Intel has finally begun shipping a 10nm processor, though this first Cannon Lake Core i3 is a tidbit ahead of what will be the main event next year. The chip-maker has been struggling to meet its own 10nm manufacturing promises, discovering that transitioning to the new architecture was trickier than it initially expected.
Indeed, Cannon Lake – the first chips to use the 10nm process – was previously expected back in 2017. At the start of last year, Intel CEO Brian Krzanich was promising 10nm within twelve months. Even that was a delay in itself: the chip company had first started suggesting that 10nm silicon would come in 2015.
Last month, Intel conceded that, again, its roadmap was more ambitious than it could actually deliver on. Rather than a full line-up of Cannon Lake processors, it admitted, there would be just one. “Intel is currently shipping low-volume 10 nm product and now expects 10 nm volume production to shift to 2019,” the company said at the time, though declined to say where it was shipping and who its hardware partner might be.
However the details have now emerged. The first Cannon Lake processor is the Intel Core i3-8121U, an 8th Generation Core chip that runs at 2.2 GHz. It has two cores and four threats, and supports Intel Turbo Boost 2.0 for up to 3.2 GHz bursts of speed.
There’s 4 MB of SmartCache, and a 15W TDP. The processor supports up to 32 GB of memory, with up to two memory channels up to 41.6 GB/s of memory bandwidth. It’ll work with DDR4-2400 and LPDDR4/x-2400 memory types. Intel says it supports Optane Memory, too, and up to sixteen PCI Express Lanes are supported, versus twelve for the old chip.
Clearly, though, this is just a morsel ahead of the main Cannon Lake meal. For a start, the Lenovo IdeaPad 330 notebook that currently uses this particular processor is only available in China at the moment. Its counterpart in the US uses existing chips.
Intel has blamed its ambitious generation transition for the issues getting Cannon Lake into production, with Krzanich admitting that the company has “bit off a little too much” with its goal of increasing density 2.7x over its current 14nm chips. In contrast, the chief executive argued, most industry transitions aim for sub-2x density improvements.
Krzanich also blamed problems with the photolithography technique that Intel is using for the new production, and suggested it had contributed to lackluster yields. The company had a solution, he said, but needed time to actually implement it. As such, we’re not expecting a significant influx of Cannon Lake until 2019.