shadermark 2.1 released

tEd · Sep 11, 2004

http://www.shadermark.com

hovz · Sep 11, 2004

benchmarks asap

trinibwoy · Sep 11, 2004

Forceware 65.76 Shadermark 2.1

(1024x768) X8R8G8B8 (D24X8) vram used 131072000
HAL (pure hw vp): NVIDIA GeForce 6800 (Anti-Detect-Mode: off, gamma correction: DAC)

Shader PS2.0 PS3.0

2 511 511
3 425 425
4 412 421
5 340 340
6 392 391
7 235 236
8 170 177
9 322 322
10 271 268
11 388 388
12 275 274
13 224 170
14 255 255
15 160 187
16 123 123
17 187 187
18 32 32
19 84 92
20 29 29
21 47 47
22 66 68
23 73 75
24 48 48
25 53 49
26 52 48

Any ideas on why some of my results suck compared to Dave's 6600GT review?

hovz · Sep 11, 2004

can people with xtpes and ultras post results. i wanna see how 2.0b and 3.0 compare

hovz · Sep 11, 2004

m4trix posted these results with a pro at stock speeds on rage3d with 2.0b

2 764
3 552
4 552
5 454
6 551
7 507
8 317
9 1050
10 683
11 534
12 180
13 265
14 293
15 215
16 260
17 331
18 31
19 103
20 34
21 not supported
22 160
23 not supported
24 112
25 89
26 89

hovz · Sep 11, 2004

trinibwoy said:
Forceware 65.76 Shadermark 2.1

(1024x768) X8R8G8B8 (D24X8) vram used 131072000
HAL (pure hw vp): NVIDIA GeForce 6800 (Anti-Detect-Mode: off, gamma correction: DAC)

Shader PS2.0 PS3.0

2 511 511
3 425 425
4 412 421
5 340 340
6 392 391
7 235 236
8 170 177
9 322 322
10 271 268
11 388 388
12 275 274
13 224 170
14 255 255
15 160 187
16 123 123
17 187 187
18 32 32
19 84 92
20 29 29
21 47 47
22 66 68
23 73 75
24 48 48
25 53 49
26 52 48

Any ideas on why some of my results suck compared to Dave's 6600GT review?

on the shadermark page is says u have to use the 66.xx drivers for nvidia cards and 4,9 for ati

ERP · Sep 11, 2004

6800GT Stock

ShaderMark v2.1 - DirectX 9 HLSL Pixel Shader Benchmark - ToMMTi-Systems (http://www.tommti-systems.com)

video mode / device info
(1024x768) X8R8G8B8 (D24S8) vram used 123731968
HAL (pure hw vp): NVIDIA GeForce 6800 GT (Anti-Detect-Mode: off, gamma correction: DAC)

options

pixel shader version: Pixel Shader 2_a

results:
shader 1 ( Pixel Shader Precision Test): pixel shader 1.1 full precision: s10e8
( Pixel Shader Precision Test): pixel shader 1.4 full precision: s10e8
( Pixel Shader Precision Test): pixel shader 2.0 full precision: s23e8
( Pixel Shader Precision Test): pixel shader 2.0 partial precision: s10e5
( Pixel Shader Precision Test): pixel shader 3.0 full precision: s23e8
( Pixel Shader Precision Test): pixel shader 3.0 partial precision: s10e5

shader 2 ( Per Pixel Diffuse Lighting - ps_2_a): 780 fps 1.2813 mspf 3903 rendered frames
shader 3 ( Per Pixel Directional Light Shader (Phong) - ps_2_a): 636 fps 1.5716 mspf 3182 rendered frames
shader 4 ( Per Pixel Point Light Shader (Phong) - ps_2_a): 640 fps 1.5634 mspf 3199 rendered frames
shader 5 ( Per Pixel Spot Light Shader (Phong) - ps_2_a): 585 fps 1.7105 mspf 2924 rendered frames
shader 6 ( Per Pixel Anisotropic Lighting - ps_2_a): 600 fps 1.6673 mspf 3000 rendered frames
shader 7 ( Per Pixel Fresnel Reflections - ps_2_a): 551 fps 1.8164 mspf 2753 rendered frames
shader 8 ( Per Pixel Car Surface Shader - ps_2_a): 358 fps 2.7931 mspf 1791 rendered frames
shader 9 ( Per Pixel Environment Mapping - ps_2_a): 893 fps 1.1197 mspf 4466 rendered frames
shader 10 ( Per Pixel Environment Bump Mapping - ps_2_a): 720 fps 1.3897 mspf 3599 rendered frames
shader 11 ( Per Pixel Bump Mapping - ps_2_a): 600 fps 1.6670 mspf 3000 rendered frames
shader 12 ( Per Pixel Shadowed Bump Mapping - ps_2_a): 400 fps 2.4997 mspf 2001 rendered frames
shader 13 ( Per Pixel Veined Marble Shader - ps_2_a): 352 fps 2.8375 mspf 1763 rendered frames
shader 14 ( Per Pixel Wood Shader - ps_2_a): 398 fps 2.5125 mspf 1991 rendered frames
shader 15 ( Per Pixel Tile Shader - ps_2_a): 280 fps 3.5696 mspf 1401 rendered frames
shader 16 ( Per Pixel Refraction and Reflection Shader with Phong Lighting - ps_2_a): 279 fps 3.5794 mspf 1397 rendered frames
shader 17 ( Per Pixel BRDF-Phong/Anisotropic Lighting - ps_2_a): 383 fps 2.6076 mspf 1918 rendered frames

