With nVidia ruling the roost for so long, what could possibly happen if the tide turns and AMD can dominate one area of the market and what is available for manufacturers to take advantage? Can AMD’s business decisions finally pay off for themselves and other companies willing to move quickly into the new markets.
This is a follow on article from my previous post.
AMD’s Present Tech.
A few years ago AMD’s “Eyefinity” really made gaming on multiple monitors an achievable goal for the average techie gamer. It was possible before but was more complicated and usually had a detrimental effect on the gaming experience. Where as Eyefinity allowed you to spread your desktop over multiple monitors or use additional monitors for web browsing etc while gaming, which was possible before but AMD had made it easy. By increasing the ease of setting up multi-monitors the sales of monitors increased and manufacturers improved their design reducing the size of the bezel to improve that wide view.
AMD freesync technology allows the Monitor and the GPU to synchronise their output to stop an update during the scan process on a monitor that can be seen by ‘tearing’ where the image displayed does not match up which is noticeable in a fast-moving scene. Both nVidia and AMD have similar technologies available, nVidia charges monitor manufacturers to use their G-sync technology, where as AMD made their technology open source. As nVidia has dominance over the graphic market both enthusiast and main stream it made sense for AMD to go open source and not charge manufacturers to limit their options when producing monitors designed for use AMD GPU’s.
Mantle was AMD’s API that was designed to reduce the API overhead and give developers more direct access to AMD’s GCN architecture basically everything above the 6000 series of their mainstream cards. After the long development process AMD basically handed over the reins to the Kronos Group and it changed to the Vulkan API. The Kronos Group developed OpenGL which if you where around during the early years of 3D acceleration you would be familiar with. OpenGL became less important to PC gamers as Microsoft’s DirectX matured and became the main API for graphics on the Window’s platform. But OpenGL was still very important on many other platforms and architecture’s, helped by its open source development and its cross-platform utility. Vulkan is to inherit this cross-platform mantle (sorry couldn’t resist!) so allowing development on PC’s, Mobile etc. Getting major manufactures like nVidia, AMD, Intel and Qualcomm onboard will help push this API. So even though nVidia is working with Kronos for drivers for their cards as it’s based on mantle AMD should have an advantage there. With the unreal engine also been able to use Vulkan that could help Linux users who are interested in gaming.
AMD’s GCN architecture had been designed to use asynchronous computing which with the release of DirectX 12 makes a difference to benchmarks. Asynchronous computing allows the GPU to fill in spare compute cycles with out of phase processing. This is not used in DirectX 11 and had not been utilised until now. The early benchmarks for the PC game ‘Ashes of the Singularity’ showed surprising results in early DirectX 12. Because AMD’s GCN architecture has been in development for such a long time it means that many owners will get a bonus when playing DirectX 12 games.
AMD as a whole.
It’s no secret that AMD has not only been lagging behind both INTEL and nVidia in terms of technology but is also suffering financially because of the lack of market share. AMD has been posting losses for some time now even with their dominance in the console market. Both the present generation of Playstation and Xbox use AMD processors, the upcoming mid generation releases of the PS Neo and Xbox Scorpio and also the Nintendo NX will also use AMD chips. These chips used for consoles gave AMD a bit of a lifeline, console chips are high volume but are low margin products so it has allowed AMD to keep all their infrastructure together ready for a change of fortunes. If AMD hadn’t had this life line it would have had to spin-off more business’s to keep the main areas afloat. As the British nuclear industry and government has found out if you lose too much technical knowledge you have to go to other countries/companies and pay a premium for the technology you helped create and pay a premium for the privilege.
AMD is poised to release 3 important technologies this year! On the 29th June (tomorrow!) the Polaris architecture is released in the form of the Radeon RX 480. In late 2016 both the Zen APU architecture and Vega GPU are to be released. In both the APU(CPU) and GPU lines AMD have been only iterating on old technology for the last few years with neither architecture been very good in the first place. The GPU’s where hot running and power-hungry while the APU where underpowered. The Polaris release is aimed at the main stream market with a price point at around $200 and according to rumours will be the entry-level for VR with the 8Gb version. So if Polaris does deliver on power efficiency and performance and defines the main stream entry-level for VR it has a chance of acquiring a chunk of the market from nVidia. The Vega release at the moment is scheduled around the end of October and according to all the rumours is aimed at a more high-end market but building on Polaris with high levels of power efficiency. The Zen APU is also rumoured to be going for the power efficiency/performance sweet spot.
