version 2.2.0 nvidia gpu

VVeerrssiioonn 22..22..00

NNVVIIDDIIAA GGPPUU 22..22..00

NVIDIA , , , , , ( ) . NVIDIA , , , , .

. , NVIDIA Corporation 3 . NVIDIA Corporation . . . NVIDIA Corporation NVIDIA Corporation .

NVIDIA, NVIDIA , GeForce, NVIDIA Quadro NVIDIA Corporation . .

2004 by NVIDIA Corporation. All rights reserved.

2.2.0 11/16/2004

ps_3_0

2.1.0 07/20/2004

2.0.4 07/15/2004 MRT

2.0.3 06/25/2004 GPU

2.0.0 06/01/2004 NV40(GeForce 6 )

NVIDIA GPU

1.0.0 07/14/2003 GeForce FX

2

NVIDIA GPU

3

Chapter 1. .......................................................................................8

1.1. .........................................................................................8

1.2. ...............................................................................9

Chapter 2. ..............................................................10

2.1. ................................................................... 10

2.2. ......................................................................... 11

2.2.1. 11

2.2.2. 12

2.2.3. NVPerfHUD 13

2.3. : CPU......................................................................... 14

2.4. : GPU......................................................................... 15

Chapter 3. GPU ......................................................................17

3.1. ..................................................................... 17

3.2. ................................................................................ 20

3.2.1. 20

3.3. ......................................................................... 20

3.3.1. Indexed Primitive Call 20

3.4. .................................................................................... 21

3.4.1. 21

3.4.2. ps_2_a 21

3.4.3. 22

3.4.4. 23

3.4.5. Interpolant 24

3.4.6. 24

3.4.7. 25

3.4.8. 25

3.4.9. 26

3.4.10. 26

3.4.11. 27

3.4.12. mul() 27

3.4.13. saturate() D3DTADDRESS_CLAMP ( GL_CLAMP_TO_EDGE) 27

3.4.14. Interpolant 28

3.5. ................................................................................. 28

3.5.1. 28

3.5.2. 28

3.5.3. 29

3.6. ....................................................................................... 31

3.6.1. 2 Z 31

3.6.2. Z 32

3.6.3. 32

3.6.4. 33

3.7. (Antialiasing) ....................................................... 33

Chapter 4. GeForce 6 ................................................35

4.1. Shader Model 3.0 ............................................................. 35

4.1.1. Pixel Shader 3.0 36

4.1.2. Vertex Shader 3.0 37

4.1.3. (Dynamic Branching) 37

4.1.4. 38

4.1.5. (Instancing) 38

4.1.6. 39

4

NVIDIA GPU

5

4.2. sRGB ............................................................................ 39

4.3. .................................................................. 39

4.4. ............................................................... 40

4.5. .................................................................. 41

4.5.1. 41

4.6. MRT(Multiple Render Target) .................................................... 41

4.7. ...................................................................... 43

4.8. ............................................................... 44

4.9. .................................................................................... 45

Chapter 5. GeForce FX .........................................................47

5.1. ......................................................................... 47

5.2. ..................................................................... 47

5.3. DirectX ................................................................. 48

5.4. OpenGL ................................................................ 48

5.5. 16 ......................................................... 49

5.6. ............................................................... 50

5.7. DirectX ps_2_x ps_2_a ......................................... 51

5.8. ...................................................... 52

5.9. .................................................................................... 52

5.10. .................................................................. 52

5.11. ....................................................................................... 53

Chapter 6. ..................................................................................55

6.1. GPU ................................................................................ 55

6.2. .................................................................. 56

Chapter 7. NVIDIA SLI GPU ..............................................61

7.1. SLI ? ...................................................................... 61

7.2. CPU ...................................................................... 63

7.3. VSync ( ) ............................................................. 63

7.4. 2 (Lag) ............................................... 64

7.5. .................. 65

7.6. ............................................... 66

7.7. D3DPOOL_MANAGED .................................. 66

Chapter 8. ..............................................................................67

8.1. ? ...................................................... 67

8.2. ............................................................... 68

8.3. ........................................................ 68

8.3.1. 68

8.3.2. 69

8.3.3. 69

8.3.4. 3D 2D 70

8.3.5. 70

8.3.6. (Dirty Rectangle) 70

8.3.7. 70

8.3.8. 70

8.3.9. 71

8.3.10. 71

8.3.11. 71

8.3.12. 71

8.3.13. 71

8.3.14. 71

8.3.15. 72

8.4. .................................................................. 72

8.4.1. 72

8.4.2. Out of the Monitor 72

8.4.3. 72

8.4.4. 73

8.4.5. 73

6

NVIDIA GPU

7

8.5. API ........................................................................... 73

8.6. ................................................................................ 74

Chapter 9. .................................................................................75

9.1. NVPerfHUD.............................................................................. 75

9.2. NVShaderPerf .......................................................................... 76

9.3. FX Composer ........................................................................... 76

9.4. NVIDIA Melody ........................................................................ 77

9.5. ................................................... 77

Chapter 1.

1.1.

