AMD announced it will design processors based on 64-bit ARM technology in addition to its x86 processors for the cloud and data center servers. AMD will be the first company to provide both x86 and ARM solutions to the server market, and the first ARM-based server CPU is scheduled for 2014.
There aren’t much details at this time, but AMD did say it will be take advantage of AMD SeaMicro Freedom supercompute fabric in order to link CPU, Ethernet and Storage and the processor will be sold as an AMD Opteron CPU. Interestingly, the company does not plan to design 64-bit ARM processor for the client market (e.g. Windows RT) for the time being.
Red Hat and AMD partnered for the server software, and Red Hat has started to work to bring support for the next generation of 64-bit ARM server processors to the Fedora Project include 64-bit ARM OpenJDK. More details are available over here on Red Hat 64-bit ARM plans with AMD (and others), but it’s mostly a lot of blabla.
If you are interested in this announcement, there’s a 1 hour webcast with more details and a panel discussing the news with ARM, Facebook, Amazon, Dell, Red Hat and AMD. I watched it, and the part I found the most interesting was the slide explaining where ARM or x86 servers would be used.
ARM based servers would mainly be used for data center and cloud applications where compute/dollars and compute/watt are critical. APU (CPU + GPU + Fabric) would also be used in the server market, as the GPU can be used to accelerate media rendering (GPGPU). Compute Clusters would still be reserved to the most powerful x86 servers.
If you are looking for technical details, don’t watch the webcast (it’s really boring). But if you are interested to know more about the business side as well as the technical and business reasons for which companies may want to switch to ARM/AMD server it might be interesting to you.
|Support CNX Software - Donate via PayPal or become a Patron on Patreon|