harmattan
Supreme [H]ardness
- Joined
- Feb 11, 2008
- Messages
- 5,129
Kyle
A suggestion on reviews. I find it really frustrating reading video card reviews because they are just not very applicable to the consumer purchashing the card. Especially for a video card it is extremely unreasonable to assume that the person reading the review is building a new PC, rather it us likely he is upgrading his PC. I understand the concept if removing the bottlenecks of CPU and memory and testing them on a high end PC, but for the most part this is a more real world synthetic benchmark. A review of this card should use a midrange PC of at least one generation old, and the last two generations of video cards should be included. Showing results on an i5-3xxx or i5-4xxx and including say a 670 and 460 (sweet spots of previous generati ons) would give a much more real world and valuable benchmark on whether the purchase is worth it. It answers the question if I upgrade my current video card in the same system to the RX480 what is a real world expectation of increase in video speed
It's certainly an interesting and novel approach, but opens up more problems than it solves, I think. In any to comparative test, there needs to be some sort of normalization. Also, what you're testing here are the video cards - not how one performs with a bottleneck.
I would say a one-off where the review shows different cards performing on different CPUs in different games might be interesting - something I think [H] did a while back - but i suspect it would be practically hard and the payoff not worthwhile. It's a fair assumption that CPUs and other bottlenecks affect performance similarly across GPUs.