Despite regular new announcements about the potential viability of wonder material graphene, the senior VP of AMD’s datacentre group has dismissed a move away from silicon as the bedrock of performance computing anytime soon. Forrest Norrod was asked the question aboutour post-silicon future and suggested we’re going to be sticking with tradition for maybe another decade.
The key players in the PC industry have already been laying out their fresh plans for world computing domination, and the Game Developers Conference and Graphics Technology Conference haven’t even started in earnest yet. The latest round of computing announcements have come from the Rice Oil and Gas HPC conference, where Norrod also suggested that, without 3D stacking, AMD’s chiplet technology is just a quick fix for the failure of Moore’s Law.
Intel has also used the conference to announce that its 10nm production process is back on track and ahead of where it expected to be, with products coming “sooner to market.” This latest news, however, is looking much further into the future.
In response to the silicon transition question from the audience at the Rice Oil and Gas HPC conference in Houston, Norrod says:
“The simple answer to the question is over the next seven to ten years… no, I don’t see that. I think there’s a clear path on using traditional silicon down to about 3nm geometry. So another couple of process nodes, which we’ll get to in about 5 or 6 years. I think after that it gets a little fuzzier, obviously we’ve got the dice roll on quantum off to the side.”
Read more: The companies making quantum computing a reality
And the quantum question throws everything into flux. And my brain. The whole quantum computing thing fries my lobes every time I think I’ve got even the most vague handle on it. But it’s happening, and there are even cloud-based quantum machines you can book time on right now, and they look beautiful too.
So, when does AMD reckon we’re going to have quantum computing fully realised? Norrod’s a little more vague about his expectations here…
“Quantum will mature in ten to 100 years,” he says. “I would put it probably… it’s not ten it’s probably more like 20 or 30 years.”
That’s either a long time away, then, or not very far away. Or maybe both at the same time. Who knows, anything’s possible in the quantum game. Whatever, it looks like we’ll be stuck using standard silicon CPUs for another ten years or so.
But that’s not so bad, it’s served us pretty well so far.