AMD A8-3850 APU with Radeon HD Graphics Processors

[Arkanna]

was the variation on this like a considerable difference? I'd been using that site he linked for benchmarks recently and thought I'd found a gold-mine, but that's a little worrisome if there's much of a variation in what their single scores tell and what the sites who run multiple benchmarks are showing

Single benchmarks only tell part of a story. To get the entire story, you must look at a full suite of benchmark tests, since some benchmarks show better performance depending on CPU architecture, number of cores, etc. For instance, usually gaming benchmarks favor faster-clocked but smaller number of cores, while encoding benchmarks usually favor more cores over faster cores.

The best sources I've found for performance analysis remain HardOCP and Anandtech.

[Caiyuo]

thanks for that, getting into the specifics really helped.

[Isiolia]

was the variation on this like a considerable difference? I'd been using that site he linked for benchmarks recently and thought I'd found a gold-mine, but that's a little worrisome if there's much of a variation in what their single scores tell and what the sites who run multiple benchmarks are showing

I'd say so, yes. Look, for example, at the Anandtech review and see the performance as compared to the Athlon II X4 635. They note that the performance is similar - and the various tests run seem to confirm that. At best, the Cinema4D rendering is like a 7% improvement - other tests show hardly, if any, performance difference.

You can also use their Bench page to compare those two CPUs across even more benchmarks here. Outside of a couple of the Sysmark ones (which I wonder if the values were swapped, as other Sysmark results don't show that), the two are relatively similar.

On the PassMark list, the A8-3850 shows as being nearly twice the performance of the X4 635, with a listed score of 6204 as compared to 3362 for the Athlon, or a ranking of 57 over 201.
It doesn't appear to be some odd result - the A6 3650 isn't too far behind, and likewise the X4 640 and all are logically tiered compared to the 635. My assumption would just be that however the PassMark benchmark works, it really favors the CPU architecture of the Fusion chips. If there are actually real-world apps that show similar gains, then it'd be a great deal for the performance. I just don't see it in the actual application benchmarks though.

[Caiyuo]

shit, anandtech is really earning all these mentions, thanks for that breakdown. I'm glad to have a more concise idea of how to answer questions about comparisons, too, not as ambiguous of a process as I always thought it was.

[Isiolia]

One thing I did see when trying to see if anyone had broken down how PassMark worked is that their big list there is (as they state) comprised of user results. Not everyone's results are going to show the potential of a given CPU. It could be that at least some of the disparity is due to some of the 700-odd people who benchmarked the X4 635 having something screwed up, thus dragging the average score down. Since there's no real data there, it's hard to say.

[Arkanna]

thanks for that, getting into the specifics really helped.

I apologize if my response was too generic. :-/ Just tryin to help.

[Caiyuo]

I apologize if my response was too generic. :-/ Just tryin to help.

I was thanking you for that! lol Sorry if it came off sarcastic, I meant in regards to the difference between gaming and encoding benchmarks, I tend to just skip to written conclusions whenever I read about any of this so I had no idea.

[Arkanna]

Oh my mistake >< Tone doesn't traverse the internet too well, I'm afraid.