, API, (GPU) . . [email protected] .

.

1 ( ) .

2 .

3 . , .

4 NVIDIA GeForce 6 NV4X Quadro FX GPU . , .

5 NVIDIA GeForce FX NV3X Quadro FX GPU .

mailto:[email protected]://www.nvidia.com/page/geforce6.htmlhttp://www.nvidia.com/page/quadrofx_family.htmlhttp://www.nvidia.com/page/fx_desktop.htmlhttp://www.nvidia.com/page/quadrofx_family.html

NVIDIA GPU

9

1.2.

, .

6 NVIDIA GPU , , GPU .

7 NVIDIA SLI(Scalable Link Interface) . SLI GPU .

8 . .

9 NVIDIA .

10 GPU .

[email protected] .

mailto:[email protected]

Chapter 2.

.

2.1.

, . , NVPerfHUD ( NVPerfHUD ) (ms) ..

.

. , DiretX Debug , .

. , DLL, DirextX .

( ) .

. , . .

http://developer.nvidia.com/object/nvperfhud_home.html

NVIDIA GPU

. .

. , CPU, GPU . .

11

2.2.

2.2.1.

, . . . . . .

1.

CPU, AGP , GPU 1 ) . . .

CPU GPU . NVPerfHUD (NVPerfHUD 9.1 ) GPU . GPU 1 , CPU . GPU 1 , CPU GPU , CPU . GPU GPU .

2.2.2.

. , .

. . . Windows , AMD CodeAnalyst(http://www.amd.com/us-en/Processors/DevelopWithAMD/0,,30_2252_3604,00.html) VTune (http://www.intel.com/software/products/vtune/) . , .

CPU GPU . CPU , BIOS . , . CPU , CPU .

12

http://www.amd.com/us-en/Processors/DevelopWithAMD/0,,30_2252_3604,00.htmlhttp://www.amd.com/us-en/Processors/DevelopWithAMD/0,,30_2252_3604,00.htmlhttp://www.intel.com/software/products/vtune/

NVIDIA GPU

13

2.2.3.

GPU . Coolbit( 6 ) . , , , (, ).

GPU . Coolbit( 6 ) . , (GPU ).

, CPU , GPU , GPU CPU , GPU . CPU n% n% , CPU . GPU n% n% , GPU .

NVPerfHUD

CPU Null Hardware(Null HW) . ( ), GPU . Null HW , CPU .

NVPerfHUD 2.0 (draw call) Null Hardware Driver . CPU . . NVPerfHUD .

NVPerfHUD GPU , GPU . NVPerfHUD GPU , GPU .

NVPerfHUD . http://developer.nvidia.com/object/nvperfhud_home.html .

2.3. : CPU

CPU , CPU . CPU .

( DLL )

(nv4disp.dll, nvoglnt.dll)

DirectX (d3d9.dll)

DirectX (hal32.dll)

CPU CPU , CPU . . , . CPU .

, . hal32.dll, d3d9.dll , nvoglnt.dll CPU , API . CPU , ? (batch) . .

http://developer.nvidia.com/docs/IO/8230/BatchBatchBatch.ppt

http://download.nvidia.com/developer/presentations/GDC_2004/Dx9Optimization.pdf

14

http://developer.nvidia.com/object/nvperfhud_home.htmlhttp://developer.nvidia.com/docs/IO/8230/BatchBatchBatch.ppthttp://download.nvidia.com/developer/presentations/GDC_2004/Dx9Optimization.pdfhttp://download.nvidia.com/developer/presentations/GDC_2004/Dx9Optimization.pdf

NVIDIA GPU

15

2.4.

NVPerfHUD . ( ), .

CPU .

Lock ? Lock CPU GPU , GPU CPU . CPU , . lock CPU .

CPU GPU ? CPU GPU , GPU ! CPU CPU .

CPU GPU . GPU ?

. , 2 2 . Shader Model 3.0 , draw overhead . Shader Model 3.0 4.1 .

: GPU

GPU . GPU , . http://developer.nvidia.com/docs/IO/4449/SUPP/GDC2003_PipelinePerformance.ppt .

NVPerfHUD GPU . , LOD 2 2 . , .

http://developer.nvidia.com/docs/IO/4449/SUPP/GDC2003_PipelinePerformance.ppthttp://developer.nvidia.com/docs/IO/4449/SUPP/GDC2003_PipelinePerformance.ppt

NVPerfHUD .

GPU , 3 .

16

Chapter 3. GPU

GeForceFX GeForce 6 GPU . . . .

GPU , GPU : , , (Programming Techniques, Tips, and Tricks for Real-Time Graphics) . .

http://developer.nvidia.com/object/gpu_gems_samples.html .

3.1.

GeForceFX GeForce 6 GPU . , .

http://developer.nvidia.com/object/gpu_gems_samples.html

GPU

CPU

. . http://developer.nvidia.com/object/nv_texture_tools.html

DirectX , Instancing API SetMatrix, .

GPU

Indexed Primitive Call . Use DirectX 9 [ID3DXMesh:OptimizeInplace() ID3DXMesh:Optimize()] .

