projects / platform / upstream / mesa.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3f50dde) | patch
intel/fs: Lower dword integer multiplies on XeHP.
authorRafael Antognolli <rafael.antognolli@intel.com>
Wed, 30 Jan 2019 00:34:30 +0000 (16:34 -0800)
committerMarge Bot <eric+marge@anholt.net>
Fri, 16 Apr 2021 08:27:35 +0000 (08:27 +0000)
commit49b2d9f428cc7a8d4010c78c4a1ae5238434f413
tree568eb22980f2cad5e7aa2150bbf97b9b67cda0a4tree | snapshot
parent3f50dde8b35a117137617a862b3f0a6c4e1d35e0commit | diff
intel/fs: Lower dword integer multiplies on XeHP.

From the BSpec:

 "When multiplying DW X DW, resulting dst can only be QW precision. If
 DW precision is required at output than MUL/MACH macro must be used."

So for now simply lower it. We might want to revisit it later.

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10000>
src/intel/compiler/brw_fs.cpp diff | blob | history
Domain: Graphics System / GL;