Nvidia’s recent release of the GTX 1070 and 1080 while both very desirable cards are both high-end cards with the GTX 1080 retailing at over $600 and the 1070 at over $300 these are for the reference card, or the standard release. Which according to nVidia marketing are worth paying more for than waiting for board partners to release boards with improved cooling and overclocked out of the factory. With nVidia having a high-end product due to their manufacturing process they have had low amounts of stocks available therefore making the card more valuable. nVidia has been accused of allowing price gauging by keeping the availability scarce and by over pricing the reference card they are allowing other board manufacturers to inflate the price of their products. This will give nVidia low volume but high profit, as they have market share this is not a problem. AMD are going for the opposite approach with the RX 480 release, according to many suppliers in the USA the supply’s of the card are plentiful. The price point is very enticing and is at the upper edge of what is deemed as budget cards without been conned into buying a stunted monstrosity (you know those horrible £/$30 cards that have no right been aimed at gamers!)
INTEL has for many years had a Tick-Tock approach to CPU releases, where they have a major chip update followed by a smaller iteration 18 months later. Recently they announced that this will no longer happen. Due to technological limitations the size of the transistors is getting small enough that bigger problems are presenting themselves. AMD has only been iterating over the last few years even though the architecture name has changed it has just been iterations on the previous design of chip. Zen is a whole new architecture and has had a very long in terms of CPU developement research and development cycle. But is too far into the future to make a big enough impact on any market decisions.
If the RX 480 lives up to the hype and allows AMD to claw back some market share it hopefully stimulate other manufacturers to deliver products based on AMD’s technology. For example free sync monitors could see a big benefit from an increase in AMD GPU’s. If a manufacturer can hit the price/performance mark a new monitor that takes advantage of another feature of a new GPU is enticing. With free sync been open source and not charging a premium for use of the technology this could be a boon to monitor manufacturers. If the 8Gb version of the RX 480 is a good entry-level for VR this could start opening up this market to more than just the techie enthusiasts and creating a bigger market. With both Oculus and Vive competing for your money a value for money entrance requirement could accelerate uptake. Oculus having DRM on their headset (but recently removed) will maybe be forgotten by the time more people decide to take up VR. I believe VR still needs to drop in price as it’s still a niche market. The open area needed to play games that involve movement is an impediment, but the use of a headset in simulator type games is awesome. The ability to head-track using IR was a big step in immersion but taking the next step to be sat in the cockpit of your Spitfire, Cobra Mk III, Sopwith Camel or Apache Gunship is according to VR proponents going to change the way we game. Playstation think so by announcing their own console that will greatly reduce the entry requirement/price for VR (running on AMD tech!) VR also needs a killer game, On the Xbox and 360 it was the Halo games. Though there are many games available for VR no game has caught the imagination of the gaming masses, this could be the problem of representing a 3D world in a simple 2D image in a review. If a Playstation game can ignite this interest especially if it’s not a console exclusive this could push the uptake of this new form of gaming. With a VR environment the input device needs to be appropriate, gamepads due to their size and ease of use are an obvious solution. For Simulator games a HOTAS set-up is great but the learning curve of muscle memory to use all the buttons when you can’t physically see the stick and controls is yet to be overcome.
AMD’s release is not going to allow them to become the majority in the GPU market, but hopefully will allow them to take a chunk out of nVidia’s share forcing nVidia to realise they are not the only horse in the race. Intel and nVidia have become a little complacent having market dominance for so long and been able to charge a premium for their top of the range products (GTX 1080 and the new Broadwell-E chips.) Without AMD in the game other companies have a monopoly and can charge what they want for their products there is not a need to really develop new technology against what a competitor could unveil. So from this perspective I am hoping that the RX 480 is the card the market wants and AMD needs it to be. The card could be the first step in AMD re-entering the race giving the public and corporations choices when buying technology and creating an exciting arm’s race between tech companies. The tech market has lost most of its excitement and ‘turbulence’ with the race been down to just 2 players the same as the CPU market when Cyrix was dissolved in 1997. It is highly unlikely a new competitor will enter either the CPU or GPU market so it would be unfortunate if we lost one of the last players in the field. It’s likely to be an exciting few days for tech
See you on the other side…