performance impact of heavy alpha blending

shader 18 ( Fur Shader With Anisotropic Lighting - ps_2_a): 47 fps 21.1115 mspf 237 rendered frames

performance impact of multiple shaders switches

shader 19 ( Combination Effect - ps_2_a): 155 fps 6.4330 mspf 778 rendered frames

performance impact of dynamic flow control

shader 20 ( Dual Layer 8x8 PCF Shadow Mapping without Flow Control - ps_2_a): 46 fps 21.6443 mspf 232 rendered frames
shader 21 ( Dual Layer 8x8 PCF Shadow Mapping with Flow Control - ps_3_0): 73 fps 13.7537 mspf 364 rendered frames

performance impact of floating point filtering

shader 22 ( High Dynamic Range Shader - low quality version without filtering - ps_2_a): 105 fps 9.5597 mspf 524 rendered frames
shader 23 ( High Dynamic Range Shader - high quality with fp filtering - ps_2_a): 116 fps 8.6092 mspf 581 rendered frames
shader 24 ( High Dynamic Range Shader - high quality without fp filtering - ps_2_a): 70 fps 14.2267 mspf 352 rendered frames

performance impact of multiple render targets

shader 25 ( Per Pixel Edge Detection And Hatching Shader using 1 RT and 2 Passes - ps_2_a): 80 fps 12.5195 mspf 400 rendered frames
shader 26 ( Per Pixel Edge Detection And Hatching Shader using 2 RTs and 1 Pass - ps_2_a): 76 fps 13.1605 mspf 380 rendered frames

ERP · Sep 11, 2004

FWIW the generally fastest path on my 6800 is the 2_a path.
3.0 is generally within a few frames either way with the obvious exception of the dynamic branch version of the shadow shader which is much faster.

forcing fp16 makes very little difference, forcing fp32 does slow things on some of the tests, I'm assuming from the tests affected that it has to do with the free 16 bit normalize.

hovz · Sep 11, 2004

ERP said:
6800GT Stock

ShaderMark v2.1 - DirectX 9 HLSL Pixel Shader Benchmark - ToMMTi-Systems (http://www.tommti-systems.com)

video mode / device info
(1024x768) X8R8G8B8 (D24S8) vram used 123731968
HAL (pure hw vp): NVIDIA GeForce 6800 GT (Anti-Detect-Mode: off, gamma correction: DAC)

options

pixel shader version: Pixel Shader 2_a

results:
shader 1 ( Pixel Shader Precision Test): pixel shader 1.1 full precision: s10e8
( Pixel Shader Precision Test): pixel shader 1.4 full precision: s10e8
( Pixel Shader Precision Test): pixel shader 2.0 full precision: s23e8
( Pixel Shader Precision Test): pixel shader 2.0 partial precision: s10e5
( Pixel Shader Precision Test): pixel shader 3.0 full precision: s23e8
( Pixel Shader Precision Test): pixel shader 3.0 partial precision: s10e5

