Testing AMD’s Mantle
Thief
“Garrett, the Master Thief, steps out of the shadows into the City. In this treacherous place, where the Baron’s Watch spreads a rising tide of fear and oppression, his skills are the only things he can trust. Even the most cautious citizens and their best-guarded possessions are not safe from his reach….As an uprising emerges, Garrett finds himself entangled in growing layers of conflict. Lead by Orion, the voice of the people, the tyrannized citizens will do everything they can to claim back the City from the Baron’s grasp. The revolution is inevitable. If Garrett doesn’t get involved, the streets will run red with blood and the City will tear itself apart.” From ThiefGame.com
To test Thief we simply used the game’s built-in benchmark. We ran the game in Mantle and DX11 mode, restarting each time the API is changed (something that had to be done for all the games). We used the Very High preset and the 1920 x 1080 resolution to test.
Thief showed some incredibly gains yet again: 13 FPS on both the FX 4100 and FX 8350 from switching the API. For the FX 8350 the frame-rate was also so much more consistent meaning not only was there a higher average frame-rate but a more fluid gameplay experience: no juddering from low minimum frame-rates.
How about with core i3 or Pentium Anniversary edition with mantle? i wonder how much gain they can get…
well you will get a small performance boost, but mantle is most efficent with AMD gear. I remember i saw some video (linustechtips) where they compared AMD some 4k or something APU series and a R9 2xx and some intel i7 i think it was with a r9 2xx and the AMD combo was significantly better
Yes its about overcoming CPU bottlenecks so AMD CPUs always gain the most as they always perform worse so there is more bottleneck to be removed by Mantle.
core i3 and pentium anniversary is intel low-mid end offering so i just wonder if the gain will be enough to close the gap against core i7. given im still running core i3-530, how much of gain is what im trying to find out but most review out there prefer to compare a10 apu and core i7…
Well we do not have either of those CPUs but we will expand testing to include the anniversary edition CPU when AMD add more Mantle games later on. I’d presume the pentium anniversary would make a 10-15% gain. On your i3 530 which is clarkdale and fairly old I’d say you’ll see similar gains to the FX 8350: 20-30%.
not likely. even with cpu bottleneck removed, AMD cpu’s are still behind by about 2 cores to 1 margin. I know AMD loves to pull the card that their competitor is doing things to make games run worse on their hardware. So could the test be setup to run worse on intel chips?
Without Intel CPUs this article is not so much. We all know AMD CPUs cause bottleneck. Lets see some amd and intel comp.
I agree, AMD cpu’s being slower per core means more likely to be the bottle neck using directx.
bottleneck my dick i can play BF3 on ultra @ 60fps on my 7850k
Also dont forget to test mantle with 2 gb amd vga card.
Huh, funny thing with theif test, my 4770k at stock clocks and a gtx780 avg’s almost 70fps using directx. Yet an 8350 + 290x using mantle is 15% slower. yes using very high settings as well.