Batch up draws into triangle fans as large as possible when drawing convex
edge AA polys, so we minimize state changes and GL calls. This requires
querying GL for the maximum number of fragment uniforms. It also makes the
shader generator produce custom shaders for the number of relevant edges.
This gives a ~5X speedup on the "Shapes" SampleApp.
Review URL: http://codereview.appspot.com/4536070/
git-svn-id: http://skia.googlecode.com/svn/trunk@1380 2bbb7eff-a529-9590-31e7-b0007b416f81
diff --git a/gpu/src/GrGpuGLShaders.cpp b/gpu/src/GrGpuGLShaders.cpp
index 08845a9..bbf9719 100644
--- a/gpu/src/GrGpuGLShaders.cpp
+++ b/gpu/src/GrGpuGLShaders.cpp
@@ -193,7 +193,7 @@
idx = (int)(random.nextF() * (kNumStages+1));
pdesc.fFirstCoverageStage = idx;
- pdesc.fUsesEdgeAA = (random.nextF() > .5f);
+ pdesc.fEdgeAANumEdges = (random.nextF() * (getMaxEdges() + 1));
for (int s = 0; s < kNumStages; ++s) {
// enable the stage?
@@ -442,16 +442,17 @@
void GrGpuGLShaders::flushEdgeAAData() {
const int& uni = fProgramData->fUniLocations.fEdgesUni;
if (GrGLProgram::kUnusedUniform != uni) {
- float edges[18];
- memcpy(edges, fCurrDrawState.fEdgeAAEdges, sizeof(edges));
+ int count = fCurrDrawState.fEdgeAANumEdges;
+ Edge edges[kMaxEdges];
// Flip the edges in Y
float height = fCurrDrawState.fRenderTarget->height();
- for (int i = 0; i < 6; ++i) {
- float b = edges[i * 3 + 1];
- edges[i * 3 + 1] = -b;
- edges[i * 3 + 2] += b * height;
+ for (int i = 0; i < count; ++i) {
+ edges[i] = fCurrDrawState.fEdgeAAEdges[i];
+ float b = edges[i].fY;
+ edges[i].fY = -b;
+ edges[i].fZ += b * height;
}
- GR_GL(Uniform3fv(uni, 6, edges));
+ GR_GL(Uniform3fv(uni, count, &edges[0].fX));
}
}
@@ -701,7 +702,7 @@
desc.fColorType = GrGLProgram::ProgramDesc::kAttribute_ColorType;
}
- desc.fUsesEdgeAA = fCurrDrawState.fFlagBits & kEdgeAA_StateBit;
+ desc.fEdgeAANumEdges = fCurrDrawState.fEdgeAANumEdges;
for (int s = 0; s < kNumStages; ++s) {
GrGLProgram::ProgramDesc::StageDesc& stage = desc.fStages[s];