Is My Comp Going To be able to run this?

[oldkawman1]

I couldn't find where the Anandtech review specified that it was a 512MB version. With the Gamespot review it could be a typo, but they did list it twice--both in the chart itself and then down lower in the page in the system specs.

Now in this review over at Tom's Hardware, they mention in the text earlier that the 8600GTS uses 256MB, but difficult to tell how it compares in this discussion since there are no lower cards to compare it. LOL

Then there was a recent international review (I think it might have been German) with performance charts that might have listed the different cards, but I didn't bookmark it.

Right, had to Sherlock that one. Anandtech did not specify, but, since the benchmark result in this one from Tom's Hardware,

http://www.tomshardware.com/2007/10/...gt/page12.html

did specify 512MB and the FPS was 38.3 while this benchmark from anandtech did not specify gpu memory

http://www.anandtech.com/video/showdoc.aspx?i=3128&p=3

but it's results had 41 fps, with both tests at 1600x1200 everything max I deduced it had to be the 512MB version. I figured no way the 256MB version of the 8600GTS could outpeform the 512MB version of itself.

I would like to see a UT3 benchmark program we can download and run ourselves, like ut2004 had. I do not use AA because it reduces my fps. It may be nice to know how much and have a common benchmark program so results can be compared.

Every benchmark runs everything at max, mostly resulting in poor fps results, which has very limited utility as no one will actually run like that. It only answeres the question if you can run max out at this res with that card.

[cel4145]

Right, had to Sherlock that one . . .

You certainly could be right. However, notice that the 8600GT and 8800GTX when compared in the Anandtech and Gamespot reviews at the same res are showing significant performance increases in a much higher percentage than the 8600GTS increase: 8600GT, 15% increase; 8600GTX, 24% increase. Meanwhile, the 8600GTS difference in FPS is only 5%.

If you look at Toms Hardware review chart and compare the 8600GTS 256MB and 8600GTS 512MB, they are neck and neck with each other in many games. For example, look at Oblivion at 1600x1200. Both 8600GTS are right there with the 7900GTX.

Now in comparison to the 7900 series with all of those tests in the Toms Hardware vga charts, 86000GTS is all over the place where one could certainly conclude that overall, the 7900 series beats out the 8600GTS in a wide variety of games. Yet, in those tests the two versions of the card are almost always close together in performance. This is why I decided to go with the 256MB version over the 512 version (that and I was updating mine and my son's computer at the same time, buying 2 PSUs, a mobo, processor, 4GB ram, and a new cpu--lol). It didn't seem like the 128bit processing was giving much advantage to the extra memory in many test cases with games on Toms Hardware (and a suggestion I read elsewhere in reference to the 128 bit limiting access to the 512MB, but can't remember where I read it for now).

Since Gamespot is using the retail game instead of the demo and newer Nvidia beta drivers, it is possible that the game and Nvidia are even better optimized for the 8000 series in some way. And I don't know enough about chipsets, but is there something in the 8000 series that might allow for additional processing over the GPU in the 7900 series, something that Nvidia and UT3 are taking advantage of? Oblivion seems to be doing it in that Toms Hardware test. That would explain why the 8600GTS 512MB did so well against the 7900GTX in the Anandtech review and could certainly make a case for Gamespot's testing of the 256MB as they claim.