Intel is attempting to bring in a truly new CPU core design after a gap of nearly five years. However, it appears the company’s challenges with 10nm could be limiting its design and deployment choices significantly. Intel’s new microarchitecture, internally referred to as ‘Willow Core’, could simply be an adaption of an older microarchitecture, indicates a Tweet from a reliable source.
Intel’s ‘Sunny Core’ will soon be succeeded by the supposedly new ‘Willow Cove’ CPU core. Needless to add, this new microarchitecture for the next-gen Intel CPUs is supposed to be a big milestone. In fact, the Intel Willow Cove CPU design will be the company’s first truly new CPU core design in close to 5 years. However, instead of confidently moving on to the next evolutionary step in CPU die size, the company is merely attempting to optimize the older generation of die sizes.
In simple words, Intel is attempting to play with the tried and tested, but increasingly obsolete 14nm production process, instead of moving onto the 10nm fabrication process. It is important to note that while Intel is clearly struggling with several issues pertaining to the shift to the newer fabrication techniques, its main rival has already moved on to a much more advanced fabrication process featuring significantly reduce silicon die sizes.
Intel Is Working To Adapt “Willow Cove” CPU Cores Onto A 14nm Microarchitecture?
Although the Willow Cove CPU core design succeeds Sunny Cove, it appears Intel is working to adapt the Willow Cove CPU cores onto a 14 nm microarchitecture. Intel’s Sunny Core CPU microarchitecture is implemented in the 10 nm “Ice Lake” microarchitecture. Additionally, the Willow Cores CPU microarchitecture was expected to debut with the 10 nm+ “Tiger Lake” CPUs. But it seems that might not happen, at least in the near future.
— TechPowerUp (@TechPowerUp) December 2, 2019
A quite popular and reportedly reliable Twitter User, with Twitter handle @chiakokhua, attempted to make sense of some highly technical documents. The Twitter user, a retired VLSI engineer with a proven track record about CPU microarchitecture news, claims that “Rocket Lake” is essentially a 14 nm adaptation of “Tiger Lake.” Moreover, the iGPU on this supposedly new generation of CPUs has been shrunk significantly. Apparently, Intel made the design choice to make room for larger CPU cores. In other words, Intel’s inability to move onto 10nm doe sizes is forcing the company to take some unfavorable design choices.
The Gen12 iGPU on “Rocket Lake-S” will feature just 32 Execution Units (EUs). Needless to add, this is exceptionally lower than Tiger Lake CPUs. With 96 EUs, the Tiger Lake CPUs have three times the power. Incidentally, “Rocket Lake” also replaces “Tiger Lake’s” FIVR (fully-integrated voltage regulation) with a conventional SVID VRM architecture. As experienced readers and those who closely follow CPU news will quickly realize, Intel appears to have compromised heavily.
Intel Rocket Lake CPUs To feature Lower Core Count For Higher IPC Gains?
It has been previously reported that “Rocket Lake-S” silicon has up to 8 CPU cores. Even the predecessor of this CPU, the “Comet Lake-S,” sported up to 10 cores. This is quite a step back, especially after considering that these CPUs are still being produced on the 14nm fabrication process. However, Intel justifies that the lowered core count helped the company achieve some much-needed IPC gains.
— Tom's Hardware (@tomshardware) December 2, 2019
Intel’s Rocket Lake-S CPUs could be Intel’s first major IPC uplift on the Intel platform since “Skylake.” However, it is not clear how well these CPUs, which are still being manufactured on the 14nm process, be received by the consumers.
It is amply clear that it is Intel that has to play catch-up with its main rival AMD. The Ryzen and Threadripper CPUs, as well as the eagerly anticipated ‘Milan’ could help ensure AMD continues to lead the race of high-end CPUs for quite some time.