shader 2 ( Per Pixel Diffuse Lighting - ps_2_a): 780 fps 1.2813 mspf 3903 rendered frames
shader 3 ( Per Pixel Directional Light Shader (Phong) - ps_2_a): 636 fps 1.5716 mspf 3182 rendered frames
shader 4 ( Per Pixel Point Light Shader (Phong) - ps_2_a): 640 fps 1.5634 mspf 3199 rendered frames
shader 5 ( Per Pixel Spot Light Shader (Phong) - ps_2_a): 585 fps 1.7105 mspf 2924 rendered frames
shader 6 ( Per Pixel Anisotropic Lighting - ps_2_a): 600 fps 1.6673 mspf 3000 rendered frames
shader 7 ( Per Pixel Fresnel Reflections - ps_2_a): 551 fps 1.8164 mspf 2753 rendered frames
shader 8 ( Per Pixel Car Surface Shader - ps_2_a): 358 fps 2.7931 mspf 1791 rendered frames
shader 9 ( Per Pixel Environment Mapping - ps_2_a): 893 fps 1.1197 mspf 4466 rendered frames
shader 10 ( Per Pixel Environment Bump Mapping - ps_2_a): 720 fps 1.3897 mspf 3599 rendered frames
shader 11 ( Per Pixel Bump Mapping - ps_2_a): 600 fps 1.6670 mspf 3000 rendered frames
shader 12 ( Per Pixel Shadowed Bump Mapping - ps_2_a): 400 fps 2.4997 mspf 2001 rendered frames
shader 13 ( Per Pixel Veined Marble Shader - ps_2_a): 352 fps 2.8375 mspf 1763 rendered frames
shader 14 ( Per Pixel Wood Shader - ps_2_a): 398 fps 2.5125 mspf 1991 rendered frames
shader 15 ( Per Pixel Tile Shader - ps_2_a): 280 fps 3.5696 mspf 1401 rendered frames
shader 16 ( Per Pixel Refraction and Reflection Shader with Phong Lighting - ps_2_a): 279 fps 3.5794 mspf 1397 rendered frames
shader 17 ( Per Pixel BRDF-Phong/Anisotropic Lighting - ps_2_a): 383 fps 2.6076 mspf 1918 rendered frames

performance impact of heavy alpha blending

shader 18 ( Fur Shader With Anisotropic Lighting - ps_2_a): 47 fps 21.1115 mspf 237 rendered frames

performance impact of multiple shaders switches

shader 19 ( Combination Effect - ps_2_a): 155 fps 6.4330 mspf 778 rendered frames

performance impact of dynamic flow control

shader 20 ( Dual Layer 8x8 PCF Shadow Mapping without Flow Control - ps_2_a): 46 fps 21.6443 mspf 232 rendered frames
shader 21 ( Dual Layer 8x8 PCF Shadow Mapping with Flow Control - ps_3_0): 73 fps 13.7537 mspf 364 rendered frames

performance impact of floating point filtering

shader 22 ( High Dynamic Range Shader - low quality version without filtering - ps_2_a): 105 fps 9.5597 mspf 524 rendered frames
shader 23 ( High Dynamic Range Shader - high quality with fp filtering - ps_2_a): 116 fps 8.6092 mspf 581 rendered frames
shader 24 ( High Dynamic Range Shader - high quality without fp filtering - ps_2_a): 70 fps 14.2267 mspf 352 rendered frames

performance impact of multiple render targets

shader 25 ( Per Pixel Edge Detection And Hatching Shader using 1 RT and 2 Passes - ps_2_a): 80 fps 12.5195 mspf 400 rendered frames
shader 26 ( Per Pixel Edge Detection And Hatching Shader using 2 RTs and 1 Pass - ps_2_a): 76 fps 13.1605 mspf 380 rendered frames

2 780
3 636
4 640
5 585
6 600
7 551
8 358
9 893
10 720
11 600
12 400
13 352
14 398
15 280
16 279
17 383
18 47
19 155
20 46
21 73
22 105
23 116
24 70
25 80
26 76

for easier comparison

DSN2K · Sep 11, 2004

Code:

video mode / device info
(1024x768) X8R8G8B8 (D24X8) vram used 132120576
HAL (pure hw vp): NVIDIA GeForce 6800 (Anti-Detect-Mode: off, gamma correction: DAC)  Driver:66.31

 PS3.0  PS2.0a

2  636     725
3  510     588
4  510     587
5  475     510
6  503     552
7  305     510
8  221     330
9  480     846
10 329     660
11 510     567
12 365     394
13 298     340
14 340     361
15 250     256
16 149     255
17 250     339
18 46       48
19 126     148
20 39       41
21 61       64
22 90       98
23 100     111
24 63       68
25 70       72
26 68       70

Pete · Sep 11, 2004

For easy reference:

Code:

