When Intel launched Lakefield back in June 2020, it felt like the beginning of something new. ARM may have deployed big.Little a decade ago, but Intel and AMD had previously stuck to multi-core CPUs built on a single CPU design. Lakefield combined one Ice Lake “big” CPU cores with four Tremont “little” CPU cores. Lakefield targeted the low-power market, with a TDP of just 7W, and was intended to compete with ARM products featuring similar power consumption envelopes. Now, Intel has announced an early retirement for the CPU. Lakefield is entering end-of-life barely a year after launch.
Intel’s explanation for the cancellation is straightforward: “Market demand for the products listed in the ‘Products Affected/Intel Ordering Codes’ table below have shifted to other Intel products.” But Lakefield was unique, as far as Intel was concerned. This isn’t the first time Intel has struggled to compete in the low-power market.
Intel’s power/performance curves for Lakefield versus Ice Lake.
In the run-up to the Windows 8 launch, Intel expected its Clover Trail platform to quickly seize market share in what was supposed to be a rapidly expanding landscape for Windows 8-based tablets and convertible devices. The explicit thinking at the time was to leave the low-end market to ARM and establish Intel as the upper-tier vendor of choice for x86 tablets.
Part of the reason this plan failed is undoubtedly due to the poor reception Windows 8 received, but Clover Trail-powered devices were often priced similarly to mainstream, more powerful x86 laptops. Intel took had an x86 tablet business for a little while, but it sustained its market share by shipping products contra revenue. It was a deliberate choice to lose money on shipments in exchange for building market share. When Intel moved away from this strategy, the company’s tablet share fell swiftly.
Lakefield may have suffered from some of the same problems. Intel and its OEM partners sold the chip into expensive systems priced between $800-$1,200. That sets the chip up to compete against high-end Intel and AMD systems with much larger batteries and far better performance. Compared with these systems, Lakefield can’t win. Preliminary data suggests Windows 11 performs better on hybrid CPUs, but that wasn’t enough to make Lakefield viable, apparently.
Alder Lake will be Intel’s second stab at the hybrid model, this time with up to eight “big” cores and small cores based on Gracemont. Gracemont is expected to field a 64KB L1 cache (up from 32KB on Tremont) with DDR5, PCIe 4.0, and AVX/AVX2 support.
Intel may have better luck positioning Alder Lake than Lakefield. Switching to little cores for idle power should save energy compared with conventional Intel CPUs, and the systems OEMs build are less likely to face off with hardware that dramatically outclasses them at the same price point. The fact that Lakefield is being canned after such a short time suggests very little uptake or interest in the chip. That doesn’t mean Alder Lake will run into similar problems, but it speaks to a consistent issue Intel has had winning market share with low-power chips. It’s not that Intel can’t build them. It’s that they don’t seem to wind up in the hardware people actually want to buy at competitive prices.
Now Read:
- Does Intel’s Lakefield SoC Measure Up?
- Foveros, EMIB, ODI: Intel’s Architecture Day Was All About Interconnects
- Does Intel’s Lakefield SoC Measure Up?
No comments:
Post a Comment