HD 2900XT 1GB DDR4 Benchmarks

[Munky]

Originally posted by: Equinox
The way that 3DMark judges processor "scores" almost seems arbitrary. I swear I've benched a single core and a dual core AMD64 with fraps running, and they honestly seemed to be getting the same framerates in the CPU tests - yet the X2 had a much higher score. It's kind of meaningless the way that processors are scored.

The dual core really runs the cpu test faster, but they both run so slow that at <1 fps you can hardly tell the difference.

 

[Cookie Monster]

We all know the R600XT is a monster when it comes to 3dmark, beating even the 8800 Ultra. But is that so in terms of real world gaming performance?

Thought so.

The clocks of this thing is 825mhz/2100mhz i think. I actually wish it actually performed faster than a GTX. I mean look at the friggn specs on paper. 1GB of GDDR4, 512bit memory interface, 825/2100mhz, 64 vec5 shaders (320ALUs) etc etc.

 

[Oyeve]

Time will tell.

 

[thilanliyan]

Originally posted by: keysplayr2003

Originally posted by: Fenixgoon
3dmark benches do not count anymore, IMO. i want to see real game-based benches, not theoretical testing suites.

Agreed. It would have taken them an additional 4 minutes? Maybe? to run a game bench. Why only include 3DMark which is proving to be truly less and less indicative of graphics card performance? 2900XT scores way higher than it should in 3DMark06 but this does not convey to actual gaming. So what does this tell us about 3DMark? That's right. Useless.

Those scores are from someone at XS ("Denny" something) and you know how much they love their 3DMark. So you can't really blame the store for not testing games.