'New' Benchmark

S

Sunday

Newcomer
Hi lads!
Did anyone tried new NBENCH benchmark? Is it good?
Is it really that much reliable in GPU benchmarking (insulating the CPU)?
And, is it unbiased in terms of CPU favoring?
Here is PR:

About N-Bench 3.0
N-Bench is a Benchmark program that is capable of CPU performance and displaying 3D graphics. It consists of a "Demo mode" that plays back beautiful 3DCG movies and a "Benchmark mode" that measures CPU performance. Read this file carefully before using N-Bench.
Special Features
• New 3DCG engine based on Microsoft DirectX 9.0 technology.
• All 3DCG, music, and special effects are renewed.
• "Speed Bench", measure the time to complete the task.
• Corresponds to 3DNow! Professional technology and SSE2 instruction of the Intel Pentium 4.
• 4 CPU benchmark test which is not influenced by the graphic performance.
• 4 3D benchmark test which work on XGA.
Working Environment
• Computer equipped with the 1GHz AMD Athlon processor or 1GHz Intel Pentium III processor or greater
• Microsoft Windows XP
• Microsoft DirectX 9.0 or later (required)
• Free 128MB RAM or more (required)
• Color Display with 1024*768 pixels (required)
• 150MB of hard disk space or more (required)
• DirectX 9.0-compatible video card with 3D accelerator function with 32MB VRAM or more (required)
• DirectX 9.0-compatible sound card
Recommended Environment
• AMD Athlon XP 2800+ processor or greater
• 256MB RAM
• Microsoft DirectX 9.0a or later
• nVIDIA "GeForceFX" GPU based video card or ATI RADION 9500, with 32 MB VRAM or more
• Creative Technology Ltd. "Sound Blaster Live!"

http://pc.watch.impress.co.jp/docs/2003/1128/nbench3.lzh
 
Given that AMD is behind this benchmark I would say it's likely biased towards their processors in some way or another. Could still be a useful bench though. Nice to see that this version didn't have so incredibly poorly programmed graphics, unlike the previous two.
 
Humus said:
Given that AMD is behind this benchmark I would say it's likely biased towards their processors in some way or another.
Click to expand...

You guessed it. Lookin at the results in the post above, the 2400+ gets much better scores than a 2.6P4
 
9700 Pro

------------------------------------------------------------------------------
[3D Benchmark Score]
------------------------------------------------------------------------------
<Ninja 1>
N-Bench Score 2947 Marks
FPS Average 40.56 FPS
Total Polygon Count 30088 Counts
Total Texture Size 18.4 MB

<Ninja 2>
N-Bench Score 2882 Marks
FPS Average 50.64 FPS
Total Polygon Count 46254 Counts
Total Texture Size 21.9 MB

<Grand Touring>
N-Bench Score 2698 Marks
FPS Average 40.33 FPS
Total Polygon Count 86274 Counts
Total Texture Size 16.8 MB

<Star Fighter>
N-Bench Score 2564 Marks
FPS Average 125.86 FPS
Total Polygon Count 99502 Counts
Total Texture Size 29.6 MB

<3D Overall>
N-Bench Score 2772 Marks
FPS Average 64.35 FPS

------------------------------------------------------------------------------
[Overall Score]
------------------------------------------------------------------------------
N-Bench Score 2473 Marks
FPS Average 64.35 FPS
 
Well since we are in the 3Dtechnology forum I thought that part was most appropriate ;)

To make Xmas 'merry' :) I ran another with my 9700 Pro overclocked to 370/330.

N-Bench Ver 3.0 Results File

------------------------------------------------------------------------------
[Info]
------------------------------------------------------------------------------
2003/11/28 14:39:41
5800468756

------------------------------------------------------------------------------
[System]
------------------------------------------------------------------------------
Processor Type AMD Athlon(tm) XP 2800+
Processor Speed 2.204 GHz
Free Memory 551016 KB
Total Memory 1048048 KB
Display Adapter RADEON 9700 PRO
Free VRAM 127288 KB
Operating System Windows XP 5.01.2600 Service Pack 1
DirectX Version DirectX 9

------------------------------------------------------------------------------
[Extensions]
------------------------------------------------------------------------------
MMX Yes
SSE Yes
SSE2 No
3DNow! Yes
Enhanced 3DNow! Yes
3DNow! Professional Yes

------------------------------------------------------------------------------
[Options]
------------------------------------------------------------------------------
<Benchmark Options>
Continuous Loop No
BGM Enable No
Protect Mode No
Challenge Mode No
3D Benchmark Tests Yes
CPU Benchmark Tests Yes

