From 1852ebd65df1748f071947204e440f1dece30f31 Mon Sep 17 00:00:00 2001
From: mobserveur <mobserveur@gmail.com>
Date: Fri, 28 Jun 2024 11:23:08 +0200
Subject: Apple GPU Optimisations

This commit massively improves 2D UI, 3D UI and 3D world rendering performance and smoothness.
---
 indra/llrender/llrender.h | 26 +++++++++++++++-----------
 1 file changed, 15 insertions(+), 11 deletions(-)

(limited to 'indra/llrender/llrender.h')

diff --git a/indra/llrender/llrender.h b/indra/llrender/llrender.h
index 020682f526..555234356f 100644
--- a/indra/llrender/llrender.h
+++ b/indra/llrender/llrender.h
@@ -416,11 +416,15 @@ public:
 
     void begin(const GLuint& mode);
     void end();
-    void vertex2i(const GLint& x, const GLint& y);
-    void vertex2f(const GLfloat& x, const GLfloat& y);
-    void vertex3f(const GLfloat& x, const GLfloat& y, const GLfloat& z);
-    void vertex2fv(const GLfloat* v);
-    void vertex3fv(const GLfloat* v);
+
+    LL_FORCE_INLINE void vertex2i(const GLint& x, const GLint& y) { vertex4a(LLVector4a((GLfloat)x,(GLfloat)y,0.f)); }
+    LL_FORCE_INLINE void vertex2f(const GLfloat& x, const GLfloat& y) { vertex4a(LLVector4a(x,y,0.f)); }
+    LL_FORCE_INLINE void vertex3f(const GLfloat& x, const GLfloat& y, const GLfloat& z) { vertex4a(LLVector4a(x,y,z)); }
+    LL_FORCE_INLINE void vertex2fv(const GLfloat* v) { vertex4a(LLVector4a(v[0],v[1],0.f)); }
+    LL_FORCE_INLINE void vertex3fv(const GLfloat* v) { vertex4a(LLVector4a(v[0],v[1],v[2])); }
+
+    void vertex4a(const LLVector4a& v);
+
 
     void texCoord2i(const GLint& x, const GLint& y);
     void texCoord2f(const GLfloat& x, const GLfloat& y);
@@ -440,9 +444,9 @@ public:
     void diffuseColor4ubv(const U8* c);
     void diffuseColor4ub(U8 r, U8 g, U8 b, U8 a);
 
-    void vertexBatchPreTransformed(LLVector3* verts, S32 vert_count);
-    void vertexBatchPreTransformed(LLVector3* verts, LLVector2* uvs, S32 vert_count);
-    void vertexBatchPreTransformed(LLVector3* verts, LLVector2* uvs, LLColor4U*, S32 vert_count);
+    void vertexBatchPreTransformed(LLVector4a* verts, S32 vert_count);
+    void vertexBatchPreTransformed(LLVector4a* verts, LLVector2* uvs, S32 vert_count);
+    void vertexBatchPreTransformed(LLVector4a* verts, LLVector2* uvs, LLColor4U*, S32 vert_count);
 
     void setColorMask(bool writeColor, bool writeAlpha);
     void setColorMask(bool writeColorR, bool writeColorG, bool writeColorB, bool writeAlpha);
@@ -500,7 +504,7 @@ private:
     bool                mCurrColorMask[4];
 
     LLPointer<LLVertexBuffer>   mBuffer;
-    LLStrider<LLVector3>        mVerticesp;
+    LLStrider<LLVector4a>        mVerticesp;
     LLStrider<LLVector2>        mTexcoordsp;
     LLStrider<LLColor4U>        mColorsp;
     std::array<LLTexUnit, LL_NUM_TEXTURE_LAYERS> mTexUnits;
@@ -512,8 +516,8 @@ private:
     eBlendFactor mCurrBlendAlphaSFactor;
     eBlendFactor mCurrBlendAlphaDFactor;
 
-    std::vector<LLVector3> mUIOffset;
-    std::vector<LLVector3> mUIScale;
+    std::vector<LLVector4a> mUIOffset;
+    std::vector<LLVector4a> mUIScale;
 
 };
 
-- 
cgit v1.2.3