Bug 111566 - [REGRESSION] [BISECTED] Large CS workgroup sizes broken in combination with FP64 on Intel.
Summary: [REGRESSION] [BISECTED] Large CS workgroup sizes broken in combination with F...
Status: NEW
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: 19.2
Hardware: All All
: not set normal
Assignee: Jason Ekstrand
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-09-05 21:45 UTC by Francisco Jerez
Modified: 2019-09-07 01:59 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Francisco Jerez 2019-09-05 21:45:20 UTC
Large enough workgroup sizes that cause the Intel compiler back-end to generate SIMD32 code are currently broken on master in combination with some shader features that cause the back-end to allocate VGRFs larger than 16 GRFs, which aren't supported by the register allocator and other back-end compiler passes, and lead to assertion failures like:

"shader_runner: ../src/intel/compiler/brw_fs.cpp:2020: void fs_visitor::split_virtual_grfs(): Assertion `offset <= MAX_VGRF_SIZE' failed.
Aborted (core dumped)"

I've written the following Piglit test that reproduces the issue:

https://gitlab.freedesktop.org/currojerez/piglit/commit/331b8afb47b9b79e5a6f697d4fe1fc2be6702617

I bisected the regressions to change:

"commit f4ef34f207d15bcade7aed644328035dd0f2cc16
Author: Jason Ekstrand <jason@jlekstrand.net>
Date:   Wed May 29 17:46:55 2019 -0500

    intel/fs: Add an UNDEF instruction to avoid excess live ranges
    
    With 8 and 16-bit types and anything where we have to use non-trivial
    strides registersto deal with restrictions, we end up with things that
    look like partial writes even though we don't care about any values in
    the register except those written by that instruction.  This is
    particularly important when dealing with loops because liveness sees
    is_partial_write and the fact that an old version from a previous loop
    iteration may be valid at that point and extends all purely partially
    written values to the entire loop.
    
    This commit adds a new UNDEF instruction which does nothing (the
    generator doesn't emit anything) but which does a fake write to the
    register.  This informs liveness that we don't care about any values
    before that point so it won't consider those registers to be falsely
    live.  We can safely emit UNDEF instructions for all SSA values that
    come in from NIR and nearly all temporaries generated by various stages
    of the compiler.  In particular, we need to insert UNDEF instructions
    when we handle region restrictions because the newly allocated registers
    are almost guaranteed to be partially written.
    
    No shader-db changes.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110432
    Reviewed-by: Matt Turner <mattst88@gmail.com>"

I had a half-baked fix for this but it seems like Jason is in the mood of fixing SIMD32 regressions -- So here you got another one.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.