Skip to content

intel/compiler: Run extra fp64 lowering pass on devices that don't support int64.

Francisco Jerez requested to merge currojerez/mesa:mtl-int64-lower_doubles-pass into main

In some cases nir_lower_int64 will emit fp64 operations which aren't natively supported on any Intel hardware (e.g. ftrunc, frem). An extra pass of nir_opt_algebraic (for frem) and nir_lower_doubles is required in order to take care of them. This fixes several int64 test-cases on MTL hardware.

Merge request reports