Indexed Primitive Call NVTriStrip . http://developer.nvidia.com/object/nvtristrip_library.html

GPU

. , . , .

ps_2_* , ps_2_a profile .

. half float .

.

.

18

http://developer.nvidia.com/object/nv_texture_tools.htmlhttp://developer.nvidia.com/object/nvtristrip_library.html

NVIDIA GPU

19

.

.

. .

FXComposer .

sincos, log, exp , .

GPU

. .

. . http://developer.nvidia.com/object/nv_texture_tools.html

. , .

GPU

2 Z Z(Z-cull)

(Antialiasing)

http://developer.nvidia.com/object/nv_texture_tools.html

GPU

3.2.

3.2.1.

3.3.

(batching) API , API . API . API . , . . GPU , .

3.3.1. Indexed Primitive Call

Indexed Primitive Call GPU . , . .

DirextX , ID3DXMesh class OptimizeInPlace() Optimize() .

NVTriStrip . NVTriStrip , http://developer.nvidia.com/object/nvtristrip_library.html .

20

http://developer.nvidia.com/object/nvtristrip_library.html

NVIDIA GPU

21

3.4.

. , . . . GPU 100% . GeForce FX GPU 20 ! GeForce 6 GPU .

3.4.1.

3.4.2.

. , 8 , ps_2_0 .

ps_2_a

HLSL (fxc.exe) . GeForce FX GPU ps_2_0 , ps_2_a . ps_2_a GeForce FX ps_2_0 . ps_2_a ps_2_0 . ps_2_a 2003 7 HLSL .

, fxc (DirectX 9.0c ) . . GeForce 6 GPU , .

GPU

3.4.3.

. GeForce FX GeForce 6 GPU 32 16 ( float half ) 12 (fixed ) . float s23e8 IEEE . half IEEE , s10e5 . 12 fixed [-2,2] , ps_2_0 . fixed DirextX ps_1_0~ps_1_4 , OpenGL NV_fragment_program extension Cg .

.

fixed , .

, half flat . half 3 , 99% 32 1 LSB() !

, float ..

/Gpp (2003 7 HLSL ) 1/2 . , . , . , 1/2 (/Gpp ).

half fixed , , , , . ps_2_0 DirectX , _pp .

OpenGL ARB_fragment_program , ARB_precision_hint_fastest

22

NVIDIA GPU

23

NV_fragment_program (http://www.nvidia.com/dev_content/nvopenglspecs/GL_NV_fragment_program_option.txt ).

fixed half (: tex2D*diffuseColor ).

OpenGL GeForce FX , ( ) .

, 1/2 . 1/2 , .

, , 1/2 .

3.4.4.

, . .

. p = sqrt(Rx + Ry + (Rz + 1) ) 2 2 2

. p = sqrt(Rx2 + Ry2 + Rz2 + 2Rz + 1)

(3.4.8 3.4.6 ), 1.0 . . p = sqrt(2 * (Rz + 1)) = 1.414*sqrt(Rz + 1)

1.414 (3.4.8 ), .

http://www.nvidia.com/dev_content/nvopenglspecs/GL_NV_fragment_program_option.txthttp://www.nvidia.com/dev_content/nvopenglspecs/GL_NV_fragment_program_option.txt

GPU

(normalize(N), normalize(L)) .

(N/|N|) dot (L/|L|) . 2 rsq(reciprocal square root) .

. (N/|N|) dot (L/|L|) = (N dot L) / (|N| * |L|) = (N dot L) / (sqrt( (N dot N) * (L dot L) ) = (N dot L) * rsq( (N dot N) * (L dot L) ) rsq .

3.4.5.

3.4.6.

Interpolant

. , , 3 q . .

// Bad practice

tangent = float4(tangentVec, viewVec.x)

binormal = float4(binormalVec, viewVec.y)

normal = float4(normalVec, viewVec.z)

interpolant .

. , . float3 reflect(float3 I, float3 N) {

return (2.0*dot(I,N)/dot(N,N))*N I;

}

. ,

24

NVIDIA GPU

25

3.4.7.

3.4.8.

. , reflect() , , . .

( ) . , . . , 1.0 .

, . , 1.0 ( ), .

. ( ) , . . half4 main(float2 diffuse : TEXCOORD0,

uniform sampler2D diffuseTex,

uniform half4 g_OverbrightColor) {

return tex2D(diffuseTex, diffuse) * g_OverbrightColor * 2.0;

}

g_OverbrightColor CPU 2.0 , .

. , HLSL CPU .

GPU

, materialColor * lightColor . , CPU .

GPU CPU . . Zpr (pack row-major) Zpc (pack column-major) .

3.4.9.

3.4.10.

0, 1, 255 . . .

, . , CPU, AGP , . .

? , . .

? , . .

? , . () 20 ,

26

NVIDIA GPU

27

3.4.11.

3.4.12.

3.4.13.

GeForce FX . , . ( 3.4 .) NVShaderPerf . GeForce 6 .

, . , . .

