AMD 790FX Vs NVIDIA 780a Chipsets

[Comment Bot]

Please feel free to comment about our story entitled "AMD 790FX Vs NVIDIA 780a Chipsets"

[nanohead]

A decent enough comparison, however....HOWEVER....

The actual REAL performance differences between the 2 products is NON EXISTENT. As usual with review sites, the scale on the graphs is designed to show HUGE differences, when in fact, the differences do not exist (with very few exceptions in this review). Most REAL statisticians would THROW out the differences in performance you document as noise in the testing system.

The Premier Elements part, if you look at the graph, looks like a HUGE difference, bit is in fact 0.7%, which is not relevant by any measure, and could be that there was some dust in the power supply fan.

You use terms like "Nvidia bested" when the actual difference was .002%, but the graph makes it look huge because of a manipulated scale. I personally don't care which one is faster, but your conclusions appear manipulated, when in fact, your test is pretty good.

The Less than 1% difference between these 2 products could be just as easily attributed to the time of day and whether the power company was supplying 222 volts as opposed to 219 volts to the wall socket. Silly.

For all we know, you shut off windows indexing for the Nvidia tests, and left it on inadvertently for the AMD test. Not saying anything like this occured, please don't misunderstand. TweakTown has some of the more enlightened technical and product reviews out there. BUT, and a big BUT, these kinds of presentations call into question methodology and baseline testing approach.

[deanjo]

Incomplete. How can you compare chipsets without seeing how well the overclocking abilities are on them? Far too many times we have seen in the past where even though chipset A vs chipset B are supposed to be roughly equivalent in specs and features one of them will pull away from the other when overclocking is thrown into the mix.

[sublifer]

Originally Posted by nanohead

A decent enough comparison, however....HOWEVER....

The actual REAL performance differences between the 2 products is NON EXISTENT. As usual with review sites, the scale on the graphs is designed to show HUGE differences, when in fact, the differences do not exist (with very few exceptions in this review). Most REAL statisticians would THROW out the differences in performance you document as noise in the testing system.

The Premier Elements part, if you look at the graph, looks like a HUGE difference, bit is in fact 0.7%, which is not relevant by any measure, and could be that there was some dust in the power supply fan.

You use terms like "Nvidia bested" when the actual difference was .002%, but the graph makes it look huge because of a manipulated scale. I personally don't care which one is faster, but your conclusions appear manipulated, when in fact, your test is pretty good.

The Less than 1% difference between these 2 products could be just as easily attributed to the time of day and whether the power company was supplying 222 volts as opposed to 219 volts to the wall socket. Silly.

For all we know, you shut off windows indexing for the Nvidia tests, and left it on inadvertently for the AMD test. Not saying anything like this occured, please don't misunderstand. TweakTown has some of the more enlightened technical and product reviews out there. BUT, and a big BUT, these kinds of presentations call into question methodology and baseline testing approach.

Agreed. 3 points out of over 6300 on pcmark is a joke. Even though you wrote in the bottom that the difference was minor, the visual relative comparison is VERY misleading. There are some that just look at graphs and pretty pictures and don't bother to read or even look at the numbers.