At the Global Foundries Technology Conference, AMD’s CTO Mark Papermaster announced that the company will be transitioning “graphics and client products” from the Global Foundries 14nm LPP FinFET process it uses today to the new 12nm LP process in 2018. Global Foundries also announced that 12LP will begin production in 1Q18.
We followed up with Papermaster in person and confirmed directly that the company will transition both Vega GPUs and the Ryzen line of processors to the 12nm LP process. However, it’s still not clear whether or not he meant that 12nm LP will be a shrink of Ryzen in 2018 (a "tick" equivalent if you will) or if Zen+/Zen 2 will also be using the 12LP process. Previously, AMD has implied that Zen 2 will use the 7nm process. The company has used both "Zen+" and "Zen 2" to refer to its next-generation die.
We expected AMD's next-generation 7nm Zen products to slug it out with Intel's 10nm Ice Lake processors next year, but it appears we will see 12nm LP against Intel's 10nm instead.
Earlier, we learned from Raja Koduri's letter to his team that we should expect "new wave of product excitement" in early 2018, which implies that is when we might see the first news of Vega with the 12nm LP process to emerge.
Nvidia’s Volta architecture is already shipping on TSMC’s 12nm FFN process this year, so on the GPU side Nvidia seems to be ahead on adopting new process technology right now. Of course, on the CPU side we also have Intel, which promised “real” 10nm chips for next year that should also be significantly ahead in performance and power efficiency compared to the 12LP process.
However, this is nothing really new, as Intel has been ahead in process technology for quite some time, and if anything, third-party foundry companies such as Samsung/Global Foundries and TSMC have closed the gap a little once they adopted the FinFET structure for their transistors. AMD's relatively quick transition to the 12nm LP processor is vital for the company as it seeks to compete with Intel's Coffee Lake processors and Nvidia's Volta GPUs.
MORE: Everything Zen: AMD Presents New Microarchitecture At Hot Chips
The transition to 12nm LP also reaffirms AMD's commitment to its partnership with Global Foundries. The companies signed a five-year wafer supply agreement recently, but there has been speculation that AMD might seek other partners for its forthcoming product generations.
We're digging in deeper for more information and will follow up as required.
If that was the case, can Vega use DDR4 system memory when it was designed for HBM2? As an engineer, do you just swap out the memory controllers and you're good to go?
For example, GCN 3rd Gen was used in Tonga (R9 285, 380(X) w/ GDDR5), Fiji (Fury (X)/ Nano w/ HBM), and Bristol Ridge APUs (DDR4).
Yes but wasn't the new architecture designed for HBM? The onboard graphics CPU from AMD uses DDR4 system memory which is slower than HBM...I'm also curious if there will be any memory on die dedicated to the graphics chip...if not, will performance be much lower using DDR4?
Yeah technology moves damn fast...makes me wonder why reprogrammable CPUs haven't been explored... imagine being able to recode older CPUs with newer instructions...that could extend the life of a CPU for many years...I do believe Intel explored this possibility but it was abandoned pretty quickly.
Vega's high bandwidth cache controller can already directly access system ram, and potentially storage, or even other pci-e devices. There's also already a Vega professional card coming that has 2TB of onboard NAND flash connected via AMD's infinity fabric.
But that doesn't magically extend the life of the CPU. You're still stuck on the same process (meaning same or similar clock speeds) with the same number of transistors. You could in theory reprogram parts of the FPGA to provide hardware acceleration for very specific tasks (which I believe is the idea behind the FPGA module on Xeons), but you can't just reconfigure the whole thing and get a faster CPU out of nowhere. Also, I'm pretty sure the FPGA required to implement a contemporary CPU would be much more expensive than the regular CPU would be.
It's also a BUDGET solution anyway since the amount of GPU die space is limited so it makes no sense to increase the cost.
Now for MOBILE devices that have no memory expansion it MAY make sense to use an APU or SoC (system on chip) that has HBM memory acting as a shared CPU/GPU memory though I suspect there's no current market for that approach