, . , (cube map) , .

mul()

mul() . , interpolant .

saturate() D3DTADDRESS_CLAMP ( GL_CLAMP_TO_EDGE)

saturate() GPU . , [0..1] .

GPU

3.4.14.

3.5.

Interpolant

(TEXCOORD ) . TEXCOORD0 TEXCOORD1, TEXCOORD2 .

3.5.1.

, . . 33% . 3D . 30~40% .

. , . , . , . (NVIDIA ) . http://developer.nvidia.com/object/nv_texture_tools.html .

3.5.2.

, . . , . , . ( , ),

28

http://developer.nvidia.com/object/nv_texture_tools.html

NVIDIA GPU

29

. .

Adobe Photoshop . . http://developer.nvidia.com/object/nv_texture_tools.html . .

3.5.3.

. OTF(on-the-fly) . GeForce FX , . FX Composer . FX Composer http://developer.nvidia.com/FXComposer .

. , log exp ps_2_0 , .

3.5.3.1.

2D

. (N dot L) (N dot H) 2D . (u, v) . max(N dot L,0) + Ks*pow((N dot L>0) ? max(N dot H,0) : 0), n)

Blinn .

1D ARGB

http://developer.nvidia.com/object/nv_texture_tools.htmlhttp://developer.nvidia.com/FXComposer

GPU

(N dot H) 1D ARGB . (N dot H) . .

( (N dot H)4, (N dot H)8, (N dot H)12, (N dot H)16 )

, 4 . GeForce 4- .

3D

3D (mix) . 2D () .

. .

3.5.3.2.

ps_1_* shader , . 16 . x y , z .

V 1 . V . , Taylor (x=1 / sqrt(x)) 1 /||V|| .

1 / sqrt(x) ~ 1 + (1 - x)

.

V / ||V|| = V / sqrt(||V||^2) = V + V (1 - ||V||^2)

2 .

dp3_sat r1, r0, r0

mad_d2 r1, r0, 1-r1, r0_x2

30

NVIDIA GPU

31

r0 V r1 V / ||V||. .

_x2 ps_1_4 . , . dp3_sat r1, r0_bx2, r0_bx2

mad r1, r0_bias, 1-r1, r0_bx2

r0 (V + 1) . (V + 1) V [ -1, 1 ] [ 0, 1 ], .

GeForce 6 GPU fp16 1/2 . fp16 . nrmh .

Normalization Heuristics and Bump Map Compression . .

http://developer.nvidia.com/object/normalization_heuristics.html

http://developer.nvidia.com/object/bump_map_compression.html

3.5.3.3. sincos()

, GeForce FX GPU . sincos . .

3.6.

3.6.1. 2 Z

GeForce FX GeForce 6 GPU 2 . .

.

http://developer.nvidia.com/object/normalization_heuristics.htmlhttp://developer.nvidia.com/object/bump_map_compression.html

GPU

- .

Texkill .

(oDepth, texm3x2depth, texdepth) .

.

.

.

3.6.2.

3.6.3.

Z

Z ( z-cull ) . , z-cull . z-cull .

( , texkill ).

( , GPU ).

GPU z-cull .

2 . , 2 ( ) . . , . z-cull . .

32

NVIDIA GPU

33

3.6.4.

3.7.

, . 2 . z-cull .

Deferred Shading . NVSDK 7.1 .

(thrashing) , .

1.

(width * bpp) . . .

2.

3.

(Antialiasing)

GeForce FX GeForce 6 . . .

. .

DirectX9.0b . StretchRect() .

GPU

, 100 100 4x , 200 200 . 100 100 , StretchRect() , GPU .

, 100 100 .

(200 200) (100 100) .

34

Chapter 4. GeForce 6

GeForce 6 Quadro FX 4xxx GPU . , .

4.1. Shader Model 3.0

Microsoft DirectX 9.0 ( 2.0 3.0) . Shader Model 2.0 2002 , GPU Shader Model 2.0 . Shader Model 2.0 , . .

Pixel Shader 2.0 Vertex Shader 2.0 Shader Model 3.0 . , .

GeForce 6

4.1.1. Pixel Shader 3.0 Pixel Shader 2.0 3.0 .

Shader 2.0

Shader 3.0

96 65535+ , ,

, .

8

32

4 ,

8

fp16-fp32

Shader Model 3.0 /

8 10

36

NVIDIA GPU

37

4.1.2. Vertex Shader 3.0 Vertex Shader Model 2.0 3.0 .

Shader 2.0

Shader 3.0

256

65535

4

,

4.1.3. (Dynamic Branching)

Shader 3.0 ( ) (Dynamic Branching). , . , , . . , .

. , ( ) . . .

GeForce 6

4.1.4.

4.1.5.

Pixel Shader 2.0 . , , . Shader Model 3.0 , / . .

(Instancing)

Shader Model 3.0 Microsoft DirectX Instancing API . , . , CPU . , , . .

, , . , 1000 , 200 , , . . 200 , .

http://download.nvidia.com/developer/SDK/Individual_Samples/samples.html .

38

http://download.nvidia.com/developer/SDK/Individual_Samples/samples.html

NVIDIA GPU

39

4.1.6.

