[Mesa-dev] nir/algebraic: Generalize an optimization - RADV / DXVK 0.93+0.94 regression with LS2019

Dieter Nützel Dieter at nuetzel-hh.de
Thu Dec 20 03:14:41 UTC 2018


Replying myself:

FIXED with Ian's

nir/algebraic: Don't put quotes around floating point literals
commit #96c4b135e34d0804e41bfbc28fc1b5050c49d71e

Thanks!

Dieter

Am 19.12.2018 06:36, schrieb Dieter Nützel:
> Sorry guys,
> 
> but my son got a RADV + DXVK 0.93 and 0.94 (D3D11) regression running
> brand new LS2019
> (https://www.farming-simulator.com/?lang=en&country=us) on our Polaris
> 20. All heaven textures are gone and some others are wrong.
> 
> git bisect bad 445867c80d
> git bisect good e4f9a37ace
> 
> /opt/mesa> git bisect good
> 615cc26b97ad520b90a8d3b3f9bdaa49c78dfda5 is the first bad commit
> commit 615cc26b97ad520b90a8d3b3f9bdaa49c78dfda5
> Author: Jason Ekstrand <jason.ekstrand at intel.com>
> Date:   Thu Dec 6 12:56:33 2018 -0600
> 
>     nir/algebraic: Generalize an optimization
> 
>     This just makes it nicely scale across bit sizes.
> 
>     Reviewed-by: Eric Anholt <eric at anholt.net>
>     Reviewed-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
>     Tested-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
> 
> :040000 040000 ced4299ecdd6d76755839c9a0af11417e5f4973a
> 321e0c0e1c985c49c4e0b37f766c4a68c4ca145f M src
> 
> But git revert 615cc26b97 alone do NOT help.
> 
> Have to file a ticket after some sleep...;-)
> 
> Greetings,
> Dieter
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev


More information about the mesa-dev mailing list