Testing AMD’s Mantle

Final Thoughts

I think it is fair to say our results with Mantle have been highly impressive. Using an XFX R9 290X GPU paired with a low-end CPU (the FX-4100) and a high-end CPU (the FX-8350) we’ve been able to produce sizeable gains in performance on both:

• In Battlefield 4 Mantle produced a 13.6% boost in average frame-rate for the FX-8350 and a 17.3% boost for the FX-4100
• In Plants Vs Zombies Garden Warfare Mantle produced a 29.2% boost in average frame-rate for the FX-8350 and a 42.7% boost for the FX-4100
• In Thief Mantle produced a 28.8% boost in average frame-rate for the FX-8350 and a 37.6% boost for the FX-4100

The trend among our results seems to be that the “lower-end” the CPU is, the larger the gains in frame-rate are. This is great news for the majority of the PC Gaming market because most people use fairly low-end CPUs rather than high-end ones. Gamers rocking processors like dual core Intel Pentiums, quad core AMD Athlons or quad core AMD FX processors will be able to get significant gains in performance by opting for the Mantle API in Mantle-ready game titles.  From what I could see there isn’t really a down-side to Mantle. In short it offers up extra performance for nothing, simply by changing an option in one of your games.

The future for Mantle is very bright, even if Mantle itself may not be part of the future. Mantle isn’t just an API, it’s a path of progress and development for the PC gaming industry. Some may say the move to a low-overhead API was a selfish move by AMD, done to give them bragging rights on the first low-overhead API or done to ensure that their CPUs can better compete with Intel’s CPUs in gaming. However, selfishly done or not, the Mantle concept will benefit everyone. We’ve all seen the progress that’s been made with GPUs over the last few years – it has been immense. Compare the GTX 580 to the GTX 780, or the HD 6970 to the R9 290X to get an idea of what I mean – performance has more or less doubled. Has the same progress been made with CPUs? No not really. Compare the Core i7 2600K to the Core i7 4770K and you’ll notice a 10-20% difference at best. That said the future of improved gaming experiences lays with GPUs, not CPUs, and so a low-overhead API makes sense for everyone: game developers, hardware vendors and consumers.

Whether Mantle survives the future is another whole other debate. The key determinant to this will be seeing what happens when DirectX 12 hits the market, AMD are in a rather disadvantageous position compared to Microsoft for a number of reasons, and so it seems more likely that DirectX 12 will phase out Mantle, rather than Mantle beating DirectX 12 or both of them coexisting. AMD represent competition to Intel and Nvidia, whereas Microsoft do not, AMD do not have the deep pockets a huge corporation like Microsoft have. AMD do not have the long history of API development that Microsoft also have, neither do AMD have the deep knowledge and understanding of Windows operating systems like Microsoft have. Furthermore, Microsoft are big game developers, they have a significant game developing and publishing division active in both the PC and console markets, if anyone is going to get a graphics API right, it surely is going to be Microsoft. However, even with all that said, DirectX 12 might as well be called “DirectX 12: Inspired By Mantle”. AMD’s Mantle is a ground-breaking concept that works in practice, I just wish more games would support low-overhead APIs like Mantle.

Thank you to AMD for providing the games that made this content possible.