To everyone griping about this being "unfair" or "unrealistic," this chart, to me, shows what you can do, best-case-scenario, with the platforms available. AMD vs Intel isn't comparing apples to apples, and most hardware junkies know that. AMD has a good implementation strategy (key being strategy), but not a real-world impact on par with Intel's theoretically flawed quad-core design.
AMD is a budget system. Look at the price of the most expensive AMD chip, the Phenom 9950 compared to Intel's top, the Extreme QX9775. AMD's is $290. Intel's is $1500. Is that really a valid comparison? No. But compare what you get from Intel for $290 versus what you get from AMD for $290, and you're in the right ballpark.
Furthermore, each application references the processor and system resources differently, so while you see a huge gap between quad- and dual-core systems on some threaded applications, games waver between dual-core and quad-core equally. Some, like Crysis, show a big preference for Intel chips' clock speed, while others like Supreme Commander yield a negligible performance difference, even though the primo Intel chips cost $1200 more.
That's why articles like this are so damn good. You can't buy every chipset, every processor, and every memory standard, then test every possible combination. That would yield about 2,000 different setups, not including graphics cards -- a number that would probably make any writer or benchmarker quit their job. What you can do, though, is change as little as possible between tests and qualify your results based on your controlled factors. So, Tom's did their job beautifully in that regard.
How you use data like this is up to you. Personally, I look at the game FPS results and some application benchmarks like file decoding and compressing, consider what factors go into that number (memory type, processor load, graphics load, etc.) and take those into account in addition to how I'm going to use my system. Eventually, I make a judgment call based on my knowledge and projections. Personally, I think the AMD quad-core solutions are a good budget option, worthy of considering alongside the Intel E8400. You can argue that til the cows come home, but the numbers justify it in most cases. For now, I have an E8400 slightly overclocked on a nvidia 600-series chipset and likely will not be upgrading that for a couple of years.
AMD is a budget system. Look at the price of the most expensive AMD chip, the Phenom 9950 compared to Intel's top, the Extreme QX9775. AMD's is $290. Intel's is $1500. Is that really a valid comparison? No. But compare what you get from Intel for $290 versus what you get from AMD for $290, and you're in the right ballpark.
Furthermore, each application references the processor and system resources differently, so while you see a huge gap between quad- and dual-core systems on some threaded applications, games waver between dual-core and quad-core equally. Some, like Crysis, show a big preference for Intel chips' clock speed, while others like Supreme Commander yield a negligible performance difference, even though the primo Intel chips cost $1200 more.
That's why articles like this are so damn good. You can't buy every chipset, every processor, and every memory standard, then test every possible combination. That would yield about 2,000 different setups, not including graphics cards -- a number that would probably make any writer or benchmarker quit their job. What you can do, though, is change as little as possible between tests and qualify your results based on your controlled factors. So, Tom's did their job beautifully in that regard.
How you use data like this is up to you. Personally, I look at the game FPS results and some application benchmarks like file decoding and compressing, consider what factors go into that number (memory type, processor load, graphics load, etc.) and take those into account in addition to how I'm going to use my system. Eventually, I make a judgment call based on my knowledge and projections. Personally, I think the AMD quad-core solutions are a good budget option, worthy of considering alongside the Intel E8400. You can argue that til the cows come home, but the numbers justify it in most cases. For now, I have an E8400 slightly overclocked on a nvidia 600-series chipset and likely will not be upgrading that for a couple of years.
Facebook
Twitter
Instagram