During the 2014 release window of Nvidia’s previous generation graphics cards – GTX 900 Series, media outlets and gamers alike were expecting to get their hands on a set of products which were capable of driving 4K displays. This remained true for the release of AMD’s R9 200 Series cards in the year prior, as well as those which followed in 2015. While all these GPUs are in-fact capable of delivering of 4K experience, the status quo of accompany PC gamers is “Maximum image settings and a 60Hz refresh rate”. Unwilling to settle for console-level image settings even if the resolution itself is 4K with a high frame rate, PC gamers were given the choice to either wait for the next series of cards, or run multi-GPU configurations in order to appease their grand appetites.
Then Vs Now
With the current series of GPUs from Nvidia, the GTX 1070 and GTX 1080 have proven to be 4K capable cards, with the GTX 1070 seeing minimal compromises to image quality in order to fulfill the 4K fantasy. As AMD hasn’t yet released a high-end GPU to address the competition, the RX 580 still delivers a pleasing 4K experience – so long as gamers are aware of the number of compromises needed to be made. Curious of the true capabilities of these cards going forward, my interest here was sparked by the potential capabilities of a recently announced game console – designed for 4K gaming.
Advertising 4K gameplay and 6TFLOPs of graphical power as a means to grab the gamer’s attention, I have no doubt that Microsoft’s upcoming Xbox One X will be able to deliver 4K gaming.While I remain solely convinced that a great number of the games will in-fact be upscaled, along with image quality settings matching the High settings for PC equivalents – 60Hz gaming is out of the question. As witnessed by the aging CPU – AMD’s Jaguar, the clock cycles and levels of efficiency required for each instruction-set remains far too slow for high refresh rate gaming. Roughly of the same performance level to AMD’s RX 580 graphics card, the GPU powering the Xbox One X is indeed a worthy upgrade over the current Xbox One carrying a GPU eyeing up the HD 7790 – albeit at 1.3TFLOPs of performance.
High-End Maxwell Vs High-End Hawaii XT
Keeping in-line with Microsoft’s belief that an AMD GPU carrying 6TFLOPs of computing power is enough for 4K, I decided to test this theory using the highest-end graphics card from AMD’s R9 200 Series GPUs. As the R9 300 Series of GPUs which followed were simply a rebranding with slight refinements to power-efficiency, performance levels remain identical to each respective card in the line-up. Seeing how the GPU in-question fairs against the previous generation high-end Nvidia card, whether or not these cards are viable 4K options for gaming will be seen. Given the architectural differences between AMD’s Hawaii XT architecture in the R9 290X, and the Maxwell architecture on the Nvidia GTX 980Ti, my initial thoughts going in remain highly positive for the Nvidia GPU.
Convinced that both graphics cards will be able to drive 4K, Nvidia’s gains and refinements within the Maxwell GPU have seen real-world performance blow past the Team Red competition, despite the Nvidia cards carrying lesser specifications – on-paper. Looking towards the GTX 970 and it’s 3.9TFLOPs of computing power, the GPU trades blows and in some instances exceeds the 5.6TFLOP AMD R9 290X. With that being said, my thoughts towards the 5.9TFLOP GTX 980Ti was of the highest expectations. This also makes me question the amount – or lack thereof – levels of optimization and GPU efficiency for AMD graphics cards, since their high-end products require an extra teraflop and a half of performance in order to keep up with the Team Green competition. Xbox One X at 6TFLOPs rivaling a GTX 1070 at 6.5TFLOPs, I don’t think so.
So how do the tested GPUs actually stack up to one another?
AMD Radeon R9 290X:
Clock speed: 1070 MHz
Memory: 4GB GDDR5
Memory speed: 5500 MHz
Memory bus: 512-bit
Memory bandwidth: 345GB/s
TFLOPs: 5.6 – 6.0
Nvidia GeForce GTX 980Ti:
Clock speed: 1400 MHz
Memory: 6GB GDDR5
Memory speed: 7010 MHz
Memory bus: 384-bit
Memory bandwidth: 336GB/s – 384GB/s
TFLOPs: 5.6 – 6.0
Configured with an Intel Core i5 3570K at 4.2GHZ and 16GB DDR3 RAM, I decided to go with hardware that’s more in-line with the real-world performance levels that most gamers are likely to be using. That’s right, PC gamers do not in-fact upgrade their configurations every year as they would like you to believe.
Consisting of Gears of War 4, Batman: Arkham Knight, Rise of the Tomb Raider, Tekken 7, and Middle Earth: Shadow of Mordor, neither game within this line-up favours one GPU brand over the other. Along with a mix of games which are poorly-optimized like Batman: Arkham Knight, and Tekken 7 which is extremely well-developed, neither title should yield benefits to one GPU nor diminish the performance of the other. Each game was set to their maximum image quality settings at a resolution of 3840x2160p.
AMD Radeon R9 290X
Nvidia GeForce GTX 980Ti
Clearly capable of delivering a 4K gaming experience at maximum image quality settings, it’s clear the performance required to reach a 60Hz refresh rate wouldn’t have been achieved even if I went with an Intel Core i7 tier processor, nor an AMD equivalent. As mentioned previously, real-world performance for what the majority of users would be running in their systems is part of the benchmarking. Along with proven testing from respectable media outlets it’s clear a multi-GPU configuration would be required, unless gamers would be willing to decrease the quality of the image. While some games may in-fact benefit from an adjustment of image quality, as they turn in little visual feedback that’s worthy of the step from High to Ultra, I suspect most gamers wouldn’t be happy with this. Gaming at 4K 30Hz is more than possible – with the R9 290X being most suited for this, while the GTX 980Ti often hovers around 40 to 50 frames-per-second for the majority of games tested, with Tekken 7 reaching 60Hz no problem.