OSDN Git Service

Unnormalized integer formats now all use the same path
authorAlexis Hetu <sugoi@google.com>
Tue, 18 Jul 2017 17:03:42 +0000 (13:03 -0400)
committerAlexis Hétu <sugoi@google.com>
Thu, 20 Jul 2017 20:20:14 +0000 (20:20 +0000)
commit91dd1c4c3e28a9485139b5bdbf1d2451e8549ebb
treedcc925ffae698e0e394425be38dea011f4ccd629
parente9c8bd76f484147cf8ed7793226846458bf1d2aa
Unnormalized integer formats now all use the same path

Integer formats are: UInt32/Int32/UInt16/Int16/UInt8/Int8
The 32 bit formats required using the floating point path for precision,
while the 16 bit and 8 bit formats were going through the short integer
path. Since all formats need to be treated as Int within a 32 bit Float
in the end, this cl unifies all integer formats under the same path.

Related dEQP tests are: functional.texture.format.sized.2d.*

Bug swiftshader:76

Change-Id: Ia2e0972243290e8b1139081222a3f19ab9e65c9f
Reviewed-on: https://swiftshader-review.googlesource.com/10748
Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
src/Renderer/Surface.cpp
src/Shader/SamplerCore.cpp
src/Shader/SamplerCore.hpp