1. Graphics Graphics Lab @ Korea University cgvr.korea.ac.kr 1 Using Vertex Shader in DirectX 8.1 강 신 진 2002. 02. 05

2. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 2 Overview What is Vertex Programming? Vertex Programming Assembly Language Sample Programs Instruction Set

3. Graphics Graphics Lab @ Korea University cgvr.korea.ac.kr 3 What is Vertex Programming? Part 1

4. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 4 Traditional Rendering Pipeline Traditional Graphics Pipeline frame-buffer anti-aliasing texture blending setup rasterizer transform & lighting

5. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 5 Vertex Shader Rendering Pipeline SetVertexShader() frame-buffer anti-aliasing texture blending setup rasterizer transform & lighting Vertex Program Switch from standard T&L mode to Vertex Program mode

6. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 6 What is Possible? Complete control of transform and lighting HW Complex vertex operations accelerated in HW Custom vertex lighting Custom skinning and blending Custom texture coordinate generation Custom texture matrix operations Custom vertex computations of your choice Offloading vertex computations frees up CPU

7. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 7 What is Possible? Custom transform, lighting, and skinning Directional Light Bump Point LightingKeyframe Interpolation

8. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 8 Vertex Attributes Vertex Program Vertex Output Program Parameters Temporary Registers Read/Write-able 16x4 registers 128 instructions 96x4 registers 12x4 registers Read-only Vertex Programming Conceptual Overview

9. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 9 Vertex Programming Conceptual Overview Vertex Attribute Registers Vertex Program Vertex Result Registers Program Parameter Registers Temporary Registers v[0] v[1] … v[15] c[0] c[1] … c[95] R0 R1 … R10 R11 oPos, oD0 r r r/w w Address Register A0.x

10. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 10 Sample Code Position & Constant Color reg c0= (0,0.5,1.0,2.0) reg c4-7= WorldViewProj matrix reg c8 = constant color reg v0= position ( 4x1 vector ) reg v5= diffuse color const char SimpleVertexShader0[] =“ vs.1.0 //Shader version 1.0 m4x4 oPos, v0, c4 //emit projected position mov oD0, c8 //Diffuse color = c8”

11. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 11 What is Vertex Programming? Vertex Program Assembly language interface to T&L unit GPU instruction set to perform all vertex math Reads an untransformed, unlit vertex Creates a transformed vertex Optionally creates  Lights a vertex  Creates texture coordinates  Creates fog coordinates  Creates point sizes

12. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 12 What is Vertex Programming? Vertex Program Does not create or delete vertices  1 vertex in and 1 vertex out No topological information provided  No edge, face, nor neighboring vertex info Dynamically loadable

13. Graphics Graphics Lab @ Korea University cgvr.korea.ac.kr 13 Vertex Programming Assembly Language Part 2

14. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 14 Instruction Format: Opcodedst, [-]s0 [,[-]s1 [,[-]s2]];#comment Instruction name Destination Register Source0 Register Source1 Register Source2 Register Assembly Language Format

15. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 15 Instruction Format: Opcodedst, [-]s0 [,[-]s1 [,[-]s2]];#comment Instruction name Destination Register Source0 Register Source1 Register Source2 Register Example: MOVr1, r2 R1 x y z w R2 x y z w Assembly Language Format

16. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 16 Simple Example: MOVR1, R2; R1 x y z w 7.0 3.0 6.0 2.0 R2 x y z w 7.0 3.0 6.0 2.0 R1 x y z w 0.0 R2 x y z w 7.0 3.0 6.0 2.0 beforeafter Assembly Example

17. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 17 Source registers can be negated: MOVR1, -R2; beforeafter R1 x y z w 0.0 R2 x y z w 7.0 3.0 6.0 2.0 R1 x y z w -7.0 -3.0 -6.0 -2.0 R2 x y z w 7.0 3.0 6.0 2.0 Assembly Example

18. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 18 Destination register can mask which components are written to… R1  write all components R1.x  write only x component R1.xw  write only x, w components Masking

19. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 19 Destination register masking: MOVR1.xw, -R2; beforeafter R1 x y z w 0.0 R2 x y z w 7.0 3.0 6.0 2.0 R1 x y z w -7.0 0.0 -2.0 R2 x y z w 7.0 3.0 6.0 2.0 Masking

20. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 20 There are 17 instructions in total ARL MOV MUL ADD MAD RCP RSQ DP3 DP4 DST MIN MAX SLT SGE EXP LOG LIT All Instructions