Card       X800P    6800GT    6800
Owner      m4trix   ERP       DSN2K
Drivers    4.9?     66.xx?    66.xx?
Test/SM    2_b      2_a       3_0
2           764      780       636
3           552      636       510
4           552      640       510
5           454      585       475
6           551      600       503
7           507      551       305
8           317      358       221
9          1050      893       480
10          683      720       329
11          534      600       510
12          180      400       365
13          265      352       298
14          293      398       340
15          215      280       250
16          260      279       149
17          331      383       250
18           31       47        46
19          103      155       126
20           34       46        39
21          n/a       73        61
22          160      105        90
23          n/a      116       100
24          112       70        63
25           89       80        70
26           89       76        68

If you want to keep things neat, quote my reply and edit it to include your numbers. It'll be easier to compare horizontally, rather than scrolling back and forth among one column x 30 row posts.

Heh, hovz, someone had to do get the ball rolling the right way. As I don't have a card to contribute numbers, I figured this'd help.

BTW, is test #12 ("per pixel shadowed bump mapping") the same as virtual displacement mapping?

hovz · Sep 11, 2004

thx pete

hovz · Sep 11, 2004

Re: For easy reference:

Code:

Card       X800P    6800GT    6800    6800U OC    xtpe
Owner      m4trix   ERP       DSN2K   jadedragon  thatdude
Drivers    4.9?     66.xx?    66.xx?  ????        4.9
Test/SM    2_b      2_a       3_0     3_0         2_b
2           764      780       636     960         984
3           552      636       510     819         725
4           552      640       510     820         723
5           454      585       475     648         600
6           551      600       503     719         724
7           507      551       305     720         669
8           317      358       221     420         376
9          1050      893       480     1018        1248
10          683      720       329     891         820
11          534      600       510     720         694
12          180      400       365     535         251
13          265      352       298     330         335
14          293      398       340     472         402
15          215      280       250     320         276
16          260      279       149     353         346
17          331      383       250     420         444
18           31       47        46     61          40
19          103      155       126     178         137
20           34       46        39     51          46
21          n/a       73        61     83          n/a
22          160      105        90     116         199
23          n/a      116       100     131         n/a
24          112       70        63     77          140
25           89       80        70     81          117
26           89       76        68     78          115

thatdude90210 · Sep 12, 2004

My modded x800 vivo running at 16 pipes and xtpe speed:

Code:

driver 4.9
pixel shader version: Pixel Shader 2_b

2    984
3    725
4    723
5    600
6    724
7    669
8    376
9   1248
10   820
11   694
12   251
13   335
14   402
15   276
16   346
17   444
18    40
19   137
20    46
21   n/a
22   199
23   n/a
24   140
25   117
26   115

hovz · Sep 12, 2004

so the ultra is now faster then the xtpe in shaders

whql · Sep 12, 2004

I only see and "Ultra OC", which is 425MHz. No Actual PE's run yet either.

hovz · Sep 12, 2004

yes, check my updated chart. all the way to the right. also someoen on rage 3d posted their xt pe results and its almost identical

edit : removing the extra 25 mhz wouldnt change the results enough to not favor the ultra

thatdude90210 · Sep 12, 2004

whql said:
I only see and "Ultra OC", which is 425MHz. No Actual PE's run yet either.

Someone posted some results from an actual PE at rage 3d, and it's a touch faster all around than my modded vivo. Might be because I'm running an early bios with what was supposed to have been more "relaxed" memory timing.

hovz · Sep 12, 2004

thatdude90210 said:
whql said:

I only see and "Ultra OC", which is 425MHz. No Actual PE's run yet either.

Click to expand...

Someone posted some results from an actual PE at rage 3d, and it's a touch faster all around than my modded vivo. Might be because I'm running an early bios with what was supposed to have been more "relaxed" memory timing.

still slower than the ultra tho

Broken Hope · Sep 12, 2004

Results for a X800 XT not the PE

943 fps
704 fps
702 fps
583 fps
702 fps
642 fps
342 fps
1176 fps
753 fps
654 fps
229 fps
303 fps
368 fps
274 fps
317 fps
411 fps
37 fps
130 fps
44 fps
195 fps
137 fps
114 fps
111 fps

shadermark 2.1 released

tEd

Casual Member

hovz

trinibwoy

Meh

hovz

hovz

hovz

ERP

ERP

hovz

DSN2K

Pete

Moderate Nuisance

hovz

hovz

thatdude90210

hovz

whql

hovz

thatdude90210

hovz

Broken Hope

Similar threads