[llvm-bugs] [Bug 40129] New: [AMDGPU][Regression, bisected] Recent Load/Store optimizer improvements break SI

via llvm-bugs llvm-bugs at lists.llvm.org
Fri Dec 21 04:56:42 PST 2018


            Bug ID: 40129
           Summary: [AMDGPU][Regression,bisected] Recent Load/Store
                    optimizer improvements break SI
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: release blocker
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: samuel.pitoiset at gmail.com
                CC: htmldeveloper at gmail.com, llvm-bugs at lists.llvm.org


The following commit introduces a bunch of Vulkan CTS failures (~1600) on SI.
Tested with RADV.

If you want to reproduce the problem, you can run
dEQP-VK.clipping.user_defined.clip_cull_distance.vert_tess_geom.1_7 or
dEQP-VK.image.image_size.1d.readonly_1 for example.

Note that a bunch of vec3 CTS tests fail too, maybe it's an alignment issue
that is SI-specific?

4778d2ba9306039e888628874cf5a6c32e6bc004 is the first bad commit
commit 4778d2ba9306039e888628874cf5a6c32e6bc004
Author: Neil Henning <neil.henning at amd.com>
Date:   Wed Dec 12 16:15:21 2018 +0000

    [AMDGPU] Extend the SI Load/Store optimizer to combine more things.

    I've extended the load/store optimizer to be able to produce dwordx3
    loads and stores, This change allows many more load/stores to be combined,
    and results in much more optimal code for our hardware.

    Differential Revision: https://reviews.llvm.org/D54042

    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@348937

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/20181221/8aa22268/attachment.html>

More information about the llvm-bugs mailing list