First, let’s take a look at the testing setup:
Intel Core 2 Duo E6850 Overclocked @ 9 x 366MHz = 3.3GHz
Foxconn P35 MARS
EVGA 8800 Ultra Superclocked
| Inno3D 8800|
Ultra 650/1080MHz (Slightly higher than stock)
|Inno3D 8800GTX 575/900MHz|
|~ USD$ 699||USD$ 599|
Drivers used: Official Forceware 158.22.
MipMap Detail setting set to maximum level of High Quality.
When one is interested in a high-end graphics card such as
this, it is only logical to compare it against the top cards in the market, so
we ran it against a normal Ultra, and a GTX, the top dog of yester month. While
the difference between the Superclocked and a normal Ultra looks small when
looking at the surface Core and Memory clocks, the big difference actually lies
in the Shader clockspeed of the 128 Stream Processors of the GPU. Below,
you see the 3 individual clockspeed of the Superclocked compared to the others.
The Shader clocks on the GPU is significantly higher than
that of an Ultra, much more so compared to a GTX. Clocked at 1663MHz, it comes
in 10% faster than a slightly higher clocked Ultra and 23% higher than a GTX. It is in fact one
of the highest clocked shader video card in the market currently. We shall see
how this advantage translate into frame rates.
A higher clocked Shader definitely brings about a hotter
running GPU, below shows the idle and load temperature of this card:
Load temperature goes up to 75C. I’d imagine it to hit close
to 80C without the back-heatsink plastered on by EVGA. As 2D clocks and voltage
is exactly the same as 3D clocks and voltage, the card idles rather warm at 60C.