Shootouts

Q3 2006 Midrange GPU Shootout

By Vincent Chang - 4 Aug 2006

Aggregate Performance

Aggregate Performance

The purpose of doing this survey of the midrange graphics segment and their performance in certain popular benchmarks and games allows us to find out which graphics card gives you the best value in this segment. It also allows us to find out how much we have progressed from the previous generation of GPUs. To answer these questions, we have aggregated our benchmarks results to form a composite index for all the cards tested. Using the midrange symbol of the last generation, the GeForce 6600 GT as the baseline, we compared how this generation of chipsets fared against it. For those who want more specific results, here is a table summarizing the performance gains moving on from the GeForce 6600 GT to each of the newer cards:

Performance Gain/Loss Comparison
Performance Gain / Loss Compared against the NVIDIA GeForce 6600 GT 128MB ATI Radeon X1800 GTO 256MB ATI Radeon X1600 XT 256MB ATI Radeon X1600 PRO 256MB NVIDIA GeForce 7600 GT 256MB NVIDIA GeForce 7600 GS 256MB NVIDIA GeForce 7300 GT 256MB DDR2
3DMark05
(No FSAA / AF)
102.96% 64.43% 21.86% 82.09% 14.06% 5.46%
3DMark05
(4x FSAA / 8x AF)
174.37% 100.63% 47.03% 125.10% 34.38% 27.90%
3DMark06
(No FSAA / AF)
84.89% 51.35% 14.92% 89.76% 29.97% 18.80%
Splinter Cell
(SM 3.0, HDR, 8x AF)
88.44% 16.82% - 11.07% 74.35% 8.07% - 2.69%
Quake 4
(No FSAA / 8x AF )
36.72% - 0.50% - 25.69% 51.91% 3.74% 7.05%
Quake 4
(4x FSAA / 8x AF)
82.32% 24.67% - 14.07% 75.92% 4.73% 11.52%
F.E.A.R
(No FSAA / 4x AF)
67.93% 24.70% - 3.94% 85.91% 22.42% 0.00%
F.E.A.R
(4x FSAA / 8x AF)
117.38% 57.14% 9.64% 91.43% 17.38% 3.57%

As for the benchmarks used to create this composite index below, we took one set of results from each of the benchmarks (to avoid overweighing the importance of any benchmark), meaning where we've more than one set, i.e. with and without anti-aliasing, we would choose the appropriate results to use. For instance in F.E.A.R, we took the set of results without anti-aliasing. This is because with the eye-candy turned up, all the cards involved in this comparison returned rather pathetic and unplayable scores, which we felt would not reflect the actual performance setting of an actual gamer. Similar considerations affected our decision to opt for the same approach for Quake 4. However, we did choose the anti-aliasing set of results for 3DMark05, since that was a synthetic test of GPU prowess and should distinguish the limits of each GPU.

Finally, we compared the performance of the cards against that of the GeForce 6600 GT and aggregated the difference (mostly performance gain) with respect of the veteran card. So in the index below, the GeForce 6600 GT assumes the baseline with a score of 1.00. The score of 1.57 for the ATI Radeon X1800 GTO for example, means that it is up to 57% faster than the GeForce 6600 GT. From our index, ATI's Radeon X1800 GTO had the best performance for our set of benchmarks, narrowly beating the GeForce 7600 GT. The much better performance of the GeForce 7600 GT compared to the Radeon X1600 XT was also the reason why ATI had to reinforce its midrange with the Radeon X1800 GTO earlier this year. In any case, the ATI's solutions had the best absolute performance when compared to NVIDIA's equivalent competitors. However, at what cost is this performance achieved? Is the performance advantage linear to its price tag? The following page focuses on the performance-to-price ratio that will answer these questions.

Join HWZ's Telegram channel here and catch all the latest tech news!
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.