projects / platform / upstream / VK-GL-CTS.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 393912c) | patch
Reduce expected precision of RGB10A2 filtering
authorGraeme Leese <gleese@broadcom.com>
Tue, 9 Apr 2019 11:25:19 +0000 (12:25 +0100)
committerAlexander Galazin <Alexander.Galazin@arm.com>
Sun, 28 Apr 2019 08:10:58 +0000 (04:10 -0400)
commit145785bb98da85c120e3a850ee0a7d02095ae648
tree803f0c4c1841bb671a3df9438f0a0a05f9aba7b0tree | snapshot
parent393912cd7758e7b05e82cd0e6ce337e9d560de2acommit | diff
Reduce expected precision of RGB10A2 filtering

The code in the YCbCr tests says that it was copied from the texture
explicit lod tests, but those tests give rgb10a2 a 'low' precision, not
'fp16'. Reduce the precision expected in the YCbCr tests.

Components: Vulkan
VK-GL-CTS Issue: 1731
Affects: dEQP-VK.ycbcr.*
         dEQP-VK.protected_memory.interaction.ycbcr.*

Change-Id: Ib0ae233fc271ded2f9987fe448dcd351e509a0c4
(cherry picked from commit 5deebf4709ed3e2c3aaad3b8bb8703f236309b97)
external/vulkancts/modules/vulkan/ycbcr/vktYCbCrUtil.cpp diff | blob | history
Domain: Graphics System / GL;