<Benchmark Tests>
Integer 1 (Landscape) Yes
Integer 2 (Mandelbrot) Yes
Float 1 (Flock) Yes
Float 2 (Rigid Body) Yes
Ninja 1 Yes
Ninja 2 Yes
Grand Touring Yes
Star Fighter Yes

------------------------------------------------------------------------------
[CPU Benchmark Score]
------------------------------------------------------------------------------
<Integer 1 (Landscape)>
N-Bench Score 2655 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

<Integer 2 (Mandelbrot)>
N-Bench Score 2007 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

<Float 1 (Flock)>
N-Bench Score 2354 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

<Float 2 (Rigid Body)>
N-Bench Score 2286 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

<CPU Overall>
N-Bench Score 2325 Marks
FPS Average n/a

------------------------------------------------------------------------------
[3D Benchmark Score]
------------------------------------------------------------------------------
<Ninja 1>
N-Bench Score 3159 Marks
FPS Average 43.45 FPS
Total Polygon Count 30088 Counts
Total Texture Size 18.4 MB

<Ninja 2>
N-Bench Score 3092 Marks
FPS Average 54.32 FPS
Total Polygon Count 46254 Counts
Total Texture Size 21.9 MB

<Grand Touring>
N-Bench Score 2905 Marks
FPS Average 43.41 FPS
Total Polygon Count 86274 Counts
Total Texture Size 16.8 MB

<Star Fighter>
N-Bench Score 2738 Marks
FPS Average 134.36 FPS
Total Polygon Count 99502 Counts
Total Texture Size 29.6 MB

<3D Overall>
N-Bench Score 2973 Marks
FPS Average 68.88 FPS

------------------------------------------------------------------------------
[Overall Score]
------------------------------------------------------------------------------
N-Bench Score 2649 Marks
FPS Average 68.88 FPS
 
The amount of texture data seems pretty small compared to what most of today's games use. The polycounts don't seem that high, either.
 
DT, the 2x fsaa should make no difference. on my 9600 pro 4xaa+8xaf gives the same results as aa/af disabled. Strange (?)

One other funny thing: after first run, I enabled "challenge mode" in benchmarks settings, and some obvious eyecandy was turned on. Since then it runs with the eyecandy regardless of the challenge mode setting.
BTW, the car looks good enough ;)
 
------------------------------------------------------------------------------
[System]
------------------------------------------------------------------------------
Processor Type AMD Athlon(tm) 64 Processor 3200+
Processor Speed 1.250 GHz :?
Free Memory 128512 KB
Total Memory 523760 KB
Display Adapter RADEON 9700 PRO
Free VRAM 125240 KB
Operating System Windows XP 5.01.2600 Service Pack 1
DirectX Version DirectX 9

------------------------------------------------------------------------------
[Extensions]
------------------------------------------------------------------------------
MMX Yes
SSE Yes
SSE2 Yes
3DNow! Yes
Enhanced 3DNow! Yes
3DNow! Professional Yes

------------------------------------------------------------------------------
[Options]
------------------------------------------------------------------------------
<Benchmark Options>
Continuous Loop No
BGM Enable No
Protect Mode No
Challenge Mode No
3D Benchmark Tests Yes
CPU Benchmark Tests Yes

<Benchmark Tests>
Integer 1 (Landscape) Yes
Integer 2 (Mandelbrot) Yes
Float 1 (Flock) Yes
Float 2 (Rigid Body) Yes
Ninja 1 Yes
Ninja 2 Yes
Grand Touring Yes
Star Fighter Yes

------------------------------------------------------------------------------
[CPU Benchmark Score]
------------------------------------------------------------------------------
<Integer 1 (Landscape)>
N-Bench Score 2608 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

<Integer 2 (Mandelbrot)>
N-Bench Score 2609 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

&lt;Float 1 (Flock)>
N-Bench Score 2054 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

&lt;Float 2 (Rigid Body)>
N-Bench Score 2082 Marks
FPS Average n/a
Total Polygon Count n/a
Total Texture Size n/a

&lt;CPU Overall>
N-Bench Score 2338 Marks
FPS Average n/a

------------------------------------------------------------------------------
[3D Benchmark Score]
------------------------------------------------------------------------------
&lt;Ninja 1>
N-Bench Score 2834 Marks
FPS Average 38.97 FPS
Total Polygon Count 30088 Counts
Total Texture Size 18.4 MB

&lt;Ninja 2>
N-Bench Score 2765 Marks
FPS Average 48.58 FPS
Total Polygon Count 46254 Counts
Total Texture Size 21.9 MB

&lt;Grand Touring>
N-Bench Score 2654 Marks
FPS Average 39.67 FPS
Total Polygon Count 86274 Counts
Total Texture Size 16.8 MB

