Cavium is all about MIPS... ARM seemed to be kind of a redheaded stepchild over there. I ran into this last year, trying to get ARM support from them for a new product. Well, that and all the SOC documentation.,, much of which just not available, even to existing customers. Freescale seemed happy to have the business, they seem to be able to give top support to more than one architecture, and they have great specs, everything we needed.
Interesting that both AMD and AppliedMicro will be sampling their 28nm chips this year though this is the second generation for Applied. I agree with Karl that Broadcom will not be in contention till late 2015 based on their announcements last year.
Whatever happened to Cavium's thunder - they were to sample early half of this year but all seems to be quiet.
Samsung, TI, Huawei are all dark-horses and might surprise everyone. I don't see why Huawei's HiSilicon cannot license A57 cores from ARM. They will first enter the chinese market.
How about nVidia? After a 2 year radio silence, Denver finally made an appearance although mobile centric. Could they make a move in the server space.
What about Qualcomm?
The space is getting crowded but it is still very early in the space. Other than Applied, no one has working silicon yet. Q4 2014 will be interesting indeed.
I don't think BRCM will ship until 2015, Rick. They are using 14nm FinFet which is not available yet to them, and of course they are doing custom cores. Takes time! But they will be a serious player as this market evolves.
There will be others, but for competitive reasons, many are holding the infomration very close.
Also, remember that this is not all about hardware, it is also about software. And, there are many in the software ecosystem that have a vested interest in seeing more than one architecture thrive. Stay tuned for more information.
"However, I do believe the ARM-based processors will have an impact on future server designs and processor ASPs by the end of the year"
I think that all depends on how much Intel will lower the ASP of its C2000 server line and on what will be Q3 14nm Denverton and 14nm Broadell SOC. If these devices will are with a good interconnection fabric build in, i don't see much momentum of ARM in server space.
X86 ecosystem is too much comfortable and tool rich, ARM needs to prove to be substantiallybetter to do a dent in comodity server space. Unfortunately they are only at the beginning and there are many others, more advanced and estabilished RISC vendors who do not wantto losepotential customers. This is not Intel versus ARM but ARM versus the whole server industry, that is a different thing. All these projections about the future ARM penetration in server market are a little ridiculous or laughable.
Just now Intel owns the microserver space because was fast to deliver S1000 and C2000 lines. I don't think Intel will stop this strong action in near future, neither will sleep the big RISC vendors to maintain their customers or gain more.
"they do have differentiation through the Freedom Fabric and the SeaMicro servers"
yes this is the only chance, still Intel is not sleeping and it has its own 3D fabric solution for large installations of low power low performance SOCs.....same is valid for others large players in theserver market.
Strange enough, in consumer it is an Intel versus the world thing, in server space an ARM versus the world thing :).
I would agree that the number of unit shipments by any of the ARM server or server processor vendors that have are will be annnouncing this year is minimal because of the process of developing and selling servers into the market, especially large installations. However, I do believe the ARM-based processors will have an impact on future server designs and processor ASPs by the end of the year.
In regard to standard vp. custom IP, there are tradeoffs to both design choices. Note that even using standard CPU cores does not rule out customization through the other functions of the chip. In AMD's case, they do have differentiation through the Freedom Fabric and the SeaMicro servers.
Replay available now: A handful of emerging network technologies are competing to be the preferred wide-area connection for the Internet of Things. All claim lower costs and power use than cellular but none have wide deployment yet. Listen in as proponents of leading contenders make their case to be the metro or national IoT network of the future. Rick Merritt, EE Times Silicon Valley Bureau Chief, moderators this discussion. Join in and ask his guests questions.