Intel wants to go back to tick-tock model

These days, it's almost hard to remember but a decade ago Intel was famous for its "tick-tock" model. The basic premise here was a differentiation between a new node and a new architecture. A tick was the shrinking of the current CPU architecture to a new node, while the tock consisted of rolling out a new CPU architecture on the slightly more mature node. For many years, this model worked extremely well.

But then Intel's 10nm issues started. Initially, tick-tock became tick-tock-tock but many delays later the model has been completely abandoned. The failure to bring 10nm to the market has been one of Intel's biggest failures and has allowed AMD to catch up with the chip giant.

The 10nm chips were supposed to be available in 2016 but even today 10nm production is quite limited. High volume production isn't expected until later this year, with the arrival of the 10nm Alder Lake architecture.

Gelsinger promises return of Tick-Tock

During the Q&A session of yesterday's Engineering the Future event, new Intel CEO Pat Gelsinger promised the Tick-Tock model will be re-established as it's seen as a critical part of getting back to unquestioned CPU leadership. AnandTech says Gelsinger talked about yearly process node improvement, coupled with regular microarchitecture updates. It doesn't seem like an exact replica of the old model, but more of a commitment to get back to a regular cadence of process node and CPU architecture improvement.

On top of this commentary, Pat Gelsinger also stated that Intel’s CPU roadmaps are already baked in through 2021, 2022, and 2023. The company is thus looking to 2024/2025 for ‘unquestioned CPU leadership performance’, which traditionally means the fastest processor for single thread and multi-thread workloads. This is for sure a laudable goal, however Intel will also have to adapt to a changing landscape of chiplet processor designs (coming in 2023), enhancing on-die accelerators (GNA already present), and also what it means to have leadership performance – in the modern era, leadership performance doesn’t mean much if you’re also pushing lots of Watts. Intel stated that its 7nm process is now comfortably on track to deliver Meteor Lake, a client CPU using tiles/chiplets, in 2023, however we are likely looking to a 7nm variant or even external processes for a 2024/2025 product. Intel has also stated that it is looking to consider the core of its leading edge compute on external foundry processes, although one might argue that this doesn’t explicitly say ‘CPU’.

First look at Ponte Vecchio supercomputer chip

Intel talked about the tile-based design of the Meteor Lake processor, which will deliver a 7nm compute tile in 2023. Other than this, there was not a lot of talk about the chip giant's near-term roadmap. Unfortunately, we didn't hear any news about Alder Lake, a possible Alder Lake Refresh, or the Xe-HPG gaming GPUs.

Intel did show off its Ponte Vecchio supercomputer chip, VideoCardz offers some analysis. With Ponte Vecchio, Intel will compete with NVIDIA's Hopper, and to a smaller extend AMD's CDNA2, in the supercomputer market. One common theme here is that all three designs are based on chiplets, or tiles as Intel likes to call them.

Ponte Vecchio features the Xe-HPC architecture and will feature a total of 47 tiles. Combined, the tiles pack over 100 billion transistors. Important to note is that Ponte Vecchio is not a processor. The chip will be used in tandem with the "Sapphire Rapids" Xeon server processor.

Intel has working silicon of Ponte Vecchio, it will be rolled out later this year for the Aurora supercomputer:

Intel Ponte Vecchio is an advanced supercomputing accelerator that will be used in tandem with future Xeon architecture codenamed Sapphire Rapids. Intel already announced that the first supercomputer to feature Ponte Vecchio is Aurora. This supercomputer for Argonne National Laboratory is now expected to completed by the end of 2021. It will offer around 1 exaFLOPs of computing power.

Peta-flops in your palm. An ambitious dream. @intel silicon engineers along with our packaging alchemists made it come true in just 2 years! @pgelsinger's infectious energy for technology is a great accelerant for our march onwards to bring this magic to a cloud near you pic.twitter.com/C4LMuNrY2F

— Raja Koduri (@Rajaontheedge) March 24, 2021