4.2.

, DirectX 9.0 Shader Model 3.0 , , . , . , CPU .

sRGB

sRGB . DXT , .

GeForce 6 GPU , . .

sRGB . sRGB .

4.3.

GeForce 6 GPU . , . .

GeForce 6

4.4.

GeForce 6 GPU .

2D 3D MIP sRGB R8G8B8 N N N N N N N N N A8R8G8B8 Y Y Y Y Y Y Y Y N X8R8G8B8 Y Y Y Y Y Y Y Y N R5G6B5 Y Y Y Y Y Y Y Y N X1R5G5B5 Y Y Y Y Y Y Y Y N A1R5G5B5 Y Y Y Y Y Y N N N A4R4G4B4 Y Y Y Y Y Y N N N R3G3B2 N N N N N N N N N A8 Y Y Y Y Y N/A N N N A8R3G3B2 N N N N N N N N N X4R4G4B4 N N N N N N N N N A2B10G10R10 N N N N N N N N N A8B8G8R8 N N N N N N N N N X8B8G8R8 N N N N N N N N N G16R16 Y Y Y Y Y N N N N A2R10G10B10 N N N N N N N N N A16B16G16R16 N N N N N N N N N A8P8 N N N N N N N N N P8 N N N N N N N N N L8 Y Y Y Y Y N N N N L16 Y Y Y Y Y N N N N A8L8 Y Y Y Y Y N N N N A4L4 N N N N N N N N N float 2D 3D MIP sRGB R16F N N N N N N/A N N N G16R16F Y Y Y Y Y N/A Y N N A16B16G16R16F Y Y Y Y Y N/A Y Y N R32F Y Y Y Y N N/A Y N Y G32R32F N N N N N N/A N N N

40

NVIDIA GPU

41

A32B32G32R32F Y Y Y Y N N/A Y N Y 2D 3D MIP sRGB D24X8 Y N N Y Y N/A Y N/A N D24S8 Y N N Y Y N/A Y N/A N D16 Y N N Y Y N/A Y N/A N

4.5.

GeForce 6 GPU . 16 (fp16) 32 (fp32) .

/

3D

Non-power-of-2 Textures

16

32

4.5.1.

4.6.

R16F . G16R16F . G16R16F R23F A16B16G16R16F . G16R16F .

MRT(Multiple Render Target)

GeForce 6 GPU MRT . 4 . 4 float MRT .

GeForce 6

MRT GPGPU . (deferred shading) float4 . , , , . .

DirectX caps bit NumSimultaneousRTs . GeForce 6 GPU caps bit 4 . MRT SetRenderTarget(index, pRenderTarget) API . . oC0, oC1, oC2, oC3 . MRT 0 3 NULL .

MRT GPU . , MRT . , , , . GeForce 6 , , 32 64 (, A16R16G16B16F ) 4 , 128 (, A32R32G32B32F ) 4 .

, D3DMISCCAPS_MRTPOSTPIXELSHADERBLENDING caps bit USAGE_QUERY_POSTPIXELSHADERBLENDING , , , , MRT .

GeForce 6 MRT . , R32F, G16R16F, and A32R32G32B32F . GeForce 6 GPU(GeForce 6200 ) A16R16G16B16F . DirectX MRT . MRT ( 1 3

42

NVIDIA GPU

43

), oC0.a 4 .

, MRT . MRT . , 4 A32R32G32B32F A8R8G8B 16 ! , GeForce 6 3 .

. , MRT MRT . , . MRT (3.6.4 ).

MRT 3 , . , MRT 4 , MRT 3 . , , (: ). SDK 7.1 ftp://download.nvidia.com/developer/Movies/NV40-LowRes-Clips/Deferred_Shading.avi .

4.7.

GeForce 6 GPU , RAM . . . . .

ftp://download.nvidia.com/developer/Movies/NV40-LowRes-Clips/Deferred_Shading.aviftp://download.nvidia.com/developer/Movies/NV40-LowRes-Clips/Deferred_Shading.avi

GeForce 6

http://developer.nvidia.com/object/using_vertex_textures.html Using Vertex Textures .

4.8.

GeForce 6 . .

. GeForce 6 ( ) 4 . . .

. GeForce 6 GPU 2 . , GeForce 6 fp16 . 16 . , . , .

. 4.1.3 , . , ( , 30 x 30 ).

44

http://developer.nvidia.com/object/using_vertex_textures.html

NVIDIA GPU

45

4.9.

, GeForce 6 CPU .

N.z = sqrt( 1 N.x*N.x N.y*N.y );

3-5 . .

, GeForce 6 GPU DirectX D3DFMT_V8U8 OpenGL GL_LUMINANCE8_ALPHA8 . 16 /, 2:1 .

. (+) . (-) ( ) (http://developer.nvidia.com/object/mipmapping_normal_maps.html ) .

http://developer.nvidia.com/object/bump_map_compression.html Bump Map Compression .