21. Graphics Graphics Lab @ Korea University cgvr.korea.ac.kr 21 Sample Program Part 3

22. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 22 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

23. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 23 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

24. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 24 Step 1 Declare the vertex data struct CUSTOMVERTEX { FLOAT x, y, z; DWORD diffuseColor; }; #define D3DFVF_CUSTOMVERTEX (D3DFVF_XYZ|D3DFVF_DIFFUSE) CUSTOMVERTEX g_Vertices[]= { { -1.0f, -1.0f, 0.0f, 0xffff0000 }, { +1.0f, -1.0f, 0.0f, 0xff00ff00 }, { +1.0f, +1.0f, 0.0f, 0xff0000ff }, { -1.0f, +1.0f, 0.0f, 0xffffff00 }, };

25. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 25 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

26. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 26 Basic Instruction (Step 2) DP3: Three-Component Dot Product Function: Computes the three-component (x,y,z) dot product of two source vectors and replicates the result across the destination register. Syntax: DP3dest, src0, src1;

27. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 27 Basic Instruction (Step 2) beforeafter DP3 Example: DP3R1, R6, R6; x y z w R6 3.0 2.0 1.0 R1 0.0 x y z w R6 3.0 2.0 1.0 R1 14.0

28. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 28 Sample Code – 1 (Step 2) Position & Diffuse Color reg c0= (0,0.5,1.0,2.0) reg c4-7= WorldViewProj matrix reg c8 = constant color reg v0= position ( 4x1 vector ) reg v5= diffuse color const char SimpleVertexShader1[] =“ vs.1.0 //Shader version 1.0 dp4 oPos.x, v0, c4 //emit projected x position dp4 oPos.y, v0, c5 //emit projected y position dp4 oPos.z, v0, c6 //emit projected z position dp4 oPos.w, v0, c7 //emit projected w position mov oD0, v5 //Diffuse color = vertex color ”

29. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 29 Sample Code – 2 (Step 2) Position & Texture reg c0= (0,0.5,1.0,2.0) reg c4-7= WorldViewProj matrix reg c8 = constant color reg v0= position ( 4x1 vector ) reg v5= diffuse color reg v7= texcoords ( 2x1 vector ) const char SimpleVertexShader2[] =“ vs.1.0 //Shader version 1.0 dp4 oPos.x, v0, c4 //emit projected x position dp4 oPos.y, v0, c5 //emit projected y position dp4 oPos.z, v0, c6 //emit projected z position dp4 oPos.w, v0, c7 //emit projected w position mov oT0.xy, v7 //copy texcoords”

30. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 30 Sample Code – 3 (Step 2) Position & Lighting reg c0= (0,0.5,1.0,2.0) reg c4-7= WorldViewProj matrix reg c8 = constant color reg v0= position ( 4x1 vector ) reg v5= diffuse color reg v7= texcoords ( 2x1 vector ) const char SimpleVertexShader3[] = vs.1.0 //Shader version 1.0 dp4 oPos.x, v0, c4 //emit projected x position dp4 oPos.y, v0, c5 //emit projected y position dp4 oPos.z, v0, c6 //emit projected z position dp4 oPos.w, v0, c7 //emit projected w position dp3 r0.x, v3, c12 //N dot L in world space mul oD0, r0.x, v5 //Calculate color intensity mov oT0.xy, v7 //copy texcoords

31. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 31 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

32. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 32 Step 3 Check for vertex shader support D3DCAPS8 caps; m_pd3dDevice->GetDeviceCaps(&caps); if( D3DSHADER_VERSION_MAJOR( caps.VertexShaderVersio n ) < 1 ) return E_FAIL;

33. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 33 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

34. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 34 Step 4 Declare the shader registers DWORD dwDecl[] = { D3DVSD_STREAM(0), D3DVSD_REG(D3DVSDE_POSITION, D3DVSDT_FLOAT3), D3DVSD_REG( D3DVSDE_DIFFUSE, D3DVSDT_D3DCOLOR ), D3DVSD_END() };

35. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 35 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

36. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 36 Step 5 Create the shader TCHAR strPath[512]; LPD3DXBUFFER pCode; DXUtil_FindMediaFile( strPath, _T("VertexShader.vsh") ); D3DXAssembleShaderFromFile( strPath, 0, NULL, &pCode, NULL ); m_pd3dDevice->CreateVertexShader( dwDecl, (DWORD*)pCode- >GetBufferPointer(), &m_hVertexShader, 0 ))) pCode->Release();

37. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 37 Vertex Shader Frame Work in DX 8.1 Step 1: Declare the vertex data Step 2: Design the shader functionality Step 3: Check for vertex shader support Step 4: Declare the shader registers Step 5: Create the shader Step 6: Render the output pixels

38. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 38 Step 6 Render the output pixels m_pd3dDevice->SetVertexShaderConstant( 0, &mat, 4 ); float color[4] = {0,1,0,0}; m_pd3dDevice->SetVertexShaderConstant( 4, &color, 1 ); m_pd3dDevice->SetStreamSource( 0, m_pQuadVB, sizeof(CUSTOMVERTEX) ); m_pd3dDevice->SetVertexShader( m_hVertexShader ); m_pd3dDevice->DrawPrimitive( D3DPT_TRIANGLEFAN, 0, 2 );

39. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 39 Demo Program

40. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 40 Performance For Optimal performance Be clever Exploit vector parallelism (Ex. 4 scalar adds with a vector add) Swizzle and negate away (no performance penalty for doing so) Use LIT and DST effectively

41. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 41 Summary – Vertex Programs Increased programmability Customizable engine for transform, lighting, texture coordinate generation, and more. Facilitates setup for per-fragment shading. Allows animation/deformation through key-frame interpolation and skinning. Accelerated in Future Generation GPUs Offloads CPU tasks to GPU yielding higher performance.

42. Graphics Graphics Lab @ Korea University cgvr.korea.ac.kr 42 The Instruction Set Appendix

43. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 43 There are 17 instructions in total ARL MOV MUL ADD MAD RCP RSQ DP3 DP4 DST MIN MAX SLT SGE EXP LOG LIT All Instructions

44. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 44 RCP RCP: Reciprocal Function: Inverts the value of the source and replicates the result across the destination register. Syntax: RCPdest, src0.C; where ‘C’ is x, y, z, or w

45. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 45 RCP beforeafter RCP Example: RCPR1, R2.w; x y z w R2 7.0 3.0 6.0 2.0 R1 0.0 x y z w R2 7.0 3.0 6.0 2.0 R1 0.5

46. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 46 RSQ RSQ: Reciprocal Square Root Function: Computes the inverse square root of the absolute value of the source scalar and replicates the result across the destination register. Syntax: RSQdest, src0.C; where ‘C’ is x, y, z, or w

47. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 47 RSQ beforeafter RSQ Example: RSQR1.x, R5.x; x y z w R5 -4.0 3.0 7.0 9.0 R1 0.0 x y z w R5 -4.0 3.0 7.0 9.0 R1 0.5 0.0

48. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 48 SLT SLT: Set On Less Than Function: Performs a component-wise assignment of either 1.0 or 0.0. 1.0 is assigned if the value of the first source is less than the value of the second. Otherwise, 0.0 is assigned. Syntax: SLTdest, src0, src1;

49. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 49 SLT beforeafter SLT Example: SLTR1, R2, R3; x y z w R3 2.0 2.1 5.0 7.0 R2 7.0 3.0 6.0 2.0 R1 0.0 x y z w R3 2.0 2.1 5.0 7.0 R2 7.0 3.0 6.0 2.0 R1 0.0 1.0

50. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 50 LIT LIT: Light Coefficients Function: Computes ambient, diffuse, and specular lighting coefficients from a diffuse dot product, a specular dot product, and a specular power. Assumes: src0.x = diffuse dot product(N L) src0.y = specular dot product(N H) src0.w = power(m)

51. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 51 LIT LIT: Light Coefficients Syntax: LITdest, src0 Result: dest.x = 1.0 (ambient coeff.) dest.y = CLAMP(src0.x, 0, 1) = CLAMP(N L, 0, 1) (diffuse coeff.) dest.z = (see next slide…) (specular coeff.) dest.w = 1.0

52. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 52 LIT LIT: Light Coefficients Result: (Recall: src0.x  N L) if ( src0.x > 0.0 ) dest.z = (MAX(src0.y,0)) (ECLAMP(src0.w,-128,128)) = (MAX(N H,0)) m where m in (-128,128) otherwise, dest.z = 0.0 (dest.z is specular coeff. as defined by OpenGL)

53. CGVR Graphics Lab @ Korea University cgvr.korea.ac.kr 53 LIT beforeafter LIT Example: LITR1, R7; x y z w R7 0.3 0.8 0.0 2.0 R1 0.0 R7 0.3 0.8 0.0 2.0 x y z w R1 1.0 0.3 0.64 1.0 (ambient) (diffuse) (specular) (Good to 8+ bits)