bloodyguard said:
romxxii said:
I responded to him because he was using synthetics to grade GPU's, thing that nobody does nowadays. In Linus's videos he adds the 7970 and the 680 in later ones and by drivers I mean the Never Settle, which boosted the performance of the whole line-up by a significant amount. Also, he tested mostly Gaming Evolved titles because they were more than the Way It's Meant to be Played ones. And I'm pretty sure I said that AMD beats nVidia in their respective price bracket, so no 7970Ghz vs Titan.
I agree that synthetics are a poor measure of GPU performance ranking. However, I believe that Linus' methodologies are not as neutral as they can be. For instance, there are about an equal number of games optimized for Nvidia and AMD, so you should use as large a sample of test games as possible to get a legitimate performance benchmark. Otherwise, you run into the problem of painting an incomplete picture, one that favors a side. In this case, because they choose Gaming Evolved titles, obviously will perform better on AMDs. Now, run the same test, but use only Nvidia optimized titles. The picture will change completely, with AMD losing to Nvidia. Get it?
Personally, I'd prefer benchmarks from guys like Tom's Hardware, who not only test under multiple titles, but also test for frame-time variance. Then there's TechPowerUp, who reviews under multiple resolutions and settings. There's also Xbit Labs, who do performance comparison charts with over 20 games.
