commit 06e9f324ef5b723622647a6f111c38cb2d479869
author Arun <arun.demeure@imgtec.com> Mon Jan 23 11:59:21 2017 +0000
committer Chris Craik <ccraik@google.com> Thu Jun 01 13:42:12 2017 -0700
tree 1d4e45635bf9eb8c9c18e494f5be25711678ed84
parent 142a392b75e7bf8e4e880b5cc49d111fd8d82533

Optimised hwui rounded corners shader

This is a ~2x performance increase for the rounded corners shader which
was previously shader limited even on many high-end GPUs. It simplifies
the calculations by pre-dividing by the radius, which also allows to
maintain mediump/FP16 precision (this improves performance and might
improve image quality on GPUs that don't support highp e.g. Mali-400).
It also uses a conditional in the shader to save much of the work on
the vast majority of pixels which are not close to the corners.

Test: manual - visual inspection on fugu (nexus player)

Change-Id: I38f491cfb006e2991ef0b9d6b3d67d2c4d328d44

libs/hwui/ProgramCache.cpp

diff --git a/libs/hwui/ProgramCache.cpp b/libs/hwui/ProgramCache.cpp
index 59225e1..f70723c3 100644
--- a/libs/hwui/ProgramCache.cpp
+++ b/libs/hwui/ProgramCache.cpp
@@ -58,7 +58,9 @@
         "uniform mat4 textureTransform;\n"
         "uniform mediump vec2 textureDimension;\n";
 const char* gVS_Header_Uniforms_HasRoundRectClip =
-        "uniform mat4 roundRectInvTransform;\n";
+        "uniform mat4 roundRectInvTransform;\n"
+        "uniform mediump vec4 roundRectInnerRectLTWH;\n"
+        "uniform mediump float roundRectRadius;\n";
 const char* gVS_Header_Varyings_HasTexture =
         "varying vec2 outTexCoords;\n";
 const char* gVS_Header_Varyings_HasColors =
@@ -87,7 +89,7 @@
         "varying vec2 ditherTexCoords;\n",
 };
 const char* gVS_Header_Varyings_HasRoundRectClip =
-        "varying highp vec2 roundRectPos;\n";
+        "varying mediump vec2 roundRectPos;\n";
 const char* gVS_Main =
         "\nvoid main(void) {\n";
 const char* gVS_Main_OutTexCoords =
@@ -125,7 +127,7 @@
         "    alpha = vtxAlpha;\n";
 
 const char* gVS_Main_HasRoundRectClip =
-        "    roundRectPos = (roundRectInvTransform * transformedPosition).xy;\n";
+        "    roundRectPos = ((roundRectInvTransform * transformedPosition).xy / roundRectRadius) - roundRectInnerRectLTWH.xy;\n";
 const char* gVS_Footer =
         "}\n\n";
 
@@ -169,8 +171,8 @@
 };
 
 const char* gFS_Uniforms_HasRoundRectClip =
-        "uniform vec4 roundRectInnerRectLTRB;\n"
-        "uniform float roundRectRadius;\n";
+        "uniform mediump vec4 roundRectInnerRectLTWH;\n"
+        "uniform mediump float roundRectRadius;\n";
 
 const char* gFS_Main =
         "\nvoid main(void) {\n"
@@ -315,15 +317,18 @@
         "    fragColor = blendColors(colorBlend, fragColor);\n"
 };
 
-// Note: LTRB -> xyzw
+// Note: LTWH (left top width height) -> xyzw
+// roundRectPos is now divided by roundRectRadius in vertex shader
+// after we also subtract roundRectInnerRectLTWH.xy from roundRectPos
 const char* gFS_Main_FragColor_HasRoundRectClip =
-        "    mediump vec2 fragToLT = roundRectInnerRectLTRB.xy - roundRectPos;\n"
-        "    mediump vec2 fragFromRB = roundRectPos - roundRectInnerRectLTRB.zw;\n"
+        "    mediump vec2 fragToLT = -roundRectPos;\n"
+        "    mediump vec2 fragFromRB = roundRectPos - roundRectInnerRectLTWH.zw;\n"
 
-        // divide + multiply by 128 to avoid falling out of range in length() function
-        "    mediump vec2 dist = max(max(fragToLT, fragFromRB), vec2(0.0, 0.0)) / 128.0;\n"
-        "    mediump float linearDist = roundRectRadius - (length(dist) * 128.0);\n"
-        "    gl_FragColor *= clamp(linearDist, 0.0, 1.0);\n";
+        // since distance is divided by radius, it's in [0;1] so precision is not an issue
+        // this also lets us clamp(0.0, 1.0) instead of max() which is cheaper on GPUs
+        "    mediump vec2 dist = clamp(max(fragToLT, fragFromRB), 0.0, 1.0);\n"
+        "    mediump float linearDist = clamp(roundRectRadius - (length(dist) * roundRectRadius), 0.0, 1.0);\n"
+        "    gl_FragColor *= linearDist;\n";
 
 const char* gFS_Main_DebugHighlight =
         "    gl_FragColor.rgb = vec3(0.0, gl_FragColor.a, 0.0);\n";

libs/hwui/renderstate/RenderState.cpp

diff --git a/libs/hwui/renderstate/RenderState.cpp b/libs/hwui/renderstate/RenderState.cpp
index c4e8e4f..2a3a9f3 100644
--- a/libs/hwui/renderstate/RenderState.cpp
+++ b/libs/hwui/renderstate/RenderState.cpp
@@ -274,14 +274,20 @@
         // TODO: avoid query, and cache values (or RRCS ptr) in program
         const RoundRectClipState* state = glop.roundRectClipState;
         const Rect& innerRect = state->innerRect;
-        glUniform4f(fill.program->getUniform("roundRectInnerRectLTRB"),
-                innerRect.left, innerRect.top,
-                innerRect.right, innerRect.bottom);
-        glUniformMatrix4fv(fill.program->getUniform("roundRectInvTransform"),
-                1, GL_FALSE, &state->matrix.data[0]);
 
         // add half pixel to round out integer rect space to cover pixel centers
         float roundedOutRadius = state->radius + 0.5f;
+
+        // Divide by the radius to simplify the calculations in the fragment shader
+        // roundRectPos is also passed from vertex shader relative to top/left & radius
+        glUniform4f(fill.program->getUniform("roundRectInnerRectLTWH"),
+                innerRect.left / roundedOutRadius, innerRect.top / roundedOutRadius,
+                (innerRect.right - innerRect.left) / roundedOutRadius,
+                (innerRect.bottom - innerRect.top) / roundedOutRadius);
+
+        glUniformMatrix4fv(fill.program->getUniform("roundRectInvTransform"),
+                1, GL_FALSE, &state->matrix.data[0]);
+
         glUniform1f(fill.program->getUniform("roundRectRadius"),
                 roundedOutRadius);
     }