Consoles have long touted the phrase “close to the metal” as a means to explain that game developers have fewer software-side obstacles between their application and the hardware. One of the largest obstacles and enablers faced by PC gaming has been DirectX, an API that enables wide-sweeping compatibility (and better backwards compatibility), but also throttles performance with its tremendous overhead. Mantle, an effort of debatable value, first marketed itself as a replacement for Dx11, proclaiming DirectX to be dead. Its primary advantage was along the lines of console development: Removing overhead to allow greater software-hardware performance. Then DirectX 12 showed up.
DirectX is a Microsoft API that has been a dominant programming interface for games for years. Mantle 1.0 is AMD's abandoned API and is being deprecated as developers shift to adopt Dx12. The remnants of Mantle's codebase are being adapted into OpenGL, a graphics API that asserts minimal dominance in the desktop market.
With AMD's Mantle in dire straits and losing ongoing support, the question of timing for its inevitable death has been fresh in our minds. Microsoft's DirectX 12 promises to accomplish many of the same objectives that made Mantle appealing – namely, putting developers “closer to the metal” – while being distributed alongside the prolific Windows OS; this, we think, has already stifled Mantle's viability to developers.
DirectX 12 has been discussed by nVidia and Intel for a while now, with AMD only responding occasionally to recommit to Mantle. The API is still far away for gaming uses -- at least a year -- but it's making the rounds at SIGGRAPH 2014 in
Intel demonstrated a Haswell-equipped tablet running graphics stress test software that toggled between DirectX 11 and DirectX 12. During the demonstration, the company was able to yield nearly a 70% performance increase in Dx12 over Dx11, jumping from 19FPS to 33FPS. Intel attributes this gain largely toward reduced overhead in the API (putting developers "closer to the metal," as Mantle does), then pointed toward multi-threaded rendering optimization.
AMD's Mantle had a rocky unveil with Battlefield 4 and has seen fierce attempts at invalidation by nVidia, but the company continues to plow through difficulties. In the face of DirectX 12 -- still some 20 months out -- Mantle has just announced its partnership with "40 unique development studios pre-registered for private beta" of their Mantle SDK. Developers can use a new portal to access information pertinent to Mantle.
We've been following Star Citizen fairly extensively since its 2012 campaign. As journalists, part of the job is "discovering" games before they make it big; I always task writers with dedicating some portion of our time at PAX to discovering indie games, the hope being that one goes mainstream after we've made it in the door early. I vividly remember Star Citizen hitting the $800,000 mark on Kickstarter and feeling like I'd missed the boat for journalistic success -- it was at the height of its campaign and everyone else had already started talking about it. Even still, we linked up with CIG CEO & Chairman Chris Roberts to discuss technology in-depth (lots of hardware conversation in that link), which had been entirely unexplored up until that point. It's still one of my favorite articles I've worked on, and much of that content remains relevant through today. Funny how much I've learned since then, too.
Months later, we caught up with Roberts at PAX East 2013 shortly before a discussion panel (filmed). Fast forward to July, and we found ourselves at the Cloud Imperium Games office in Santa Monica. At this point, Roberts' next major goal was $21 million; that'd allow him the freedom of ditching private investors in favor of crowd-sourcing the entire game, he told us, and it was no longer a pipe dream to do so. Everyone in the room knew the funding target was on the horizon, it was just a matter of when. I don't think any of us could have told you that Star Citizen would be sitting at $42 million -- more than double our July meeting -- less than a year later.
Almost immediately after our press conference with nVidia concluded -- the one directly challenging the validity of Mantle -- AMD contacted us for a discussion on their R9 295X2 video card. The R9 295X2 has been spoiled for quite a while in traditional AMD marketing fashion, namely by sending extremely flattering photos to some of the major tech outlets. As of last week's call, we were able to get the full Radeon R9 295X2 specs, including TDP, fab process, memory & buses, and details on the difficulty with PSU support.
Let's get right to it with this one.
In my many years working on the journalism side of this industry, I've never seen nVidia put forth such an aggressive stance as exhibited during last week's press conference. We'll start this post with some rapid-fire catching up from the last few months.
The past months have been very AMD-intensive. AMD's Mantle API fronted momentous marketing outreach, touting a bypass to DirectX's performance overhead that has historically been a drain on CPU and GPU output. The hardware has been held back by the API, we were (somewhat accurately) told by AMD, and Mantle was the proposed solution to put developers "closer to the metal" -- or closer to the hardware-level -- similar to console development. This news came to be during a period of silence for Microsoft's DirectX API, which hadn't seen noteworthy development since 2012 with Dx11.1 (at that point). It was the ideal opportunity for an emergent API to make a big splash without significant, refreshed competition.
The name is still unknown, but what will eventually become DirectX 12 should be shown off at GDC shortly; we'll be in attendance to report on the new announcements and will also be attending the GPU Technology Conference the following week, so check back for deeper analysis as we are exposed to information. In the meantime, Microsoft's new iteration of DirectX has some between-the-lines reading for AMD's Mantle.
AMD's newest Catalyst 14.1 bet drivers introduce Mantle support on all GCN-enabled GPUs available, including modern APUs and all Radeon 7000 cards and onwards (7000, 8000, RX 200, and Kaveri APUs). In testing performed around the web, the most significant performance gains can be found in the 290 & 290X, 260 / 7790, and Kaveri APUs.
Mantle makes the most sense as a booster for low-end hardware (like APUs) moreso than the high-end stuff; with an APU, your borderline between medium and high settings at 1080p is going to be more noticeable than on a 290X (which is already more powerful than any game can truly utilize). Mantle makes the difference between playable and unplayable FPS on Kaveri chips and low-end discrete GPUs.
Despite recent patches to Battlefield 4, issues within the game are still prevalent on every platform. Issues being experienced have been addressed by DICE on their forums. Courtesy of their post updated on the 20th, known issues are listed below.
We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.