, NVIDIA Melody . Melody . Melody http://developer.nvidia.com/object/melody_home.html .

http://developer.nvidia.com/object/mipmapping_normal_maps.htmlhttp://developer.nvidia.com/object/bump_map_compression.htmlhttp://developer.nvidia.com/object/melody_home.html

Chapter 5. GeForce FX

GeForce FX . , .

5.1.

GeForce FX GeForce FX 5900 2 . . , .

. API .

5.2.

GeForce FX Direct3D 512 OpenGL 1024 . DirectX , ps_2_a ps_2_x .

GeForce FX

OpenGL , GLSL, Cg(arbfp1 fp30 ) ARB_fragment_program .

Quadro FX 2048 .

5.3. DirectX

DirectX 9 ps_2_0 24 ( GeForce FX 32 float ). 16 a _pp .

HLSL Cg . 32 float 16 half .

1/2 .

ps_1_0 ps_1_4 .

DirectX ps_2_0 . ps_1_1 ps_1_4 , .

5.4. OpenGL

ARB_fragment_program 24 . ARB_fragment_program .

NV_fragment_program. half(16 ) fixed(12 ) .

48

NVIDIA GPU

49

5.5.

ARB_precision_hint_fastest. Unified Compiler , .

ARB_precision_hint_nicest. float .

16

half , float . half 10 () 5 . float 23 8 . V . . half , 0.1%. . , V . .

half , -2048 2048 , . , . , 4096 4097 , 16 . . INF . . .

GeForce FX GPU half . half .

float . . , . ,

GeForce FX

GPU ?

. . . ( .)

. . , .

CPU , , .

5.6.

2D 3D MIP sRGB R8G8B8 N N N N N N N N N A8R8G8B8 Y Y Y Y Y Y Y Y N X8R8G8B8 Y Y Y Y Y Y Y Y N R5G6B5 Y Y Y Y Y Y Y Y N X1R5G5B5 Y Y Y Y Y Y Y Y N A1R5G5B5 Y Y Y Y Y Y N N N A4R4G4B4 Y Y Y Y Y Y N N N R3G3B2 N N N N N N N N N A8 N N N N N N/A N N N A8R3G3B2 N N N N N N N N N X4R4G4B4 N N N N N N N N N A2B10G10R10 N N N N N N N N N A8B8G8R8 N N N N N N N N N X8B8G8R8 N N N N N N N N N

50

NVIDIA GPU

51

G16R16 Y Y Y Y Y N N N N A2R10G10B10 N N N N N N N N N A16B16G16R16 N N N N N N N N N A8P8 N N N N N N N N N P8 N N N N N N N N N L8 Y Y Y Y Y N N N N L16 Y Y Y Y Y N N N N A8L8 Y Y Y Y Y N N N N A4L4 N N N N N N N N N float 2D 3D MIP sRGB R16F N N N N N N/A N N N G16R16F Y Y Y Y N N/A Y N N A16B16G16R16F* Y N N N N N/A Y N N R32F Y Y Y Y N N/A Y N N G32R32F N N N N N N/A N N N A32B32G32R32F* Y N N N N N/A Y N N 2D 3D MIP sRGB D24X8 Y N N Y Y N/A Y N/A N D24S8 Y N N Y Y N/A Y N/A N D16 Y N N Y Y N/A Y N/A N * (wrapping) DX9.0c

5.7. DirectX ps_2_x ps_2_a

GeForce FX ps_2_0 (DDX DDY ), , . 2 . ps_2_x HLSL Cg . . ps_2_a . GeForce FX .

GeForce FX

5.8.

GeForce FX 64 128 4 1 2 32 .

OpenGL , NV_float_buffer , NV_fragment_program / . RGBA8 / NEAREST_MIPMAP_NEAREST ( fp32 2 fp16 ).

GeForce FX , (texel) 32 , .

5.9.

GeForce FX, GeForce4, GeForce3 GPU 2 . GPU , CxV8U8 . Z .

http://developer.nvidia.com/object/bump_map_compression.html Bump Map Compression .

5.10.

GeForce FX .

52

http://developer.nvidia.com/object/bump_map_compression.html

NVIDIA GPU

53

5.11.

. . LOD GeForce FX .

GeForce 6 GPU float , half . GPU float/half , , .

GeForce FX GeForce 6 . GeForce FX , .

, GeForce FX ps_1_* ps_2_* . , ps_1_1 , ps_1_4 ps_2_0 .

Chapter 6.

GPU GPU .

6.1. GPU

GPU (Window ) GPU ID . ID . , GeForce 6 GPU . GPU caps bit(DirectX ) (OpenGL ) . OpenGL , NV40 FX (GeForce 6xxx Quadro FX 3400 ).

ID . ID . GPU , .

ID . ID .

GeForce FX

ID ID . .

GeForce 6 GPU . GPU TNT GPU ID . NV4X , .

ID cap . cap . , cap , .

. .

ID , [email protected] Developer Relations .

NVIDIA GPU ID http://developer.nvidia.com/object/device_ids.html .

6.2.

GeForce 3 GPU NVIDIA OpenGL DirectX . (hardware shadow mapping) % . . . , ps_2_0 .

