bugs]

http://bugs.llvm.org/show_bug.cgi?id=31964

            Bug ID: 31964
           Summary: [test][builtins] __divsc3 (and more) tests fail
                    depending on -march
           Product: compiler-rt
           Version: unspecified
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: compiler-rt
          Assignee: unassignedbugs at nondot.org
          Reporter: mgorny at gentoo.org
                CC: llvm-bugs at lists.llvm.org

I attempted to address this in D28862 [1] but I don't really have the time atm
to figure out the best solution. Long story short, some tests may fail randomly
depending on -march and a few more parameters due to inexact results.

Citing the example from D28862:

> For example, the value of (0.000001+j0.000001) / (-0.500000-j2.000000))
> computed the canonical way without specific machine flags is:
>
> z = -0x1.3bce70p-21 + j0x1.7af7bcp-22
>
> However, if -march=i386 -mfpmath=387 is used, it becomes:
>
> z = -0x1.3bce72p-21 + j0x1.7af7bcp-22

I don't really know if we should allow inexact result here (and how inexact
exactly) or if the generated code should be considered buggy.

[1]:https://reviews.llvm.org/D28862

-- 
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/20170215/eb1295e3/attachment-0001.html>