projects / platform / upstream / VK-GL-CTS.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0d16754) | patch
Add new tests for (u)int->float conversion
authorTony Wasserka <tony.wasserka@gmx.de>
Thu, 25 Mar 2021 18:26:13 +0000 (19:26 +0100)
committerAlexander Galazin <Alexander.Galazin@arm.com>
Fri, 16 Apr 2021 07:36:13 +0000 (07:36 +0000)
commit02ebdf12f30b3a2b8169557bc03b1eda37d4de89
treeebc375c25caf49a9564502031a8544ad5b8f9e92tree | snapshot
parent0d16754ac1d5e0e2364bf4b1bcbbed0f8089b4d6commit | diff
Add new tests for (u)int->float conversion

This adds a number of special values that triggered incorrect behavior
on our implementation.

Further, some of the min/max tests were too conservative and didn't
actually use the smallest/larged possible values. They now use the full
range of exactly convertible values instead.

Components: Vulkan

Affects: dEQP-VK.spirv_assembly.instruction.graphics.convert*to*.*

Change-Id: Ieadc40b40107318df4351826c5c33b4140c4f7f3
android/cts/master/vk-master-2021-03-01.txt diff | blob | history
android/cts/master/vk-master.txt diff | blob | history
external/vulkancts/modules/vulkan/spirv_assembly/vktSpvAsmInstructionTests.cpp diff | blob | history
external/vulkancts/mustpass/master/vk-default.txt diff | blob | history
Domain: Graphics System / GL;