What benchs should i do?

TSDgeos

Junior Member
May 23, 2001
17
0
0
I'm doing a review of GF4MX440 vs GF3 vs Kyro2 for my webpage and i'm thinking what benchs/game should i use.
At the moment i'm thinking on:
3dMark2001 SE
Comanche 4
Comanche 4
Dronez
Quake 3
XSMark (GLExcess)
Serious Sam 2
Unreal Tournament
Z: Steel Soldiers

Any suggestion?
 

jcmkk

Golden Member
Jun 22, 2001
1,159
0
0
What about the benchmark for the new Unreal engine? RTCW would be nice. Some Nascar 4/2002 guys might wanna see that benchmark.
 

TSDgeos

Junior Member
May 23, 2001
17
0
0
Can the new Unreal benchmark be downloaded?
My website is not as important to have direct phone line with unreal people.
 

Rand

Lifer
Oct 11, 1999
11,071
1
81


<< Can the new Unreal benchmark be downloaded?
My website is not as important to have direct phone line with unreal people.
>>



The latest Unreal Performance Test benchmark can't be downloaded yet.... though supposedly they announced it's going to be available at some point
They have of late become much more open to sharing the benchmark, so if your website is part of the mainstream online press even if it's one of the smaller mainstream sites you have a decent chance of getting a copy of it from the designers. Then again, if if that was true uyou likely wouldnt be posting this question so I'm assuming your site isnt.

RTCW would make a good benchmark, as would Nascar 2. IL2 Sturmovik would be a nice addition for the flight sim fans.
The Black & White intro makes a decent scene to benchmark, that'd be a nice addition.
There are a few tools out to benchmark Max Payne, so that'd be a decent one too.
MDK2 might be okay also.
3DMark 2000 is decent for measuring performance of older ganmes, though I would suggest running it in 32bit color rather then the default 16bit.

Regarding the ones you already mentioned... IMHO:
3dMark2001 SE: Results don't tend to scale well with reality, and it's heavily biased to over-inflate performance of DX8 hardware supporting cards so I don't like it as a benchmark but it's excellent to give n indication of theoretical capabilities in some of it's tests.
Comanche 4: Little first hand experience with it, no opinion.
Dronez: nVidia specific optimizations make this a poor benchmark to compare performance of cards from multiple manufacturers.
Quake 3: The old standby, a natural choice.
XSMark (GLExcess): Decent as a theoretical tool, and good for the 'wow' factor. But is is coded and rendered in a manner considerably different from real world games which makes it of dubious value as an indicator of real world performance. Also tends to be a bit inconsistent in repeated runs, so the margin of error is fairly high.
Serious Sam 2: Decent benchmark.
Unreal Tournament: Very much processor limited, a decent benchmark overall but not that great for measuring sheer graphics card performance.
Z: Steel Soldiers: No idea what this even is, so I've no opinion.


Out of curiosity, might I inquire as to the URL of your site?
 

TSDgeos

Junior Member
May 23, 2001
17
0
0


<< Regarding the ones you already mentioned... IMHO:
3dMark2001 SE: Results don't tend to scale well with reality, and it's heavily biased to over-inflate performance of DX8 hardware supporting cards so I don't like it as a benchmark but it's excellent to give n indication of theoretical capabilities in some of it's tests.
Comanche 4: Little first hand experience with it, no opinion.
Dronez: nVidia specific optimizations make this a poor benchmark to compare performance of cards from multiple manufacturers.
Quake 3: The old standby, a natural choice.
XSMark (GLExcess): Decent as a theoretical tool, and good for the 'wow' factor. But is is coded and rendered in a manner considerably different from real world games which makes it of dubious value as an indicator of real world performance.
Serious Sam 2: Decent benchmark.
Unreal Tournament: Very much processor limited, a decent benchmark overall but not that great for measuring sheer graphics card performance.
Z: Steel Soldiers: No idea what this even is, so I've no opinion
>>



Wow, you smashed all my benchmarks !!!

Of all the other benchmarks you suggested i only have RTCW and for some strange reason the benchmark seems not to like my 256 MB of memory and uses my HD a lot so the fps go down and doesn't is a good bench for the graphics card. 3Dmark2000 is a good option.

The URL to my site is www.secretgamezone.net, it's a VERY small spanish website. If you undestand spanish i'll like your comments.
 

TSDgeos

Junior Member
May 23, 2001
17
0
0
My impressions after some tests
Comanche 4: I think is VERY CPU limited, the GF3 gives me 31.57 fos and the GF4MX440 29.24 this on a XP 1600+, it doesn't lets me test the Kyro 2 as it says the benchmark is only for T&L cards :(
Dronez: Maybe it's nVidia specific, but my little Kyro 2 card gets reasonable fps compared to those GF3 and GF4MX440 numbers

Z. Steel Soldiers: It's a 3D strategy game from the Bitmap Brothers. On the GF3 i get 58.316fps and with the GF4MX440 39.863fps, i still have to test with the Kyro 2.
 

nemesismk2

Diamond Member
Sep 29, 2001
4,810
5
76
www.ultimatehardware.net


<< I'm doing a review of GF4MX440 vs GF3 vs Kyro2 for my webpage and i'm thinking what benchs/game should i use.
At the moment i'm thinking on:
3dMark2001 SE
Comanche 4
Comanche 4
Dronez
Quake 3
XSMark (GLExcess)
Serious Sam 2
Unreal Tournament
Z: Steel Soldiers

Any suggestion?
>>



3dMark2001 SE - Good for the latest Nvidia cards but not for the MX440 or Kyro 2
Comanche 4 - Good as a CPU benchmark but not for a video card review
Dronez - What a dull benchmark, rubbish game too!
Quake 3 - The standard benchmark, use it with MAX settings and disable TC
XSMark (GLExcess) - Used it but gave unreliable results
Serious Sam 2 - Good benchmark but some Kyro 2 problems reported
Unreal Tournament - Use the thunder demo for good results esp Kyro 2
Z: Steel Soldiers - Didn't even know you could benchmark it?

How about using Max Payne as a benchmark?