Last afternoon we had the pleasure of chatting with Tom Petersen (TAP), NVIDIA's Technical Marketing Director and a distinguished engineer. Given that NVIDIA's next-generation graphics card is in the minds of these days, it should not be a surprise that we went straight into exploring the Turing architecture that supports the GeForce RTX 2070, RTX 2080 and RTX 2080 Ti. We peppered Tom with questions about what we expect from Turing's new technologies like Ray Tracing, Tensor Cores, DLSS, and the expected performance promise of NVIDIA's former Gen GeForce GTX 1
You can check out our extensive questions and answers with TAP in the video below:
00:22 – Intro and who is Tom Petersen from NVIDIA?
05:45 – What's New With NVIDIA Turing?
9:05 – What is a GigaRay?
11:35 – Will Ray Tracing work on all NVIDIA GPUs?
13:05 – Is RT Cores (Ray Tracing) coming to CUDA?
14:16 – DLSS – What exactly is and how does AI AA work?
25:30 – How does NVIDIA RTX / Turing perform in older game titles versus GeForce 10?
30:10 – Are there any changes in the Turing GPU for overall efficiency and speed?
31:50 – Is it possible to issue CUDA, RT and Tensor core workloads at the same time?
33:28 – Turing Engineering Investment and Future Gaming Objectives (TAP)
37:02 – Can games use DLSS and Ray Tracing at the same time?
38:09 – What is NVLink and enable the new SLI / mGPU modes?
40:29 – Tickling Frame Buf
42:33 – NVLink, GPU Memory Space and Microstutter
46:25 – Why not NVLink on GeForce RTX 2070?
47:17 – Overclocking Turing – What can we expect?
49:09 – Will GeForce RTX support older SLI in older games and profiles?
49:55 – GeForce RTX prices and answers
52:02 – GeForce RTX founder Release cooling solution design details
54:48 – Liquid refrigerated GeForce RTX FE card?
55:45 – What will players be most impressed with in Turing when they plug in?
58:00 – Tom, what is the most exciting next game that takes advantage of Turing?
We have these timestamps linked to our YouTube channel, for easy scrubbing, or you can manually skip.