AMD Vega 10 Graphics Card Finally Shown

By

AMD wasn’t allowing anyone to look behind the side panel of the Ryzen/Vega powered PC when we visited with them at CES 2017, but it appears that Linus of Linus Tech Tips was and likely got paid nicely for it since the video was sponsored by AMD. He was allowed a first look at the AMD Vega pre-production card that AMD is using for testing out the new GPU. So, without further ado here is a look at the AMD Vega 10 engineering card.

AMD Vega 10 Graphics Card

The 4K 60FPS DOOM demo that was running at CES 2017 was run on a high-end Vega 10 GPU that was running about 60-70 FPS with the image quality set to Ultra running the Vulkan API. This AMD Vega engineering card has a USB port on the end of it that allows engineers inside AMD to monitor all the signals from the card. This give AMD a better picture at the power and electrical signals coming out of the GPU with very little overhead. This scene and the uncovered USB port might have all been planned in a marketing push by AMD as they had some power issues on the Radeon RX 480 that could have easily been avoided if caught before the cards were sent out to reviewers and released to the public.

AMD Vega 10 Graphics Card

You can learn more about the AMD Vega graphics card in the video below by Linus called ‘Hands LITERALLY On AMD Vega’ where he shows off the card and speaks with Raja Koduri, Senior Vice President and Chief Architect, Radeon Technologies Group, during CES 2017.

PCWorld also has a 41 minute long video with Raja Koduri where they talk a bit about Vega and he was asked by Brad Chacos if VEGA would ‘still beat the pants off the GTX 1080’ if both were running in OpenGL. Mr. Koduri had the following answer:

I think it would. Because at that resolution things are more GPU bound, less CPU bound. Thats one factor. But let me not say that too confidently because the DOOM developer had done some console optimizations like line assembly shaders and shader intrinsics and I dont think they are available in OpenGL. Thats not a fundamental OpenGL issue, its just that we didnt invest into putting all the extensions back into OpenGL which we could have [done]. I only have a finite number of engineers so theyre just focused on Vulkan.