Well some of you know that Ati will release their brand new generation of graphics cards this month; codenamed R520.
The official name is the Radeon X1800. It comes in 2 flavours at the moment, XT and Pro. The former being the most powerful. This new generation has some new features:
[ul]
[li]Shader Model 3.0 - This was not supported in the X800 range of cards however nVidias Geforce 6 series cards did.
[/li][li]A new TV Encoder
[/li][li]Crossfire support for multi GPU solutions
[/li][/ul]
The XT and XL cards will also sport 512mb of GDDR3 memory, while lower cards will have 256mb. The cards are also manufactured with a few 90nm lead free process. Another interesting thing is that the X1800XT has 600MHz core clock, meaning less fill rate than the G70.
However, with the hype it may seem that we expected too much from this new generation of cards. In fact they were more aimed to compete with the 7800GTX cards, not so much beat them in performance.
HL2 1600x1200 test
Here, the X800s can’t beat the G70s in the high res HL2 tests.
Not looking too impressive unfortunately:( Thanks for the news Addis:good: Heck the Pro is barley faster than the X850XT in HL2 come on! They’re not bad cards but it’s a long time to wait for results like this.
This website, has more details on it, and also says it will come in 3 flavours and there is also news about the x1600
i have to say, hardware analysis have odne some benchmarks of the x1800xt, and it never beats the 7800gtx, and the x850xt pe is not far behind, and the x1800xt has a bigger price tag that the 7800gtx :o, i would stick with an x800xt pe or a 7800gtx, the x1800xt should just be classed as somthing like X800GTX.
i agree with ex, the performance looks unimpressive, infact for new cards from ATI the performance is crap, X800XT PE are better, when they came out, there performance was amazing, i will not be in a hurry to upgrade to pci-express to get a new ATI card now.
I actually read on the Inquirer that ATI had not actaully released any cards yet, to any vendors or companies which would seem to make this benchmark a hoax, yet that seems pretty far out as it’s feasible enough.