Well yeah that's because nvidia still has the lead on performance, even on dx12 because of pascal's async compute. Even with GPUs because with AMD you'll be looking at almost double the TDP, a PSU upgrade and more fans and if you get the wrong motherboard it'll pull too much power from the pci slot and kill your mobo. If you're upgrading an existing build and you already have a good psu, fans and whatnot then AMD is the way to go but if you're getting a new PC nvidia is better in both performance and price because the gpus do cost more, but they don't drag a shitpile of things you have to do to be able to use them which cost money. I personally will go AMD when my gtx 780 isn't enough or shits the bed.
Justin Howard
Considering the guy who designed the architecture is the same guy who designed the last chip that got AMD ahead it's not too farfetched this time.
Jeremiah Carter
Which cards are you comparing? The real TDP of a GTX 1060 is 120 W as advertised, the real TDP of the RX 480 is 165 W so while that is false advertising (they say 150 W) it's still only a 45 W difference. It's slightly over a THIRD more. Double the TDP would mean 240 W! I would be surprised if you could find an AMD card that pulls 50% more than its closest nvidia competitor. Double is an absurd exaggeration.
Mason Ortiz
well, lets hope they're not blowing hot air. There needs to be some competition in the CPU market to stimulate innovation, from both sides, and not having it one sided as it's been for many years now.
Honestly, AMD video cards are starting to come into their own with the new upcoming graphics APIs (many mainstream engines are JUST releasing support for vulkan, so we'll be seeing a lot more of the following soon); where you can see a budget rx480 nearing the performance of a 1070 for vulkan, and surpassing the equivalently priced 1060 in a variety of games tested. This is mainly due to AMD cards being specifically built for async compute (what nvidia cards suck at, as they can only do async compute or general shading stuff like frag/vert shaders); due to fundamental differences in architecture/low-level pipelines. Note that async compute is heavily utilized by the latest graphics APIs, those being dx12/vulkan, and that's why this matters.
Gavin Cook
The only current practical application of quantum computers is approximating solutions to NP-complete problems, and then feeding the results to GPCPU. AFAIK there is only one company producing commercially available quantum computing "hardware accelerators", and that would be D-Wave.
Quantum mechanics doesn't do what you think it does.
