Article: Stuart "Veridian3" Davidson
Editor/Design: Allan "Zardon" Campbell

Last week we looked at the performance of ATI’s Radeon X1800XT videocard when at stock settings and compared it to numerous other videocard configurations including the GeForce 7800GTX. On the whole the XT was ahead of the GTX in performance when using DirectX based titles. When using OpenGL the performance leader was the GTX. What we also looked at in that article was the new GPU design that ATI were using and made comment on how it differs from other architectures.

Over the past week we’ve seen some comments from people interested in just how good the G70 would be compared to the R520 if both had the same number of pipelines and clockspeeds. So off we went to see if it was possible to get the required configuration up and running.

G70:
By default a 7800GTX is 24pipes clocked at 430mhz with 8 vertex units and memory clocked at 1200mhz. With the help of rivatuner we were able to ensure the card had 8 pipelines disabled while leaving the 8 vertex units intact.

Next up we used Powerstrip to set the 7800 to a clock which was as close as possible to the R520. The clocks we settled on which provided no artifacting etc were 450/1000.

Following the above steps we confirmed that performance had changed in the expected direction from a 24/8 or 20/7 G70 based card.

Next up was the R520 and getting it as close as possible to the above clocks. Using an internal ATI tool we were able to down clock the card to 455/1000 (due to the limitations of the clocking utility there was a 5mhz difference between cards however the performance difference caused by this should be negligible). As the R520 already uses 16 “pipelines” and 8 vertex units no further alterations were needed.

Next: Game Tests