Intel (NASDAQ: INTC) may have had a financial black hole with all the tablet and phone spends over the past few years fighting the ARM incumbency. Obviously, the outlay was so bad that the whole division had to fold under the PC Client one, leaving the latter’s boss Kirk Skaugen with the tough job on the integration – or, most likely, pushing Core microarchitecture approach further down the price scale to counter the increasingly complex ARM cores.
On the other hand, the high end Enterprise division had yet another stellar year, with little competition to bother about. Xeons are everywhere, and approaching 95% of the worldwide server and related CPU market is about as good as it gets before it becomes an absolute monopoly. The cores are mature and well tuned, as well as the ecosystem from memory to I/O to boards and everything else that matters.
Also read: Intel’s David McCloskey Looks Ahead at 2015 and Back at 2014
However, there are clouds on the horizon for the 2015: ARM vendors are persistently trying to get into the server market, starting with low margin microservers mostly running Web 2.0 stuff where the big commercial software (un)availability is less of an issue. While this topic deserves a separate story, the focus there now is on improving the core throughput as well as cache, memory and interconnect bandwidth, things that ARM was sorely lacking until now – compared to both Intel x86 and other RISCs like MIPS or POWER, or even Chinese Alpha “Shenwei”. Having said that, I do feel that ARM will start making some tangible, but still small sized, dent into the server market in 2015, but one that will be very well marketed by the alliance vendors.
On the other hand, China Government’s punishing IBM with expulsion of its high-end systems after the NSA ‘disclosure’ has resulted in an interesting side development: licensing of the POWER8 and later POWER9 architectures and IP to the Chinese, now an officially done deal.
Will we soon see inexpensive Chinese POWER machines flooding the markets? I wouldn’t say so for another two years at least, until they are tried out in the internal China market first. But, once the strategies what is to be done are made public over the next half year or so, there could be some repercussion on the Xeon positioning mindshare. Mind you, POWER8 not only has the whole shebang of high end enterprise apps, but it is also the only core more efficient per-thread than the Xeon, and it does have the complete ultra high end ecosystems for memories, interconnects, and such – including NVlink shared memory low latency links between POWER8+ and POWER9 with Nvidia Pascal GPUs by 2017. By then Intel might update its Xeon Phi offerings with direct QPI shared memory links to their own Xeons too, though.
POWER is a RISC ISA, after all, just like Alpha or MIPS or ARM, so in principle, for the same process and transistor budget, it should be able to do more. The issue of ISA did haunt Intel for quite a while, although by now they have fine tuned it to the hilt.
On that subject, Haswell brought something really important to the table, overlooked by many: its AVX2 instruction extensions now handle the more common integer, not just floating point, operations. If in the future the address calculations are added to the roster, you pretty much don’t need the old base X86 set. Most of the software that matters is already AVX optimised, and more will follow. Would a SIMD and vector style pure AVX ISA, at some point, replace the old X86 within Intel?
In summary, no one can unseat Intel from its high end throne in the coming year, either, but the attacks from both the top, if IBM decides to milk the new found Chinese partnership to the fullest, and the bottom, if the ARM finally finds its competitive spot in the server arena, will be there more than before. Watch this space for more details on all these in the coming month.