Review of ATI 3D Expression, Matrox Mystique and #9 Reality 332
Jeremy
The new PC Pro magazine (UK, October 96) has a review of the above
cards. They are also compared to the Diamond Stealth 3D and Matrox
Millenium. Thought people might be interested in an overview.
All tests were done on a P133 with Intel HX mb, Quantum Fireball, 16Mb
EDO.
2D Accel
========
PC Pro Benchmarks (Variety of windows apps run at 1024x768x16bit.
-------------------------------------------------------------------------------------------------------
Diamond Stealth 3D 2.91 ******************************
ATI 3D expression 2.52 **************
#9 Reality 332 2.73 **********************
Matrox Mystique 2.94 *******************************
Matrox Millenium 2.92 ******************************
Quake Test (presumably frame rate at a particular res)
------------------------------------------------------------------------------
Diamond Stealth 3D 18.20 ****************************
ATI 3D expression 16.36 ***************
#9 Reality 332 17.99 **************************
Matrox Mystique 17.54 ***********************
Matrox Millenium 18.24 ******************************
3D Accel
========
Direct 3D Twist (600x400) (doesn't say what options)
----------------------
Diamond Stealth 3D 131 ****************************
(see note 1 below)
ATI 3D expression 70.42 **************
#9 Reality 332 n/a No D3D support yet
Matrox Mystique 120.48 ************************
Direct 3D Tunnel (600x400)(doesn't say what options)
-------------------------
Diamond Stealth 3D 45.45 ***************************
(see note 1 below)
ATI 3D expression 24.27 **************
#9 Reality 332 n/a No D3D support yet
Matrox Mystique 50.50 *****************************
Note 1
---------
The Diamond Stealth 3D values are supplied by me as the review did not
give any Direct 3D figures for this or the Millenium. I have a P133,
32 Mb EDO, Triton 1 mb. I ran the tests at 640x480x16bit with the
following options - Gouraud shading, solid, specular highlights,
texture mapping on, perpective correction on, bilinear filtering on.
My results are using the drivers that came with the DirectX2 SDK, I
haven't yet tried the latest Diamond drivers with Tunnel and Twist (as
I no longer have the sdk installed). Also, I have an OEM Stealth 3D,
the retail version uses faster EDO DRAM, so should produce better
results.
MY Conclusions from above (please feel free to draw your own)
========================
1. Mystique doesn't appear to be much different to Stealth 3D in terms
of 3D acceleration (i.e. pretty average from what I've seen so far -
MTM, Hellbender, Descent 2, Dest. Derby), and it doesn't have bilinear
filtering.
2. #9 Virge card seems to be behind the Diamond card in performance
(the review says that #9 card uses slower EDO DRAM - 50ns). There's
also no #9 D3D drivers for the card yet.
3. ATI 3D seems the slowest of the bunch both for 2D and 3D
acceleration (although they commented on the excellent software MPEG
playback).
Hope this helps.
Jeremy
ALX
In article <3226c8e9...@news.demon.co.uk>, jez...@frimley.demon.co.uk
(Jeremy) wrote:
> The new PC Pro magazine (UK, October 96) has a review of the above
> cards. They are also compared to the Diamond Stealth 3D and Matrox
> Millenium. Thought people might be interested in an overview.
> All tests were done on a P133 with Intel HX mb, Quantum Fireball, 16Mb
> EDO.
> 3D Accel
> ========
> Direct 3D Twist (600x400) (doesn't say what options)
> ----------------------
> Diamond Stealth 3D 131 ****************************
> Matrox Mystique 120.48 ************************
> Direct 3D Tunnel (600x400)(doesn't say what options)
> -------------------------
> Diamond Stealth 3D 45.45 ***************************
> Matrox Mystique 50.50 *****************************
> 1. Mystique doesn't appear to be much different to Stealth 3D in terms
> of 3D acceleration (i.e. pretty average from what I've seen so far -
> MTM, Hellbender, Descent 2, Dest. Derby), and it doesn't have bilinear
> filtering.
Wow. This is in stark constrast to the benchmarks Matrox
published in their most recent Mystique ad. They have the
Stealth 3D at 23 FPS. This is on a P166 using only point
sampling, and at a resolution of 512x384. Either they
fudged the results quite a bit, or Diamond did one hell-of-a
job on the Driver updates.
->ALX<-
->ALX<-
T-Man
Have you had any problems with your card yet? I am buying a 3D card in the
very near future and can't decide between the Mystique, 3D Blaster or the
Diamond. Your opinion would be great...!
Jeremy
unk...@account.com (ALX) wrote:
Yes it doesn't quite seem to agree with Matrox's advertised claims for
the Mystique. BTW as I pointed out in my post the magazine review
doesn't say what options were selected when they ran the Twist and
Tunnel tests, but I listed the options I had selected for my Stealth
3D, which were all of the performance draining ones (e.g. bilinear
filtering).
I should also *restate* that I ran these tests using the S3 Virge
drivers that came with the Direct X 2 SDK and NOT the latest Diamond
ones (as I'd already deleted the SDK when I installed the latest
drivers). When I get a minute I'll reinstall the sdk demos and post
the results with the latest drivers (which should be even faster).
Perhaps someone with the full retail Stealth 3D could do similar.
Jeremy