ATi's Radeon 8500 Is Final Or Something Like That ...

Page 2 of 7:

Facts Only, Continued

Here's a quick comparison of the Radeon 8500 with its competitors from NVIDIA:

Swipe to scroll horizontally

Header Cell - Column 0	Radeon 8500	GeForce3 Ti500	GeForce3	GeForce3 Ti200
Engine	Programmable T&L (ATi "Charisma Engine II ")	Programmable T&L ("vertex & pixel shader ")	Programmable T&L ("vertex & pixel shader ")	Programmable T&L ("vertex & pixel shader ")
Core Clock	275 MHz	240 MHz	200 MHz	175 MHz
Memory Clock	550 MHz	500 MHz	460 MHz	400 MHz
Theoretical Peak Pixel Fill Rate	1100 Mpixel/s	960 Mpixel/s	800 Mpixel/s	700 Mpixel/s
Peak Memory Bandwidth	8800 MB/s (plus "HyperZ II ")	8000 MB/s (plus "light speed memory architecture ")	7360 MB/s (plus "light speed memory architecture ")	6400 MB/s (plus "light speed memory architecture ")
Typical Memory Configuration	64 MB DDR SDRAM	64 MB DDR SDRAM	64 MB DDR SDRAM	64 MB DDR SDRAM

Radeon 8500 comes with superior memory bandwidth as well as theoretical fill rate, the historically most important figures to estimate 3D-performance of a graphics card.

Here's the same for Radeon 7500:

Swipe to scroll horizontally

Header Cell - Column 0	Radeon 7500	GeForce2 Ultra	GeForce2 Ti	GeForce2 Pro
Engine	Fixed Function T&L	Fixed Function T&L	Fixed Function T&L	Fixed Function T&L
Core Clock	290 MHz	250 MHz	250 MHz	200 MHz
Memory Clock	460 MHz	460 MHz	400 MHz	400 MHz
Theoretical Peak Pixel Fill Rate	580 Mpixel/s	1000 Mpixel/s	1000 Mpixel/s	800 Mpixel/s
Peak Memory Bandwidth	7360 MB/s	7360 MB/s	6400 MB/s	6400 MB/s
Typical Memory Configuration	64 MB DDR SDRAM	64 MB DDR SDRAM	64 MB DDR SDRAM	64 MB DDR SDRAM

Radeon 7500 may have a lower theoretical fill rate than the rest, but the much more important memory bandwidth is on par with the good old GeForce2 Ultra and clearly superior to other GeForce2 cards.