Last Minute Telegrams: Jon Peddie

How much of a competition do you see with arm becoming in this market in years to come ? Also, as a secondary question if times allows it : Are the rumored rising maximum frequency of future GPU generation a marketing tactic or, are we seeing the emergence of a new performance/power sweet spot allowed by new microarchitectures and process technologies ( new nodes ).

Markus Smith

What was the impact on the professional GPU (and OS) marketshare when Apple moved to the Trashcan Mac Pro? Did it cede so much of the "professional" market that AMD "winning" (by default) the GPUs in the new Mac Pro is meaningless as far as increasing marketshare in the professional market?

TSPCFS

With the reveal of DX12 Ultimate and Ray Tracing becoming more prominent, do you think developers would utilize the mGPU feature in DX12 to improve RT performance? Would we see such a feature more widely embraced for Ray tracing similar to the early days of rasterization?

The Immortal Cameraman

What’s the industry’s position on high NA litho? From what I’ve gathered, ~2023 is when foundries will need to begin multipatterning EUV (which has been deemed not cost effective) or transition to high NA to continue transistor scaling at its historical pace. However, I’ve read that high NA litho implementation is even more difficult than EUV. Could we see the industry stall out on single patterned EUV well into the latter part of this decade (2024/2025)?

What are your thoughts on the AMD/ GloFo wafer supply agreement post Zen 3? If I recall correctly, GloFo is supplying chips well into 2023/2024-I find it hard to imagine 14/12nm chips still being sold then. Will AMD drop GloFo prematurely? What are your thoughts?

Good question. I would hope so, but I wouldn't hold my breath.

How has console gaming revolutionized the GPU industry as a whole?

DeusNightFire

With the biggest part of Nvidia's revenue coming from their gaming GPUs and AMD regaining market share in that space at a steady rate, at what point will they be unable to compete effectively? Nvidia depending 100% on GPUs and AMD depending most of the time on their CPU division to keep the revenue flowing.

Do we ever expect multi-GPU to truly return to the gaming environment? Based on recent advancements with NVLink, some highly utilized workstation software is leveraging the high bandwidth bridge to share resources almost as if they were one GPU. I’m just curious if the gaming market will develop to be able to at least share the resources like VRAM.

Brad Medlin