.

56

mailto:[email protected]://developer.nvidia.com/object/device_ids.html

NVIDIA GPU

57

http://developer.nvidia.com/object/hwshadowmap_paper.html

http://developer.nvidia.com/object/cedec_shadowmap.html

http://developer.nvidia.com/object/d3dshadowmap.html

http://developer.nvidia.com/object/Shadow_Map.html

Perspective Shadow Maps from SIGGRAPH 2002

Perspective Shadow Maps: Care and Feeding in GPU Gems: Programming Techniques, Tips, and Tricks for Real-Time Graphics (http://developer.nvidia.com/GPUGems)

Simon Kozlov GPU Gems Perspective Shadow Maps: Care and Feeding . Kozlov , . SKD 7.0 . http://download.nvidia.com/developer/SDK/Individual_Samples/samples.html .

Direct X , .

1) D3DUSAGE_DEPTHSTENCIL .

2) D3DFMT_D16, D3DFMT_D24X8 ( D3DFMT_D24S8 , ) .

3) IDirectDrawSurface9 Interface GetSurfaceLevel(0) .

4) SetDepthStencilSurface() Z .

5) DirectX , , D3DRS_COLORWRITEENABLE .

6) z .

7) .

8) z .

9) .

http://developer.nvidia.com/object/hwshadowmap_paper.htmlhttp://developer.nvidia.com/object/cedec_shadowmap.htmlhttp://developer.nvidia.com/object/d3dshadowmap.htmlhttp://developer.nvidia.com/object/Shadow_Map.htmlhttp://developer.nvidia.com/GPUGemshttp://download.nvidia.com/developer/SDK/Individual_Samples/samples.html

GeForce FX

V = Bias(0.5/TexWidth, 0.5/TexHeight, 0) *

Bias(0.5, 0.5, 0) *

Scale(0.5,0.5,1) *

ViewProjsaved * World * Object * V

CPU , .

10) ps_1.0-1.3 , D3DTTFF_COUNT4 | D3DTTFF_PROJECTED .

11) . 1.4 , .

12) x y .

13) z . , 0 ( ). 1 .

14) , 4 . .

15) .

NVIDIA ( 45.23 ) . NVDIA 52.16 . . , ps.2.0 , ( , tex2Dproj(ShadowMapSampler, IN.TexCoord0).rgb ) . w- ! , tex2D(ShadowMapSampler, IN.TexCoord0/IN.TexCoord0.w) .

58

NVIDIA GPU

59

, ps1.1-1.3 52.16 ( , SetTextureStageState(0, D3DTSS_TEXTURETRANSFORMFLAGS, D3DTTFF_PROJECTED) .

: ForceWare 61.71 (tex2D, tex2Dlod ) .

SDK DirectX OpenGL . http://download.nvidia.com/developer/SDK/Individual_Samples/samples.html .

http://download.nvidia.com/developer/SDK/Individual_Samples/samples.html

Chapter 7. NVIDIA SLI GPU

NVIDIA SLI GPU . ftp://download.nvidia.com/developer/presentations/2004/GPU_Jackpot/SLI_and_Stereo.pdf .

7.1. SLI ?

Scalable Link Interface SLI GPU . SLI x16 2 PCI Express . PCI Express . SLI , SLI -GPU . SLI -GPU 2 . 2 GPU .

ftp://download.nvidia.com/developer/presentations/2004/GPU_Jackpot/SLI_and_Stereo.pdfftp://download.nvidia.com/developer/presentations/2004/GPU_Jackpot/SLI_and_Stereo.pdf

NVIDIA SLI Multi-GPU

GPU 1.9 . 2 GPU . SLI 2 . , 256MB 256MB . GPU . , GPU 0 GPU 1 .

SLI , , (AFR) SFR .

GPU (, GPU ). , SLI .

AFR , GPU 0 n GPU 1 n+1 . n+2 GPU 0 . (, ) , AFR . , , GPU . ( , ), GPU . 2 .

SFR , GPU 0 , GPU 1 . GPU 1 , . GPU 2 . GPU 0 GPU 1 GPU .

62

NVIDIA GPU

63

7.2.

SFR . AFR . 2 , AFR .

(66.93 ) . , SFR GeForce . -GPU .

API SLI .

SLI .

CPU

CPU , . -GPU CPU . 2 CPU .

, .

7.3. VSync ( )

VSync . -GPU () . VSync .


. .

, 3 . , . .

2 .

1. ( , ).

2. ( ) .

VSync . DirectX IDirect3D9::CreateDevice() D3DPRESENT_PARAMETERS PresentationInterval D3DPRESENT_INTERVAL_IMMEDIATE .

7.4. 2 (Lag)

DirectX 3 . CPU GPU . , . . ( ) 30ms .

. , , CPU GPU . CPU

64

NVIDIA GPU

65

7.5.

GPU . .

, -GPU .

(: DirectX ) . , 1~3 .

-GPU . , n GPU n .

. GPU GPU 2 . , GPU 2 ( 15ms ), GPU 1 ( 30ms ) 30ms .

