Believe your eyes

There’s one very last motive to dispute what FCAT is reporting: It doesn’t in shape how the game seems to run on AMD hardware. The cause that Scott Wasson’s preliminary record on sub-second GPU rendering became so influential is as it crystalized and tested a problem that reviewers and game enthusiasts had observed for years. Discussions of microstutter are as vintage as multi-GPU configurations. right here’s a graph from our original Radeon HD 7990 evaluation:

That microstutter turned into truely, glaringly seen whilst benchmarking the game. It might not have proven up in a traditional FPS graph, however it popped out straight away within the FRAPS frame time statistics. searching at that graph for the primary time, I felt like I’d finally determined a manner to intuitively seize what I’d been seeing for years.

Ashes of the Singularity doesn’t seem like that on an R9 Fury X. It doesn’t look some thing just like the FCAT graph indicates it does. It appears to be similarly easy on both AMD and Nvidia hardware whilst walking at kind of the identical frame price. Granted the experience of smoothness is subjective, however the distinction in presentation among AMD and Nvidia is nothing just like the initial FCAT graph implies.

Ashes of the Singularity measures its very own body variance in a manner just like FRAPS; we extracted that data for each the GTX 980 Ti and the R9 Fury X. The graph above shows  video cards that perform identically — AMD’s body times are barely lower because AMD’s frame charge is barely higher. There are no other sizable differences. That’s what the benchmark “feels” like whilst viewed in character. The FCAT graph above indicates terrific degrees of microstutter that clearly don’t exist when playing the sport or viewing the benchmark.

AMD has instructed us that it recognizes the price of FCAT in overall performance analysis and completely intends to guide the function in a destiny motive force replace. In this example, but, what FCAT shows is happening without a doubt doesn’t in shape the revel in of the actual output — and it misrepresents AMD in the technique.