Radeon Linux Driver Shows Possible LCS based Dual GPU VEGA

Published by

Click here to post a comment for Radeon Linux Driver Shows Possible LCS based Dual GPU VEGA on our message forum
https://forums.guru3d.com/data/avatars/m/268/268700.jpg
still can not reach apex...
https://forums.guru3d.com/data/avatars/m/55/55855.jpg
The 1080 Ti competitor. :P
https://forums.guru3d.com/data/avatars/m/239/239452.jpg
Lol they need 2 gpus for compete against single nvidia gpu , not good.
https://forums.guru3d.com/data/avatars/m/270/270008.jpg
Vega @superzeus, We don't know how this performs so your conclusion is a bit premature. For all we know you could be right or this PLX setup could stomp everything Nvidia has to offer. This reminds me of AMD's approach with Ryzen by having dual CCX's. They are not making huge monolithic chips, instead they are tying two medium sized chips together in a fairly efficient way. I hope it works out as we need AMD back for innovation and/or pricing. All I know this is turning out to be a damn good year to be a tech enthusiast with Ryzen, Skylake-X, Coffee Lake 6 core news, Vega and a small possibly Volta very end of year.
https://forums.guru3d.com/data/avatars/m/261/261821.jpg
@superzeus, We don't know how this performs so your conclusion is a bit premature. For all we know you could be right or this PLX setup could stomp everything Nvidia has to offer. This reminds me of AMD's approach with Ryzen by having dual CCX's. They are not making huge monolithic chips, instead they are tying two medium sized chips together in a fairly efficient way. I hope it works out as we need AMD back for innovation and/or pricing. All I know this is turning out to be a damn good year to be a tech enthusiast with Ryzen, Skylake-X, Coffee Lake 6 core news, Vega and a small possibly Volta very end of year.
I remember a Raja interview around a year ago, during the original Polaris launch. In that interview Raja mentioned that GPU's will have to the same route as the CPU's since we're reaching the limits of single chip solutions. He said the future will be multi core gpus. I think thats why the made the Radeon Pro(or something like that) to get more developers using multi gpu's in developing.
https://forums.guru3d.com/data/avatars/m/239/239452.jpg
Well maybe you are right, butt will see soon enough!!
data/avatar/default/avatar14.webp
It's already too late for AMD: NVIDIA GTX 20 Series Hardware SPECS and release date: NVIDIA TITAN Xv - GV102 "Volta" (Fully enabled GV102) - 16.5 Billion Transistors - 42 SMs - 5,376 CUDA Cores - 336 Texture Units - 96 ROPs - 384-bit - 24 GB GDDR6 - 768 GB/s Memory Bandwidth Release Date : Q3 2018 NVIDIA GTX 2080 Ti - GV102 "Volta" (Cut down GV102) - 16.5 Billion Transistors - - 40 SMs - 5,120 CUDA Cores - 320 Texture Units - 88 ROPs - 352-bit - 22GB GDDR6 - 704 GB/s Memory Bandwidth - TSMC 12nm FFN Release Date : Q4 2018 Around The Holidays NVIDIA GTX 2080 - GV104 "Volta" (Fully enabled GV104) - 11.2 Billion Transistors - 28 SMs - 3,584 CUDA Cores - 224 Texture Units - 64 ROPs - 256-bit - 16 GB GDDR6 - 512 GB/s Memory Bandwidth - TSMC 12nm FFN Release Date : Q1 2018 NVIDIA GTX 2070 - GV104 "Volta" (Cut Down GV104) - 11.2 Billion Transistors - 21 SMs - 2,688 CUDA Cores - 168 Texture Units - 64 ROPs - 256-bit - 16 GB GDDR6 - 410 GB/s Memory Bandwidth Release Date : Q1 2018 NVIDIA GTX 2060 Ti - GV106 "Volta" (Fully enabled GV106) - 6.6 Billion Transistors - 14 SMs - 1,792 CUDA Cores - 168 Texture Units - 48 ROPs - 192-bit - 12 GB GDDR6 - 300 GB/s Memory Bandwidth Release Date : Q3 2018
https://forums.guru3d.com/data/avatars/m/197/197287.jpg
It's already too late for AMD: NVIDIA GTX 20 Series Hardware SPECS and release date: NVIDIA TITAN Xv - GV102 "Volta" (Fully enabled GV102) - 16.5 Billion Transistors - 42 SMs - 5,376 CUDA Cores - 336 Texture Units - 96 ROPs - 384-bit - 24 GB GDDR6 - 768 GB/s Memory Bandwidth Release Date : Q3 2018 NVIDIA GTX 2080 Ti - GV102 "Volta" (Cut down GV102) - 16.5 Billion Transistors - - 40 SMs - 5,120 CUDA Cores - 320 Texture Units - 88 ROPs - 352-bit - 22GB GDDR6 - 704 GB/s Memory Bandwidth - TSMC 12nm FFN Release Date : Q4 2018 Around The Holidays NVIDIA GTX 2080 - GV104 "Volta" (Fully enabled GV104) - 11.2 Billion Transistors - 28 SMs - 3,584 CUDA Cores - 224 Texture Units - 64 ROPs - 256-bit - 16 GB GDDR6 - 512 GB/s Memory Bandwidth - TSMC 12nm FFN Release Date : Q1 2018 NVIDIA GTX 2070 - GV104 "Volta" (Cut Down GV104) - 11.2 Billion Transistors - 21 SMs - 2,688 CUDA Cores - 168 Texture Units - 64 ROPs - 256-bit - 16 GB GDDR6 - 410 GB/s Memory Bandwidth Release Date : Q1 2018 NVIDIA GTX 2060 Ti - GV106 "Volta" (Fully enabled GV106) - 6.6 Billion Transistors - 14 SMs - 1,792 CUDA Cores - 168 Texture Units - 48 ROPs - 192-bit - 12 GB GDDR6 - 300 GB/s Memory Bandwidth Release Date : Q3 2018
Nice troll post