Performance per dollar based off overall system price, which CPU is you?

[Garcian Smith]

Soleron said:
Those diagrams are really useful. The i5 750 and the X4 965 are clear winners here (at the kink between the steep section and shallow). It also shows it's not really worth going for a slower CPU that that as the money saved is negligible compared to system price, and more than that is not worth the extra money.

Though Athlon IIs consistently come ahead of Clarkdale and Core 2 in that metric. Impressive when Athlon II is a 45nm quad fighting a 32nm dual. Shows how large the margin Intel really has on those things.

There's many problems with this article's methodology that render it less than scientific, and far from conclusive.

For one, I'd imagine that most people reading these boards don't care about their CPU's performance in regards to 3D modeling, video editing, file compression, or Folding@Home - all cases that favor quad-cores and/or CPUs with an L3 cache over their cheaper counterparts. General productivity and gaming have different (usually lesser) processing needs than those applications, meaning that a CPU that performs well at the price:performance standpoint for the latter will probably be overkill for the former. The i7-860, for example, is a good value for CPU-intensive tasks at $300, but performs identical to the $200 i5-750 in games.

Second, the game selection for the test is lacking. Give me a test that randomly samples at least 12-15 games and you may have me more convinced.

Finally, I'm not sure why the graphs showing "system price" are used instead of the graph showing a simple comparison based upon CPU price alone. The writer's component selection is somewhat arbitrary, and it seems to have been designed only to weight the result in favor of the Intel Core-iX line rather than AMD's Athlon II line. Given what we know about hardware reviewers and industry payola, that should be more than enough to raise suspicions.