🚀 go-pugleaf

On Thu, 19 Apr 2007 00:36:42 +0200, Bruno <dont.use@this.address.ever>
wrote:


>> The first graph is representative of the setup for a typical UT99
>> gamer.... 1024 x 768 with no AA or AF.  
>
>I thought a typical UT gamer would use way higher resolutions however I 
>did dig up some data for those playing CS, Half Life and other Valve 
>stuff and I stand corrected (providing those figures are similar for 
>UT). It seems less than 10% is using a resolution above 1280x1024 and in 
>fact about 44% is using 1024x768 or lower!

Thanks for the link.   I was already familiar with those stats
(they're quite fascinating) since I've used them for reference before.
The average system is quite modest. 

>I'm very surprised of this. I know that some years ago a screen letting 
>you go 1600x1200 cost a pretty penny, in fact the two monitors I 
>mentioned above cost me $3000+, but it has since then become very 
>affordable - even if you insist on using a TFT monitor it's a bargain 
>compared to back in time. These days one can get a 2nd hand super high 
>end monitor like my old Sony for like $100 or so, simply because so many 
>people have abandoned CRT's despite them still being the best thing for 
>gaming (and photo work if you like the colours to be correct).

There's a considerable percentage of hardcore PC gamers that would run
their games in wire frame at 8 bit 640 x 480 if they thought they
could get an extra 10 fps.  :)

>> We can see that there are 16
>> cards that have the same framerate (given a 3% margin of error).
>> Obviously we are CPU bound at that resolution, since those 16 cards
>> represent a varying amount of pixel pushing power yet they all yield
>> the same results.
>> 
>Exactly.
>
>
>> In order to "separate the pack" for a vid card shootout, you need to
>> raise the visual stakes, and they do that by using artificially high
>> AA/AF or resolution settings.  There's certainly nothing wrong with
>> that, but if you do all your gaming at "normal" resolutions and AA/AF
>> levels, then any one of those 16 cards are just as good as the other
>> in UT.
>
>Now here's the thing. The AA/AF settings aren't really artificially high 
>but more like what should be the norm, after all we deserve the games we 
>play to look the best possible.

Turning on decent AA/AF levels on my system makes a modest difference
in UT2004 ONS.  I usually keep it off though.  In a fast paced FPS,
looking at the scenery is not high on my list.  :-)

>> 
>> Now look at the results from the same article when benching Doom 3....
>> http://www.tomshardware.com/2004/10/04/vga_charts_iv/page9.html
>> 
>> Now instead of 16 cards, we have only 3.  Now which game engine would
>> you call CPU limited?
>> 
>No question Doom 3 is a very different beast. However just because it's 
>much harder on the gfx card does not mean UT is not demanding once you 
>up the resolution and gfx quality settings, the level is just different.
>
>I shall try and do some testing in the weekend. Perhaps I can make a 
>graph showing what's the limit in my system - just for the fun of it.

I've wasted entire weekends doing just that sort of thing.  Are we
geeks or what?