Unigine Is Working On Perhaps The Most Beautiful GPU Benchmark, Coming Later This Year

[Ehvis]

This chromatic aberration looks really disgusting. I hope they fixed it.

Chromatic aberration always looks ugly. It's usually fixed by allowing it to be turned off.

[bug77]

Fine, release it as a showcase piece then, not as a benchmark, because that's one thing it definitely isn't.

Well, it does show how fast the engine runs on a given video card.
What would make this a benchmark in your opinion?

[dungeon]

It does -not- do any of those things.

Are you offended by something or what?

If there were real games that used the engine, then it would make sense to benchmark those games, and they could tell you those things. But thing CANNOT be extrapolated.

Extrapolated to what, games can be here and there capped by something else then GPU... But Unigine Benchmarks are designed as Benchmark with major B so exactly designed to be GPU stress tool

[eydee]

Yeah, it is about time to have new benchmark demo... but these days it is better to be VK demo i guess

If it works like previous benchmarks, it should support all major APIs on all platforms and you can select which one to use.

[GreatEmerald]

Hm, this reminds me of those Land of the Giants/MiniMe maps in Unreal Tournament games. Fun times.

[duby229]

Are you offended by something or what?



Extrapolated to what, games can be here and there capped by something else then GPU... But Unigine Benchmarks are designed as Benchmark with major B so exactly designed to be GPU stress tool

Which won't do -anything- to tell you about how -any- game performs. It is not a benchmark.

[duby229]

Which won't do -anything- to tell you about how -any- game performs. It is not a benchmark.

EDIT: A benchmark requires requires a reference to normalize from. It's exactly the same reason why you cannot benchmark CS:GO and expect results for HL2. Sure they use the same engine, but the results aren't comparable, there is no reference. None.

[duby229]

If Unigine did have games available, then the it still wouldn't be a benchmark, because then the only thing that would make sense to benchmark would be those games as individual games.

[dungeon]

And when it comes to individual games TF2 is CPU bound everywhere... so how can anybody know which GPU is faster when there is such a cap?

Also TF2 benchmark does not extrapolate well and tell user that game lockup GPUs for about 9 months using radeonsi driver

[duby229]

And when it comes to individual games TF2 is CPU bound everywhere... so how can anybody know which GPU is faster when there is such a cap?

Also TF2 benchmark does not extrapolate well and tell user that game lockup GPUs for about 9 months using radeonsi driver

TF2 is an actual game, with actual players, that actually need to know how it performs. Big difference and is well worth benchmarking.