GPU GPU . nvCPL.dll API SLI , GPU . , NvCplGetDataInt(NVCPL_API_NUMBER_OF_SLI_GPUS, &number) SLI GPU . NVControlPanel_API.pdf NVSDK NvCpl .

-CPU GPU . , GPU , , . ,


. , . , .

GPU , -GPU . , GPU , ( ) 2 2 .

, SLI . , 1 GPU GPU .

7.6.

, , -GPU (0 ) GPU . , .

7.7. D3DPOOL_MANAGED

0 , -GPU GPU . D3DPOOL_MANAGED DirectX , D3DPOOL_DEFAULT . , .

66

Chapter 8.

NVIDIA 100% .

8.1. ?

. . ( ) , NVIDIA .

, . . .

, . .

8.2.

NVIDIA 3D Stereo Driver DirectX OpenGL . 3D Stereo Driver 2 . , . Stereo Driver Display Driver .

3D , 2 . , . .

, .

8.3.

.

8.3.1.

. 3D Stereo Driver , .

, , . , .

HUD 3D . 3D , (view frustrum) .

68

NVIDIA GPU

69

8.3.2.

8.3.3.

HUD . , HUD .

3D , , , . 3D , . 2 , 2 .

.

3D . . 3D , 3D Stereo . . . .

(, , ) , . 3D .

2D . , , . 3D , 2D . . 3D Stereo .

.

8.3.4.

8.3.5.

8.3.6.

8.3.7.

8.3.8.

3D 2D

2D 3D . . 3D Stereo . 2D 3D HUD , . 3D .

, , , . , .

(Dirty Rectangle)

, 3D . .

. 3D .

. .

70

NVIDIA GPU

71

8.3.9.

8.3.10.

8.3.11.

8.3.12.

8.3.13.

8.3.14.

D3D , RHW .

NVIDIA 3D Stereo . , 3D Stereo .

. . , .

NVIDIA 3D Stereo DirectX OpenGL . API 3D , .

. .

3D Stereo . . , . .

8.3.15.

8.4.

. .

.

8.4.1.

8.4.2.

8.4.3.

3D Stereo . , . IODisplay(http://www.i-glasses.com) . NVIDIA 3D Stereo Driver 2 . 3D , .

Out of the Monitor

. . , , . .

3D . , . , , , !

72

NVIDIA GPU

73

8.4.4.

8.4.5.

8.5.

, . , 1 , 3 , .

3D Stereo Driver . . NVDIA Developer Relations .

API

2 API .

StereoBLT API 3D , .

IStereoAPI , , . API . .

OpenGL . NVIDIA Quadro GPU , .

8.6.

NVIDIA Developer Relations [email protected] Stereo API .

http://developer.nvidia.com/object/3D_Stereoscopic_Dev.html .

74

mailto:[email protected]://developer.nvidia.com/object/3D_Stereoscopic_Dev.html

Chapter 9.

.

9.1. NVPerfHUD NVPerfHUD DirectX 4 . , . . 256 .

NVPerfHUD NVIDIA Developer http://developer.nvidia.com/object/nvperfhud_home.html .

http://developer.nvidia.com/object/nvperfhud_home.html

GPU

9.2. NVShaderPerf NVShaderPerf FX Composer Shader Perf . HLSL, GLSL, Cg, !!FP1.0, !!ARBfp1.0, ps_1_x, ps_2_x. DirectX OpenGL . , , GPU GeForce 6 GeForce FX GPU .

NVShaderPerfhttp://developer.nvidia.com/object/nvshaderperf_home.html .

9.3. FX Composer FX Composer . FX Composer , GUI .

FX Composer .

.

76

http://developer.nvidia.com/object/nvshaderperf_home.html

NVIDIA GPU

GeForce 6 GeForce FX GPU .

.

GPU , , , FPS .

.

FX Composer http://developer.nvidia.com/fxcomposer .

9.4. NVIDIA Melody , NVIDIA Melody . Melody . Melody http://developer.nvidia.com/object/melody_home.html

9.5.

. [email protected] .

77

http://developer.nvidia.com/fxcomposerhttp://developer.nvidia.com/object/melody_home.htmlmailto:[email protected]

NVPerfHUD

: CPU: GPU

GPU

Indexed Primitive Call

ps_2_a Interpolant mul() saturate() D3DTADDRESS_CLAMP ( GL_CLAMP_TO_E Interpolant

2D 1D ARGB 3D

sincos()

2 Z Z

(Antialiasing)

GeForce 6 Shader Model 3.0 Pixel Shader 3.0Vertex Shader 3.0 (Dynamic Branching) (Instancing)

sRGB

MRT(Multiple Render Target)

GeForce FX DirectX OpenGL 16 DirectX ps_2_x ps_2_a

GPU

NVIDIA SLI GPU SLI ?CPU VSync ( ) 2 (Lag) D3DPOOL_MANAGED

? 3D 2D (Dirty Rectangle)

Out of the Monitor

API

NVPerfHUDNVShaderPerfFX ComposerNVIDIA Melody

version 2.2.0 nvidia gpu

Documents