&lt;Star Fighter>
N-Bench Score 2765 Marks
FPS Average 135.72 FPS
Total Polygon Count 99502 Counts
Total Texture Size 29.6 MB

&lt;3D Overall>
N-Bench Score 2754 Marks
FPS Average 65.73 FPS

------------------------------------------------------------------------------
[Overall Score]
------------------------------------------------------------------------------
N-Bench Score 2546 Marks
FPS Average 65.73 FPS
 
I just read about another benchmark \ techdemo
http://www.openfrag.org/about.php

The link: https://sourceforge.net/project/showfiles.php?group_id=82471

Techdemo V3 is here! Yes, it is the moment you've all been waiting for (well, some of you at least!) Here's the changelog:

-------------------------------------------------
The OpenFRAG Team is proud to announce a new Techdemo!

Techdemo v3 contains better code altogether.

Version 0.6.2 of the
NeoEngine (http://neoengine.sourceforge.net/)
has been used. A big thank you to the NeoEngine
team! A special thanks goes to Arithon for his
help and support in our every need!

We've also used FMOD (http://www.fmod.org/) as
our sound engine. Thanks a lot for the free
sound engine!

We would also like to thank all contributers,
testers, debuggers and you for reading this
and/or supporting us!

Changelog:
[Techdemo v3]
* Some code from the previous Techdemo has been
copied and edited to use the new version of
the NeoEngine (0.6.2, it was written for 0.6.0)

* The rest of the code was rebuilt from scratch,
having in mind a more Object Oriented approach
then was being used before.

* The movement code has been optimized, allowing
smoother movement through the map.

* A menu has been created, for now it only let's
you exit or start the demo (there's also a
'credits' button! )

* A new type of button has been created, allowing
mouse-over events (the buttons change when the
mouse is over them ).

* Our first model has been added. The Knight model
is not fully finished yet, but most of the art
concerning the model has been done. There's no
movement in it yet though.

* 2 Songs have been added, one that plays while you
are in the menu, the other plays while you are in
the game.

* A global trigger object has been added to make it
easier to create new modules for future versions
of OpenFRAG. It is now possible for any module
to raise an event in the main application. (used
to load maps and exit the application neatly when
a module gets in trouble.

Known Bugs:
* In direct3d mode the mouse cursor will not hide,
we're looking into this... on opengl it works fine.

* It is known that on some machines the opengl version
does not work very well (they get about 1 frame per
second).

* You ABSOLUTELY NEED directx9 to run the direct3d
version!!!
-------------------------------------------------

As you can see, Collision Detection didn't make it into the Techdemo . But there is no need to feel bad about that, as there's always a NEXT Techdemo! It has not yet been decided when it will be released, but Collision Detection will be in it!

Have fun!

Here's the link again, so you don't have to scroll up: https://sourceforge.net/project/showfiles.php?group_id=82471
Click to expand...
 
reever said:
Humus said:
Given that AMD is behind this benchmark I would say it's likely biased towards their processors in some way or another.
Click to expand...

You guessed it. Lookin at the results in the post above, the 2400+ gets much better scores than a 2.6P4
Click to expand...

at least AMD is not hidding that they are behind the benchmark, unlike intel and the widely used bapco benchmarks sysmarks.

http://www.vanshardware.com/articles/2001/august/010814_Intel_SysMark/010814_Intel_SysMark.htm

http://www.vanshardware.com/reviews/2002/08/020822_AthlonXP2600/020822_AthlonXP2600.htm
 
reever said:
Humus said:
Given that AMD is behind this benchmark I would say it's likely biased towards their processors in some way or another.
Click to expand...

You guessed it. Lookin at the results in the post above, the 2400+ gets much better scores than a 2.6P4
Click to expand...

some guys AXP 2100+ @ 2170MHz & 9800 Pro - 2505 Marks
My P4 2.4C @ 3.5GHz HT enabled & 9800 Pro - 2628 Marks
:rolleyes:
 
You must log in or register to reply here.

Similar threads

J
Ghost Recon Wildlands [PC]
2 3 4
Replies
63
Views
22K
Jupiter
J
P
Portable console for all PC games - PGS
2
Replies
31
Views
13K
Cyan
Cyan
N
Deja Vu from the past :)
2
Replies
29
Views
5K
function
F
F
GroundWiz RTS Demo and Benchmark
Replies
11
Views
5K
Otto Dafe
O
A
  • Locked
News & Rumors: Xbox One (codename Durango)
613 614 615
Replies
12K
Views
2M
BRiT
BRiT
Back
Top