Because Mantle was killed off at GCN 1.1
FURY X is therefore slower in Mantle mode than Direct3D11.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
The failure that is the Fury X...how can people trust AMD?
- Thread starter Sprayingmango
- Start date
Because Mantle was killed off at GCN 1.1
FURY X is therefore slower in Mantle mode than Direct3D11.
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,746
Wasn't Mantle designed to help lift CPU bottlenecks for low/mid range rigs?
I imagine testing BF4 with a Fury X and something like a 4770k or better won't show much improvement.
The high end GCN 1.1 cards still saw a boost in Mantle games over DX11. It was simply smoother framerates, less jittery. Seeing Dragon Age Inquisition on my friend's 290X rig and A/B comparing Mantle and DX11 made me a believer in Mantle (thus my high hopes for Vulkan)
Anyway Brent already tested and posted the results in the big review thread. In BF4, FURY X was 15-20% slower in Mantle mode than DX11. Kyle mentioned they didn't include the Mantle results in the review because it wasn't faster than DX11. Seems more than fair. Had they included the Mantle results people would've just accused them of bias and crapping on the card needlessly.
Last edited:
See, now here's the problem I have. From day one people were creaming their pants about these "low-level APIs" and I said from day one that by the technical definition of the term, we (in the PC space) would be screwed going that way. Because "low-level" means targeted for each specific GPU, or maybe, if you're lucky, each GPU architecture. This is why console ports are universally trainwrecks on PC. Or, to draw an analogy, look at what you have to do in CUDA to get best performance for each card.
If Mantle was a good API you should be able to, without touching the game, run it on a newer, faster card and expect a similar performance increase.
Assuming this isn't just a one-off mistake with the Fury X, what this portends is that old games are going to run like shit on new cards, and new games are probably not going to target performance on old cards because it's too much work to cover every GPU. The PC industry will become insanely more fragmented and the quality of games will be a lot like this Batman for some cards.
I feel like before the industry rushes head-long down this path, maybe we need to examine both the pros and the cons. So far all I've heard is a list of pros and nobody is talking about the cons.